JPH04235319A - Cavitation-noise detecting apparatus - Google Patents

Abstract

PURPOSE:To detect cavitation noises stably regardless of the surrounding environment, aging and the like by dividing the broad-band signal generated in the vicinity of propellers into a high-frequency component and a low-frequency component, and using the level fluctuation ratio of both components for detecting the generation of the cavitation noises. CONSTITUTION:The broad-band signals received with a sensor 10 are divided into high-frequency components and low-frequency components with a frequency analyzing means 20. The ratio of the level fluctuations of both components is obtained with a level-fluctuation-ratio computer 30. A weight-coefficient judging device 41 outputs the weight coefficient corresponding to the level fluctuation ratio based on the level fluctuation ratio. The weight coefficient is compared with the judging threshold value outputted from a judging-threshold- vale computer 43 in a comparator 42, and cavitation noises are detected. In the judging-threshold-value computer 43, the judging threshold value is updated or continued by using the output of the comparator 42 and the level fluctuation ratio.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a cavitation noise detection device that uses signals from a sensor installed near a propeller of a ship to detect cavitation noise generated as the propeller rotates.

0002

2. Description of the Related Art Conventionally, as a method for detecting cavitation noise generated as a propeller of a ship or the like rotates, a device is known that detects cavitation noise based on the level of a wideband signal received by a sensor near the propeller. An example of its configuration is shown in FIG.

FIG. 2 is a functional block diagram showing a conventional cavitation noise detection device. This cavitation noise detection device has a sensor 1 installed near a propeller, and the sensor 1 includes an amplifier 2, a band-limiting filter 3, a sampler 4, a square calculator 5, an integrator 6, and a comparator 7.
, and output terminal 8 are connected in cascade.

When a propeller of a ship or the like rotates, a sensor 1 receives a broadband signal emitted from the propeller. The received wideband signal is amplified to an appropriate level by the amplifier 2, and unnecessary frequency components are removed by the band limit filter 3. The output of the band-limiting filter 3 is converted into a digital signal by a sampler 4 and sent to a square calculator 5.

[0005] The square calculator 5 calculates the square value of the output signal of the sampler 4 to obtain the intensity P(k) of the broadband signal generated near the propeller at time k (this is called instantaneous power), and integrates the signal. Send to vessel 6. The integrator 6 integrates the instantaneous power P(k) to obtain the long-term integrated power Pa(k) of the received broadband signal, and sends it to the comparator 7.

The comparator 7 compares the long-term integrated power Pa (k) output from the integrator 6 with a threshold value THL for determining the occurrence of cavitation noise, and determines whether the long-term integrated power Pa (k) is If the threshold value THL is exceeded,
The cavitation noise detection signal is output to the output terminal 8.

[0007]

[Problems to be Solved by the Invention] However, in the device having the above configuration, the occurrence of cavitation noise is detected using only the long-term integrated power Pa (k) of the received wideband signal, so the following problems occur. there were.

(1) When the level of background noise, such as mechanical noise from ships or underwater noise, is high, the long-term integrated power Pa (k) of the received wideband signal is compared even if cavitation noise is not generated. Threshold value THL of device 7
, and the number of false alarms increases.

(2) When the characteristics of a ship or the like change over time, the threshold value THL set at the time of manufacture no longer matches, resulting in an increase in false alarms and a decrease in the detection rate of cavitation noise.

(3) The input to the comparator 7 is the threshold value THL
When moving up and down around , detection and non-detection judgments are repeated, making the cavitation noise alarm unstable.

[0011] Therefore, it has been difficult to obtain a cavitation noise detection device that is technically fully satisfactory.

[0012] The present invention addresses the problems that the prior art had, in that false alarms increase when the background noise level is high, and that false alarms occur because the threshold value does not match the cavitation noise generation level due to secular changes in the characteristics of ships, etc. To provide a cavitation noise detection device that solves the problem of an increase in the number of alarms and a decrease in the detection rate, and the problem that an alarm for cavitation noise generation becomes unstable when the input to a comparator moves up and down around a threshold value. It is something.

