JPH08271330A - Monitoring method of equipment by acoustic analysis - Google Patents

Abstract

PURPOSE: To eliminate a false abnormal sound and detect only the true abnormal sound of an equipment by judging an abnormality only when the condition where the sound pressure level is above a prescribed value is continued for a prescribed time. CONSTITUTION: Only when the total sound pressure level or the sound pressure level of a specified frequency exceeds a prescribed value, and the state where this sound pressure level is above a prescribed value is continued for a prescribed time, an equipment to be monitored is judged abnormal. Therefore, all false abnormal sounds having high sound pressure levels such as human voice, an impact sound at the collision of a matter, exhaust sound of an air valve, and the exhaust sound of a drain are eliminated. Each sound pressure level measured value SPL of these false impact sounds is high by 10-20db. In contrast to this, a true equipment abnormal sound is a relatively continued sound, and other sounds are sounds continued only for a short time (about several seconds). Thus, when the value of secondary alarm judging frequency N is properly selected, the false abnormal sound which is not the true equipment abnormal sound can be eliminated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an acoustic analysis method for monitoring equipment or a group of equipment scattered over a wide area by using acoustic analysis, and more particularly to a method for monitoring equipment such as a human voice or a sound hit by an object. The present invention relates to a method of monitoring equipment by acoustic analysis that can eliminate false abnormal sounds and detect only true abnormal sounds of equipment.

[0002]

2. Description of the Related Art Generally, vibration, temperature, electric current, sound or the like is used for monitoring the state of equipment relating to the production of steel. Among these state monitoring methods, the method of capturing the vibration emitted from the equipment and monitoring it while diagnosing the state of the equipment (hereinafter referred to as the vibration method) is based on the analysis and evaluation technology of the detected signal. It has been established and it is known that the target equipment can be accurately monitored.

However, in such a vibration method, there is a problem in that a vibration pickup must be attached to each of the facilities to be monitored. Therefore, in the case of iron and steel production equipment, the diagnostic method using the vibration method is put to practical use only for the most important equipment to be monitored, and the diagnostic method using the vibration method is installed in incidental equipment scattered over a wide area. Didn't. That is, in the case of incidental equipment scattered over a wide area, the number of measurement points becomes enormous, so attaching vibration pickups to all of these measurement points will result in a very high total cost of the diagnostic device using the vibration method, and in terms of cost. It had not been put to practical use because it had problems.

On the other hand, a method of observing the state of the facility while diagnosing the state of the facility by utilizing the sound emitted from the facility (hereinafter referred to as acoustic method) has a considerably wide range if there is one sensor, that is, one microphone. It is known that an abnormality can be detected in the equipment and the total price of the diagnostic device can be provided relatively cheaply. It should be noted that the "sound emitted from the equipment" here is limited to the sound radiated into the space as a sound wave, and the AE signal propagated in the solid, that is, the acoustic emission signal is not considered.

As a conventional example using such an acoustic method, the following technique has been already known before the application of the present application.

That is, to accurately detect the moment when the rolling mill bites the rolled material to improve the strip passing property is disclosed in, for example, Japanese Patent Laid-Open No. 59-150611 as a method for detecting the biting of the rolling mill.
No. 6,086,045.

Further, it is known as a method for detecting the situation in the blast furnace to detect the situation of gas leakage in the furnace near the furnace wall and the situation of deposits on the inner surface of the furnace wall by detecting the situation in the blast furnace. 5
It is disclosed in Japanese Patent Publication No. 9-173212.

Further, a technique which further improves the method for extracting the sound generated in the furnace during the blowing of the converter and measuring the forming situation of the forming slag from the change of the sound level of a specific frequency is as follows.
As a converter acoustic measurement method, for example, JP-A-63-5693
It is disclosed in Japanese Patent Publication No.

[0009] Detecting an abnormality of a device, apparatus, or equipment that generates a sound wave during operation by acoustic analysis is disclosed as an abnormality detection method by acoustic analysis, for example, in Japanese Patent Laid-Open No. 3-10123. There is.

Further, detecting an abnormality in a wide range of equipment groups with a small number of sensors without touching the equipment is known as a method for diagnosing an abnormality in an equipment or equipment group, for example, as disclosed in Japanese Patent Laid-Open No. Hei 5 (1993) -58.
It is disclosed in Japanese Patent No. 187911.

[0011]

However, the above-described method for detecting the biting of the rolling mill, the method for detecting the condition inside the blast furnace,
In addition, in the conventional example such as the converter acoustic measurement method, there is a problem that the measurement target is fixed and two or more facilities cannot be monitored or diagnosed.

