JPS6343708B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention detects the presence or absence of cracks in components of equipment, equipment, machinery, etc. and the degree of destruction from the AE signal (acoustic emission signal) generated from the cracks that occur in the components. The purpose is to provide a means for detecting damage to components at an early stage and preventing accidents due to destruction of components.

It is widely known that when an object deforms or breaks, it releases part of its energy in the form of sound, which is called stress wave emission or acoustic emission. (In the present invention, this will simply be
(referred to as AE).

Therefore, AE generated by plastic deformation (hereinafter treated as an AE signal as it is detected by a sensor) is a continuous wave, whereas AE generated by cracks etc. is a discontinuous sudden waveform. is also known. The present invention is directed to discontinuous sudden signals having a larger amplitude than the background noise constantly obtained from the sensor. Various methods have been proposed for diagnosing whether or not various components are at risk of being destroyed from such AE signals.
One part is actually being implemented. However, the well-known methods have problems in quantitatively grasping the risk of component destruction, and furthermore, if the component to be measured is an object that moves at high speed or rotates, diagnosis is difficult and detection is impossible. It was hot.

The present inventors have developed a method that can accurately detect the presence or absence of damage and the degree of damage to not only stationary or fixed machines and members but also moving and high-speed rotating machines and members, that is, the method of the present invention. The gist is that the signal obtained from the AE sensor attached to the object to be measured is amplified, then the signal is divided into two circuits with different attenuation effects depending on the frequency band, and then each A method for detecting damage to an object, characterized in that after comparing attenuation amounts of objects obtained from the circuit, the signals are classified and accumulated according to attenuation ratio, and damage is detected from the tendency of accumulation.

A method to further improve accuracy is to amplify the signal obtained from the AE sensor attached to the object to be measured, and then divide the signal into two types of circuits with different attenuation effects depending on the frequency band. Then, after comparing the attenuation amounts of the signals obtained from each circuit, the signals are classified and accumulated according to the attenuation ratio, and the AEs in the classification are classified according to the preset classification criteria.
A method for detecting damage to an object, characterized in that the frequency of occurrence for each level of the attenuation ratio is determined for the signal equivalent classification, the ratio of the frequency of occurrence is compared with a set threshold value, and damage is detected by determining the magnitude of the frequency. .

A further point to be added is that the method of converting the attenuation ratio into a code signal and transmitting and receiving the code signal via a wireless circuit can be incorporated into the above method as necessary, and its characteristics are as follows. The object of the present invention is to provide a means that can be applied not only to static loads but also to all members that are subjected to dynamic loads.

Now, in general, there are large power transmission parts such as the spindle shaft of a rolling mill, which are parts for which it is particularly important to know the damage using the AE signal. AE, which is normally an extremely weak signal, is generated (treated as a noise signal because it is a
The signal gets mixed up in the noise signal.
There is a technical difficulty in that extremely complicated and expensive equipment must be used to clearly separate and extract only the AE signal, and conventional methods are not suitable for detecting damage to small parts or parts with complicated structures. At present, there is no means available.

The present invention provides a means for solving all of the above-mentioned problems using relatively simple means, and will be explained in more detail below with reference to the drawings.

The AE signal has a sudden waveform as shown in FIGS. 1a to 1c, and can be counted pulse by pulse. In addition, its frequency characteristics extend over a fairly wide band (several MHz or more), as shown by solid lines 2 and 3 in Figure 2, and at the initial stage of crack generation (solid line 2),
In the final stage (solid line 3) when the crack progresses and the member breaks, the latter tends to have relatively larger components in the high frequency range. Since the number of AE signals generated varies depending on conditions such as the material of the component and the load applied to the component, it is difficult to detect the degree of damage to the component, or in other words, the degree of danger, just from the number of AE signals generated. It is.

In addition, as shown by dotted line 4 in Figure 2, the frequency characteristics of mechanical noise signals generally have main components in the low frequency range of several tens of kilohertz or less, and include relatively high frequency components of several hundred kilohertz. However, the ratio of high-frequency components is much smaller than that of low-frequency components. moreover
When comparing the absolute value strengths of the AE signal and the mechanical noise signal, the mechanical noise signal is often much larger. Therefore, even if an attempt is made to discriminate the AE signal using a means that utilizes the difference in frequency characteristics, for example c ₂ in Fig. 2, that is, a high-pass filter with a cutoff frequency indicated by the frequency in the approximate boundary region between the AE signal and the mechanical noise signal, c ₂ as shown in Figure 3 b
Since the mechanical noise signal 5b cut off by the AE signal 1a has a much larger output than the AE signal 1a or has an output comparable to that of the AE signal 1a, it is difficult to discriminate only the AE signal.