[0013]

[Means for Solving the Problems] In order to solve the above-mentioned problems, the first invention focuses on the fact that when cavitation noise occurs, the level fluctuation of the frequency component of the broadband signal increases, frequency dividing means for dividing a wideband signal received by the sensor into a low frequency component and a high frequency component; and a level fluctuation ratio calculation means for calculating a level fluctuation ratio of the divided low frequency component and high frequency component. In the cavitation noise detection device for detecting cavitation noise generated as the propeller rotates based on the ratio, weighting coefficient determining means;
A comparison means and a determination threshold value calculation means are provided.

The weighting coefficient determining means divides the possible range of the level fluctuation ratio into a plurality of regions and assigns a weighting coefficient to each region, and calculates the level fluctuation ratio output from the level fluctuation ratio calculating means. It has a function to output weighting coefficients according to the The comparison means compares the output of the weighting coefficient determination means with a determination threshold, and detects cavitation noise based on the comparison result. The determination threshold calculation means has a function of setting the determination threshold by calculation based on the output of the level variation ratio calculation means and the output of the comparison means and providing it to the comparison means.

In the second invention, the determination threshold calculation means of the first invention is configured as follows. That is, the determination threshold calculation means compares the level fluctuation ratio at time k outputted from the level fluctuation ratio calculation means with the level fluctuation ratio stored at the time of the previous update of the determination threshold, and the comparison result of the comparison means. It has a threshold resetting function that updates or continues the determination threshold at time k+1 based on the output.

By the threshold resetting function of the judgment threshold calculation means, when the cavitation noise is detected, the level fluctuation ratio at the time k becomes 1/P of the level fluctuation ratio stored at the time of the previous judgment threshold update. (However, P is 1
Only when the value is less than or equal to the above constant), the time k+
1 is updated to the weighting coefficient at the time k, and the level fluctuation ratio at the time of the update is stored in preparation for the next resetting of the determination threshold.

[0017]

According to the first aspect of the invention, since the cavitation noise detection device is configured as described above, when cavitation noise is generated due to rotation of the propeller, its broadband signal is received by the sensor. The frequency dividing means divides the broadband signal received by the sensor into a high frequency component and a low frequency component, and the level fluctuation ratio calculating means calculates the level fluctuation ratio between the divided low frequency component and the high frequency component. Cavitation noise generation is detected by the weighting coefficient determining means, the comparing means, and the determination threshold calculating means using the level fluctuation ratio.

That is, the weighting coefficient determining means divides the possible range of the level fluctuation ratio output from the level fluctuation ratio calculating means into a plurality of regions, and assigns a weighting coefficient to each of the divided regions. When the level fluctuation ratio is output from the fluctuation ratio calculation means, a weighting coefficient corresponding to the level fluctuation ratio is outputted to the comparison means. The output of the weighting coefficient determining means and the determination threshold set by the determination threshold calculating means are compared by the comparing means, and cavitation noise is detected based on the comparison result. The determination threshold calculation means updates or continues the determination threshold by calculation based on the output of the level variation ratio calculation means and the output of the comparison means, and provides the determination threshold to the comparison means.

As a result, even if the input to the judgment threshold calculation means moves up and down around the judgment threshold, the cavitation noise generation alarm is stabilized, and the cavitation noise is stably suppressed regardless of the surrounding environment or changes over time. can be detected.

In the second invention, the threshold resetting function of the determination threshold calculation means updates or continues the determination threshold given to the comparison means, so that the determination threshold can be set accurately by simple calculation. Detection of cavitation noise becomes more stable. Therefore, the above problem can be solved.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a functional block diagram of a cavitation noise detection apparatus showing a first embodiment of the present invention.

This cavitation noise detection device has a sensor 10 such as a hydrophone provided near a propeller, and a frequency dividing means 20 is connected to the sensor 10 via an amplifier 11. The frequency dividing means 20 has a function of dividing the output of the amplifier 11 into a low frequency component and a high frequency component. This frequency dividing means 20 has a high frequency side band limiting filter 21 and a low frequency side band limiting filter 22 that remove unnecessary frequency components of the received signal,
Instantaneous power calculators 23, 24 and variance value calculators 25, 26 are connected to each output side thereof. The high-frequency side band-limiting filter 21 and the low-frequency side band-limiting filter 22 are configured with a plurality of band pass bands so as to easily capture the characteristics of cavitation noise. The band-limiting filters 21 and 22 may be configured with analog band-pass filters, or may be configured as digital filters by digitizing the output of the amplifier 11. Each instantaneous power calculator 23, 24 calculates instantaneous power based on the output of each band-limiting filter 21, 22. Each variance value calculator 25, 26 uses the instantaneous power output from each instantaneous power calculator 23, 24 to calculate the average value and the average variance of each frequency band.