In the above-described method for diagnosing abnormalities in equipment or equipment groups, sounds that are not true abnormal sounds, such as a person's talking voice or an impact sound from an object, may occur even if the sound pressure level is high. Even though it must be removed,
There was no such function.

On the other hand, the anomaly detection method by acoustic analysis described above does not have a problem that a sound that is not a true anomalous sound must be removed, and the source of the anomalous sound seems to be successfully identified by the sound collecting microphone. Looks like. However, even if a super directional microphone is used as the sound collection microphone,
Only a gain of about -2.5 dB is obtained in the 0 ° direction.
However, in actual wide area surveillance, it is necessary to install a sound collection microphone at a distance of several meters, so in the direction of 30 °,
As shown in FIG. 7, 2 corresponds to 5 (m) × tan 30 °.
The source of abnormal sound can only be identified at intervals of 8 meters.

That is, FIG. 7 is a schematic configuration explanatory view for explaining a conventional abnormality detection method for equipment or the like, which uses a super-directional microphone as a sound collection microphone and detects an abnormality in equipment to be monitored by acoustic analysis. An example is shown in which one sound collection microphone M detects an abnormal sound generated from a plurality of monitored equipments m1, m2, m3, m4, m5, m6. In this figure, the sound collecting microphone M and the equipment to be monitored m1, m2,
The distance from m3, m4, m5, m6 is 5 meters,
5 (m) × tan 3 in the direction of 30 ° from the sound collecting microphone M
The source of the abnormal sound can only be identified at an interval of 2.8 meters, which corresponds to 0 °.

Since the distance of 2.8 meters is considerably larger than the distance between the facilities (usually about 1 meter),
It is practically impossible to identify the source of the abnormal sound in the monitored equipments m1, m2, m3, m4, m5, m6 by the sound collecting microphone M.

The present invention has been made in order to solve the above-mentioned conventional problems, and eliminates false abnormal sounds such as human voices and sounds of hitting objects, and detects only true abnormal sounds of equipment. It is an object of the present invention to provide a method of monitoring equipment by acoustic analysis, which is capable of being performed.

[0017]

SUMMARY OF THE INVENTION The present invention is a method for monitoring equipment by acoustic analysis, which uses a microphone and an acoustic analyzer to monitor equipment or equipment groups scattered over a wide area. Only when the sound pressure level of the specific frequency exceeds a predetermined value and the sound pressure level exceeds the predetermined value for a predetermined time, it is determined that the monitored equipment is abnormal. By doing so, the above problems are solved.

[0018]

In the method of monitoring equipment by the acoustic method according to the present invention, the sound pressure level of the whole or the sound pressure level of the specific frequency exceeds a predetermined value, and the sound pressure level exceeds the predetermined value. Only when the state continues for a predetermined time, it is determined that the monitored equipment is abnormal. For this reason, all false abnormal sounds with a high sound pressure level, such as human voice, impact sound when an object hits, air valve exhaust sound, drain exhaust sound, etc., are excluded, and it is announced that the monitored equipment is abnormal. Only true abnormal sounds can be detected.

Next, the operation of the present invention will be described in more detail with reference to the flow chart of FIG. In this flowchart, when the analysis of the acoustic signal transmitted from the sound collecting microphone (not shown) is started in the acoustic analyzer (not shown) (step 100), the value of the flag f is initialized to zero. Then, the value of the sampling number n after the condition for determining whether or not to issue the primary alarm (so-called primary alarm determination condition) is satisfied is initialized to zero (step 104). . Next, it is determined whether or not the measured value SPL of the sound pressure level is equal to or higher than the threshold value SPL0 of the sound pressure level (step 106).

When it is determined that the measured value SPL of the sound pressure level is less than the threshold value SPL0 of the sound pressure level, the value of the flag f is reset to zero (step 10
8) After resetting the value of the sampling number n to zero (step 104), it is again determined at the next timing whether or not the sound pressure level measurement value SPL becomes equal to or higher than the sound pressure level threshold value SPL0. (Step 106). This series of functions, namely the functions shown in the loop of FIG.
The sound pressure level measurement value SPL is repeatedly executed until the sound pressure level threshold value SPL0 or more.

If the sound pressure level measurement value SPL is judged to be equal to or higher than the sound pressure level threshold value SPL0, the primary alarm judgment is made (step 110), and then the flag f
Is determined to be zero (step 11
2).

When it is determined that the value of the flag f is not zero, it is again determined at the next timing whether the sound pressure level measurement value SPL is equal to or higher than the sound pressure level threshold value SPL0. Then (step 106), the above-mentioned function is repeatedly executed. This series of functions, that is, the functions shown in the loop of FIG. 1, is repeatedly executed until the sound pressure level measurement value SPL becomes less than the threshold value SPL0 and the value of the flag f is reset to zero.