A level fluctuation ratio calculator 30, which is a level fluctuation ratio calculating means, is connected to the output sides of the variance value calculators 25 and 26, and a determining means 40 is further connected to the output side thereof. The level fluctuation ratio calculator 30 is the variance value calculator 25
, 26, calculates a level fluctuation ratio, and supplies the level fluctuation ratio to the determination means 40.

The determining means 40 is a level fluctuation ratio calculator 30.
It has a function of detecting cavitation noise based on the level fluctuation ratio output from the output terminal 50 and outputting the detection result to the output terminal 50. This determining means 40 has a weighting coefficient determining unit 41 which is a weighting coefficient determining unit, and a comparator 42 which is a comparing unit is connected to its output side, and an output terminal 50 is further connected to its output side. There is. Further, a determination threshold value calculator 43, which is a determination threshold value calculation means, is connected between the input side of the weighting factor determiner 41 and the output side of the comparator 42.

The weighting coefficient determiner 41 has a function of generating a weighting coefficient according to the level fluctuation ratio based on the level fluctuation ratio outputted from the level fluctuation ratio calculator 30 and providing it to the comparator 42. . The determination threshold calculator 43 is a circuit that sets a determination threshold for detecting cavitation noise based on the level variation ratio and the output of the comparator 42, and supplies it to the comparator 42. The comparator 42 is a circuit that compares the determination threshold from the determination threshold calculator 43 and the output of the weighting coefficient determiner 41, and detects cavitation noise based on the comparison result.

FIG. 3 is a diagram illustrating the weighting coefficient determiner 41 of FIG. 1, in which the horizontal axis represents the input level fluctuation ratio, and the vertical axis represents the output weighting coefficient ωl (l = 0, 1, . . . , N−). 1) are taken respectively.

The weighting coefficient determiner 41 divides the possible range of the level fluctuation ratio into N regions, and assigns a weighting coefficient ωl (l = 0, 1, . . . , N-1) to each divided region. Then, the weighting coefficient ωl corresponding to the input level fluctuation ratio is output. Here, the number N of divided regions takes a value of 3 to 4, for example. In FIG. 3, the case where N=3 is shown.

FIG. 4 is an operation flowchart of the determination means 40 of FIG. 1, and the operation of the cavitation noise detection apparatus of this embodiment will be explained with reference to FIGS. 1 and 3 including this figure.

In FIG. 1, when a propeller of a ship or the like rotates and a broadband signal is generated near the propeller, the broadband signal is received by a sensor 10, amplified to an appropriate level by an amplifier 11, and then converted into a high frequency band. Restriction filter 21
and is input to the low frequency side band limiting filter 22. The high-frequency side band-limiting filter 21 and the low-frequency side band-limiting filter 22 remove unnecessary frequency components of the received signal, divide it into a plurality of frequency bands, and send each band to instantaneous power calculators 23 and 24, respectively.

The instantaneous power calculator 23 on the high frequency side calculates the square value Phj(k) of the output of the high frequency side band limiting filter 21 for each band (this is called instantaneous power. j = 1, 2).
,...,Nh) and send the result to the variance value calculator 25.
Output to. The variance value calculator 25 calculates the average value mhj(k) of each band based on the instantaneous power Phj(k) from the instantaneous power calculator 23 using the following equation 1 using the exponential smoothing method.
, and the average variance δ2 hj(k).

[0031]

[Math 1]

mhj(k)=(1−α)mhj(k−1)+α・
Phj (k) δ2 hj (k) = (1-α) δ2 hj
(k-1)+α[Phj(k)-mhj(k-1)
]2 + (1-α) [m
hj(k)-mhj(k-1)]2
However, α is a constant, j = 1, 2,..., Nh Furthermore, the variance value calculator 25
Then, based on the average variance δ2 hj(k) of each band,
According to the following equation 2, the average of the dispersion on the high frequency side δ2 ah (k) is calculated by taking the average between bands, and the δ2 a
h(k) is output to the level fluctuation ratio calculator 30.

[0033]

[Math 2]

Similarly, the output of the low frequency side band limiting filter 22 is also
The instantaneous power calculator 24 calculates the square value Plj(k)(j
= 1, 2, ..., Nl) is obtained. Furthermore, the dispersion value calculator 26 calculates the mean value mlj(k) of each band, the mean dispersion δ2 lj(k), and the mean dispersion δ2 on the low frequency side.
After determining al(k), the δ2 al(k) is output to the level fluctuation ratio calculator 30.

The level fluctuation ratio calculator 30 calculates the outputs δ2 ah(k) and δ2 al(
k), the level fluctuation ratio γ(k) and its average γa
(k) is calculated using the following equation 3, and the average level fluctuation ratio γa
(k), the weighting coefficient determiner 41 and the determination threshold value calculator 43
Output to.

[0036]

[Math 3]

γ(k)=δ2 ah(k)/δ2 al(k) (or δ2 al(k)/δ2 ah(k)) γa (k)
= (1-β)・γa (k-1)+β・γ(k) However,
β: Constant Next, the operation of the determining means 40 will be explained according to FIG.

Weighting coefficient determiner 41 in determining means 40
substitutes the weighting coefficient ωl (k) corresponding to the level variation ratio average γa (k) in step S60 into CAV1 in step S61, and outputs CAV1 to the comparator 42. To make the level fluctuation ratio correspond to the weighting coefficient, for example, as shown in FIG. 3, the possible range of the level fluctuation ratio is divided into N regions, and the weighting coefficient ωl (
l = 1, 2, ..., N-1). Here, N takes a value of 3 to 4.

[0039] In step S62, the comparator 42
The output CAV1 from the weighting factor determiner 41 and the output CAV0 at time K-1 from the determination threshold value calculator 43 are compared to determine whether the following condition of Equation 4 is satisfied.

[0040]

[Math 4]

CAV1>CAV0 (When γ(k)=δ2 ah(k)/δ2 al(k)) However, CAV0; Judgment threshold (initially CAV0=ω
l) If the condition of Equation 4 is satisfied, the comparator 42 outputs the cavitation noise detection signal to the output terminal 50 in step S63, and also outputs the cavitation noise detection signal to the determination threshold calculator 43. Judgment threshold calculator 4
3, when the comparator 42 detects cavitation noise, when CAV1>CAV0, the determination threshold CAV0 at time k+1 is output to the comparator 42 as the weighting coefficient CAV1 at time k, and the average γa of the level fluctuation ratio at time k (k
) is stored as γcav0.

On the other hand, in step S62, the comparator 4
2 does not detect cavitation noise, there is no alarm output in step S65, and the process advances to step S66. In step S66, the average γa (k) of the level fluctuation ratio at time k is compared with γcav0 stored when the determination threshold CAV0 was updated, and the following equation 5 is calculated.
Determine whether or not the condition is satisfied.

[0043]

[Math 5]

[0044] However, when the condition of P: proportionality constant 5 is satisfied, that is, the average level fluctuation ratio γa (k) is the level fluctuation ratio γcav at the time of alarm output.
If the value is equal to or less than 1/P of 0, the determination threshold calculator 43 determines the determination threshold CA in step S64 in the same manner as described above.
Lower V0 to the weighting coefficient of time k and output it to the comparator 42,
The average level fluctuation ratio γa (k) at time k is γcav0
, and prepare for the next cavitation noise detection. Here, the proportionality constant P is a constant of 1 or more, and is determined by the detection level of cavitation noise, the method of dividing the region of the weighting coefficient, and the like.

If the condition of Equation 5 is not satisfied in step S66, the determination threshold value calculator 43 does not update the determination threshold value, and proceeds to the next process in step S67. Therefore, even if the average level variation ratio γa (k) varies within a range of γcav0/P or more, detection and non-detection operations are not repeated, thereby making it possible to stably detect cavitation noise.