When it is determined in step 112 that the value of the flag f is zero, 1 is added to the value of the sampling number n after the above primary alarm determination condition is satisfied (step 114). After that, this sampling number n
Is determined to be equal to the number N of times for the secondary alarm determination (step 116).

If it is determined that the sampling number n is not the same as the secondary alarm determination number N, then
It is determined whether or not the sound pressure level measurement value SPL is equal to or higher than the sound pressure level threshold value SPL0 (step 106), and the above-described function is repeatedly executed. This series of functions, that is, the functions shown in the loop of FIG.
Is repeatedly executed until it is determined that the number is equal to the number N of times for the secondary alarm determination.

When it is determined that the sampling number n is the same as the number N for the secondary alarm determination, the secondary alarm is output (step 118), and then the flag f is set to 1.
Is set (step 120), it is again determined whether the sound pressure level measurement value SPL is equal to or higher than the sound pressure level threshold value SPL0 (step 106), and the above-described function is repeatedly executed. When the flag f is set to 1, the sampling number n is not counted up,
Secondary alarms are prevented from being issued twice.

[0026]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIGS. 2 and 3 are graphs showing the results of actual measurement of the sound pressure level using this embodiment. In each figure, the horizontal axis shows the frequency in 1/3 octave display and the vertical axis shows the measurement. The generated sound pressure level SPL is shown. Also, FIG.
3A is a time chart when the sound pressure level is actually measured using the present embodiment. Hereinafter, a case where the sound pressure level is actually measured using this embodiment will be described in detail with reference to FIGS. 2 to 4 and FIG.

In the present embodiment, the analysis of the acoustic signal sent from the sound collecting microphone (not shown) is performed by the fast Fourier transform (FFT) in the acoustic analyzer (not shown) or the like. Therefore, the calculation in the signal analysis takes a relatively long time, and the sampling period t in FIG.
It will be about 1 second. Further, in FIG. 4, i is 1, 2,
The time Ti when changing to 3, 4, ... Shows each sampling position after the actual measurement is started.

From time T1 to time T4, the sound pressure level measured value SPL is the sound pressure level threshold S as shown in FIG.
It is less than PL0. Therefore, the function of the loop is repeated in the flowchart described in detail with reference to FIG.

At time T5, the state shown in FIG. 3, that is, the sound pressure level measured value SPL is equal to or higher than the sound pressure level threshold value SPL0, so that the primary alarm condition is satisfied. Therefore, while the value of the flag f remains zero, 1 is added to the value of the sampling number n, and then a secondary alarm determination is made, and the function of the loop in the flowchart of FIG. 1 is repeatedly executed. The function of this loop is repeated until time T8 when the abnormal sound continues to be generated.

However, during the period from T5 to T8, the value 4 of the sampling number n is 4 and the value of the number N of secondary alarm determinations, for example, 6
Since it is smaller, the secondary alarm output is not performed and the loop function in the flowchart of FIG. 1 is repeatedly executed. The number N for the secondary alarm determination is the number of times corresponding to the duration after the primary alarm determination when the sound pressure level measurement value SPL is a value equal to or higher than the sound pressure level threshold value SPL0, and is set arbitrarily. Value. For example, in this embodiment, N = 6 is set.

From time T9 to time T14, no abnormal sound is generated. Therefore, the loop function in the flowchart of FIG. 1 is executed, and time T15 when the abnormal sound is generated again.
Then, the function of the above loop is entered again. Also, time T20
At this point, the value of the number of times of sampling n and the value of the number of times of secondary alarm judgment N both become 6 and become the same value, and the condition of the secondary alarm is satisfied. Here, the secondary alarm is output for the first time, a flag for f = 1 is set, and the process returns to the loop function in the flowchart of FIG.

After time T21, the value of the sampling number n starts again from zero, but since the value of the flag f has already become 1, the function of the loop in the flowchart of FIG. 1 is always returned. That is, the function of adding 1 to the value of the sampling number n is not executed.

The flags of f = 0 and f = 1 are used for the following reason. That is, the time when the sound pressure level measurement value SPL becomes equal to or higher than the sound pressure level threshold value SPL0,
That is, when the time during which the primary alarm determination is established is twice or more the time corresponding to the number N of times for the secondary alarm determination, in other words, the time during which the primary alarm determination is established is This is to prevent two or more secondary alarm outputs from being output when the product of the value of the number of times n and the value of the sampling time t is twice or more.