Furthermore, in this embodiment, when the level of background noise is high or when the characteristics of a ship or the like change due to aging, the determination threshold value is updated by the determination threshold value calculator 43 accordingly, so that false alarms are avoided. This makes it possible to prevent cavitation noise and improve the detection rate of cavitation noise.

FIG. 5 is a functional block diagram of a cavitation noise detection apparatus showing a second embodiment of the present invention, in which elements common to those in FIG. 1 are given the same reference numerals.

[0048] In this cavitation noise detection device,
In place of the frequency dividing means 20 in FIG. 1, a frequency dividing means 70 having a different configuration is provided.

The frequency dividing means 70 has a function of dividing the output of the amplifier 11 into low frequency components and high frequency components, and includes a band limit filter 71 for removing unnecessary frequency components from the output of the amplifier 11. A sampler 72 that converts the output of the band-limiting filter 71 into a digital signal is connected to the output side of the band-limiting filter 71, and a frequency analyzer 73 is further connected to the output side of the sampler 72. The frequency analyzer 73 uses a technique such as fast Fourier transform on the output of the sampler 72 to analyze a plurality of frequencies bi in a predetermined analysis width.
It has a function of dividing into n, and a bin selector 74 is connected to its output side.

The bin selector 74 selects a frequency bin having characteristics of cavitation noise from the output of the frequency analyzer 73, switches the frequency bin to a high frequency side and a low frequency side, and outputs the selected frequency bin. be. bin selector 7
A dispersion value calculator 77 is connected to the high frequency side output terminal of 4 via an instantaneous power calculator 75, and a dispersion value calculator 77 is connected to the low frequency side output terminal via an instantaneous power calculator 76.
8 are connected. The instantaneous power calculators 75 and 76 are
This circuit calculates the instantaneous power from the output of the bin selector 74 and supplies the instantaneous power to the variance value calculators 77 and 78. The variance value calculators 77 and 78 are the variance value calculator 75 and the variance value calculator 75 in FIG.
This circuit is similar to 76, and the level fluctuation ratio calculator 30 is connected to its output side.

Next, the operation will be explained.

A broadband signal generated by the rotation of the propeller is received by the sensor 10 and amplified by the amplifier 11 to an appropriate level. The output of the amplifier 11 has unnecessary frequency bands removed by a band-limiting filter 71, and then digitized by a sampler 72 and sent to a frequency analyzer 73.

The frequency analyzer 73 divides the broadband signal digitized by the sampler 72 into a plurality of frequency bins at a predetermined analysis width using a technique such as fast Fourier transform, and sends the result to a bin selector 74. Output to. bin
The selector 74 selects a frequency bin having characteristics of cavitation noise from the signal from the frequency analyzer 73, and when the bin belongs to the high frequency side, the selector 74 uses the result to calculate the instantaneous power on the high frequency side. 75, and the b
When in belongs to the low frequency side, the result is output to the instantaneous power calculator 76 on the low frequency side.

Here, as an example of the frequency bin selection method, there is a method in which cavitation noise is analyzed in advance and a bin number in which the characteristics of cavitation noise appear largely is set in advance.

In the instantaneous power calculator 75 on the high frequency side, b
The signal x(fhj) from the in selector 74 (j = 1, 2
, .

[0056]

[Math 6]

Phj(k)=|x(fhj)|2 However, j = 1, 2,..., Nh Similarly, in the low frequency side instantaneous power calculator 76,
Signal x(flj) from selector 74 (j = 1, 2,...
, Nl), and the instantaneous power Plj(k) on the low frequency side is
is calculated and output to the variance value calculator 78. The level fluctuation ratio of the outputs of these variance value calculators 77 and 78 is determined by the level fluctuation ratio calculator 30, as in the first embodiment.
Cavitation noise is detected by the determining means 40. Therefore, almost the same advantages as in the first embodiment can be obtained.

Note that the present invention is not limited to the above-mentioned embodiments, and various modifications are possible. Examples of such modifications include the following.

(a) In the first embodiment, the variance value calculators 25 and 26 calculate according to equation 1 using the exponential smoothing method.
Average value mhj (k) and average variance α2 hj (k)
However, a simple averaging method or the like may be used in which N samples are collected in the time direction, and the sum is simply added and multiplied by 1/N.

(b) The frequency dividing means 25, 70 in FIGS. 1 and 5 may have a circuit configuration other than that shown. Furthermore, each circuit block in FIGS. 1 and 5 may be configured to be executed by program control using a processor, etc., instead of being configured as an individual circuit using an integrated circuit or the like. If executed under such program control, the circuit configuration can be simplified.

[0061]

Effects of the Invention As explained in detail above, according to the first invention, a sensor The broadband signal received near the propeller is divided into a high frequency component and a low frequency component by a frequency dividing means, and a level fluctuation ratio is calculated from the divided high frequency component and low frequency component by a level fluctuation ratio calculation means.
Cavitation noise is detected using the level fluctuation ratio. That is, the weighting coefficient determining means outputs a weighting coefficient from the level fluctuation ratio, the comparing means compares the weighting coefficient with a determination threshold, and based on the comparison result, cavitation noise is detected. At this time, the determination threshold given to the comparison means is updated or continued by the determination threshold calculation means using the output of the comparison means and the level fluctuation ratio.

[0062] Therefore, when the level of background noise is high or when the characteristics of a ship, etc. change due to aging, the judgment threshold changes accordingly, making it possible to prevent false alarms and improve the detection rate of cavitation noise occurrence. improves. Furthermore, when the input to the comparison means moves up and down around the judgment threshold, the judgment threshold is not updated, which prevents the instability of cavitation noise alarms and provides a highly reliable cavitation noise detection device. It will be done.

According to the second aspect of the invention, the threshold resetting function of the determination threshold calculation means updates or continues the determination threshold, making it possible to set the determination threshold more accurately, thereby reducing ambient noise. Cavitation noise can be detected more stably, almost unaffected by the level of noise and aging of the ship.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram of a cavitation noise detection device showing a first embodiment of the present invention.

FIG. 2 is a functional block diagram of a conventional cavitation noise detection device.

FIG. 3 is a diagram illustrating the weighting coefficient determiner of FIG. 1;

FIG. 4 is an operation flowchart of the determining means in FIG. 1;

FIG. 5 is a functional block diagram of a cavitation noise detection device showing a second embodiment of the present invention.

[Explanation of symbols]

Claims (2)

[Claims] 1. Frequency dividing means for dividing a wideband signal received by a sensor provided near a propeller into a low frequency component and a high frequency component, and a level for calculating a level fluctuation ratio of the divided low frequency component and high frequency component. fluctuation ratio calculation means, and detects cavitation noise generated as the propeller rotates based on the level fluctuation ratio, the cavitation noise detection device comprising: dividing a range in which the level fluctuation ratio can take into a plurality of regions; Weighting coefficient determining means assigns a weighting coefficient to each of the regions and outputs a weighting coefficient according to the level fluctuation ratio outputted from the level fluctuation ratio calculating means; and an output of the weighting coefficient determining means and a determination threshold. and a comparison means for detecting cavitation noise based on the comparison result; and based on the output of the level fluctuation ratio calculation means and the output of the comparison means, the judgment threshold is set by calculation and provided to the comparison means. 1. A cavitation noise detection device comprising: determination threshold calculation means. 2. The cavitation noise detection device according to claim 1, wherein the determination threshold value calculation means stores the level fluctuation ratio at time k outputted from the level fluctuation ratio calculation means and the previous determination threshold value at the time of updating. Based on the comparison result with the level fluctuation ratio and the output of the comparison means,
It has a threshold resetting function that updates or continues the determination threshold at time k+1, and when the cavitation noise is detected by the threshold resetting function, the level fluctuation ratio at the time k is the same as that of the previous determination threshold update. 1/P of the level fluctuation ratio memorized at the time (P is a constant of 1 or more)
Only when the value becomes less than double, the determination threshold at the time k+1 is updated to the weighting coefficient at the time k, and the level fluctuation ratio at the time of the update is stored in preparation for resetting the determination threshold next time. Cavitation noise detection device.