By the way, the sound captured by the sound collecting microphone includes not only the true abnormal sound of the equipment but also the sound of a person or an impact when an object hits, and the exhaust sound of the air valve. Each of these sounds has a sound pressure level measurement value SPL of 10 to 20.
It is higher by dB. On the other hand, the true equipment abnormal sound is
It is a relatively continuous sound, and other sounds are sounds that last only for a short time (about several seconds). Therefore, by appropriately selecting the value of the number N of times for secondary alarm determination, it is possible to eliminate a false abnormal sound, which is not a true equipment abnormal sound.

The sound collecting microphone may be a microphone having directivity or a microphone having no directivity. Further, in the case of the present embodiment, the 1/3 octave band method is generally used for the analysis of the acoustic signal sent from the sound collecting microphone, but the 1/1 octave band method and the fast Fourier transform (FFT) are used. Analysis or frequency spectrum may be used.

Next, an example of using the present invention in an oil cellar arranged under a cold rolling mill and containing utilities such as a hydraulic system and a water system around the rolling mill will be described.
Briefly explained. 5 and 6 are graphs showing the measurement results in this example.

When the equipment group is operating normally, the overall sound pressure level is 88 with a frequency pattern as shown in FIG.
It became ~ 90 dB. Set the sampling period t to 1.0
When the value of the number N of times for secondary alarm determination is set to 8 seconds, that is, when the duration (t × N) value is set to about 30 seconds, a human voice, an impact sound when an object hits, an air valve It was possible to detect only the true abnormal sound, by eliminating all the exhaust sound of the drain and the discharge sound of the drain. Note that FIG. 6 is a graph when the sound generated when the bearing of the drainage pump is abnormal is measured.

In these embodiments, when the above duration is not recognized, that is, in the flow chart of FIG. 1, the sound pressure level measurement value SPL is the sound pressure level threshold value SPL0.
If the above is exceeded, the secondary alarm output will average 3 in 5 minutes.
Most of the alarm output was unrelated to the true alarm output.

The following can be seen from the results of such an embodiment. That is, in the case of the present embodiment, the sound pressure level of the whole or the sound pressure level of the specific frequency exceeds a predetermined value, and
Only when the duration of the sound pressure level exceeds a predetermined value, it is judged that the monitored equipment is abnormal.Therefore, acoustic analysis is used for equipment or equipment groups scattered over a wide area. It can be effectively monitored, and the results of the present invention prove to be sufficient.

Although the sound to be measured is an audible sound in this embodiment, it may be an ultrasonic wave of 20 kHz or higher.

[0042]

As described above in detail, in the equipment monitoring method of the present invention, the sound pressure level of the whole or the sound pressure level of a specific frequency exceeds a predetermined value, and the duration of the sound pressure level is predetermined. Only when the value exceeds the value, the equipment to be monitored is judged to be abnormal.

Therefore, it becomes possible to monitor the facilities scattered over a wide area. In addition, all false abnormal sounds such as human voices, impact sounds when hitting objects, air valve exhaust sounds, drain exhaust sounds, etc. are eliminated, and only true abnormal sounds that indicate that the monitored equipment is abnormal. Can be detected. Furthermore, it is no longer necessary to use a directional microphone as the sound collection microphone.

Therefore, according to the present invention, a facility monitoring method by acoustic analysis is realized which can effectively monitor a facility or a group of facilities scattered in a wide area by utilizing acoustic analysis.

[Brief description of drawings]

FIG. 1 is a flow chart for explaining the operation of the present invention.

FIG. 2 is a graph showing the result of actually measuring the sound pressure level when the equipment is normal in the example of the present invention.

[Fig. 3] Similarly, a graph showing the result of actually measuring the sound pressure level when the equipment is abnormal.

FIG. 4 is a time chart for explaining an example of the present invention.

FIG. 5 is a graph showing the results of measuring normal sound in an example in which the present invention was used in an oil cellar of a cold rolling mill.

[Fig. 6] Similarly, a graph showing the results of measuring abnormal noise of the drainage pump bearing.

FIG. 7 is a schematic configuration explanatory diagram illustrating a conventional abnormality detection method for equipment or the like, in which a superdirective microphone is used as a sound collection microphone, and abnormality of the equipment to be monitored is detected by acoustic analysis.

[Explanation of symbols]

 M: Sound collection microphone mi: Equipment to be monitored

Claims (1)

[Claims] 1. A method of monitoring a facility by acoustic analysis, which monitors a facility or a group of facilities scattered in a wide area by using a microphone and a sound analyzer, and a sound pressure level of a whole or a specific frequency is predetermined. By the acoustic analysis, it is determined that the equipment to be monitored is abnormal only when the sound pressure level exceeds the predetermined value and the sound pressure level exceeds the predetermined value for a predetermined time. Equipment monitoring method.