The present inventors have succeeded in solving the above-mentioned technical problem by focusing on the signal attenuation ratio. FIG. 4 is a schematic block diagram of an embodiment of the apparatus according to the method of the present invention.

In FIG. 4, a signal obtained from an AE sensor 6 attached to an object to be measured (not shown) is amplified to a desired level by a preamplifier 7, and is amplified to a desired level by a high-pass filter 8a, _c2 having different cutoff frequencies, for example, _c1 . The water is diverted to a high-pass filter 8b having a high pass filter 8b. Said AE sensor 6
The sensor may be similar to the AE sensor used for ultrasonic damage, but it is desirable to have flat frequency characteristics over a wide band (1 to 2 MHz). Now, the cutoff frequencies c ₁ and c ₂ (in this embodiment, for example, c ₁ is set to 50 KHz and c ₂ is set to around 600 KHz) are c ₁
<c ₂ , high pass filters 8a, 8b
The output of is the signal 5 shown in Figure 3a and b above.
a, 5b, that is, the mechanical noise signal passing through the high-pass filters 8a, 8b has a low frequency range as its main component, and is therefore greatly attenuated here. On the other hand, AE signals with broadband components are only slightly attenuated.
Next, the outputs of the high-pass filters 8a and 8b pass through envelope detectors 9a and 9b and peak hold circuits 10a and 10b, respectively, to hold the peak value of the signal waveform, and then are input to the ratio measuring circuit 11.
The ratio measuring circuit 11 compares the output signals that have passed through the two circuits and measures the ratio between them. This ratio is referred to as a damping ratio in the present invention. That is, the output peak value of the peak hold circuit 10b/the output peak value of the peak hold circuit 10a = attenuation ratio In the present invention, the absolute value of the signal does not matter, but the ratio matters. Although this ratio can be converted into various signals such as voltage, pulse width, digital code, frequency, etc., a method for converting it into pulse width will be described here.

The ratio measuring circuit 11 can obtain a predetermined function by having a circuit configuration consisting of a variable resistor 23, an integrator 24, an adder 25, and a comparator 26 as shown in FIG. 5, for example.

That is, as shown in FIG. 6, when the horizontal axis represents time and the vertical axis represents input level, the peak hold circuit 1
Since the input levels 10a' and 10b' from 0a and 10b are 10a'>10b', when the input 10b' is integrated with an appropriately selected gain and time constant, the output increases linearly from zero in proportion to time. and the integrator 24
The output obtained by adding the output of and the input 10b' in the adder 25 is smaller than the input 10a' at the start of integration _t1 , but increases with time and finally reaches the same size as the input 10a'. ₂ generates output. The time from the start of integration to the output of the comparator 26 is proportional to the difference between inputs 10a' and 10b' and the integration time constant, and inversely proportional to the magnitude of input 10a', so if the integration time constant is constant, both inputs It is proportional to the ratio of 10a' and 10b'. That is, by using such a circuit, the ratio can be converted into a pulse width. This is transmitted via a modulation circuit 12, an FM transmitter 13, a transmitting antenna 14, a receiving antenna 15,
The signal is input to a pulse width measurement circuit 18 via an FM receiver 16 and a pulse width demodulation circuit 17.

Now, as mentioned above, when transmitting data, it is also free to convert it into a digital code, frequency, etc. and transmit it. Also, from the AE sensor 6 to the FM transmitter 13,
Since the equipment up to the transmitting antenna 14 can be extremely small and have a small volume, it is possible to easily obtain an AE signal by attaching it to a moving or rotating member such as a machine or equipment.

The pulse width measuring circuit 18 classifies the input pulses according to their width, and generates an output signal to one of the counting/memory circuits 19a to 19n in accordance with the pulse width. Counting/memory circuit 1
The one among 9a to 19n that receives this output increases its count value by one. This count value is accumulated until all desired measurements are completed. In this way, the count value of any one of the counting/memory circuits 19a to 19n increases by 1 every time one discontinuous sudden signal is generated. As described above, the pulse width of the signal is proportional to the attenuation ratio, so the classification result is that the sound generated in the object to be measured, that is, the AE signal and the mechanical noise signal, are classified by the attenuation ratio.

Therefore, for example, 10 counting/memory circuits 19a~
19j is provided, and 19a is to integrate when a pulse width corresponding to an attenuation ratio of 1.0 to 0.9 comes, and in the same way, 19b is to integrate a pulse width of 0.9 to 0.8, and so on. The frequency of occurrence divided by 0.1 from attenuation ratio 1.0 to 0.0 is obtained.

The contents of the counting/memory circuits 19a to 19j are converted into, for example, analog values and read out from the output circuit 20 at regular intervals independently of the integration operation, and the display device
When displayed on the CRT 21, the operator can know what kind of change is occurring in the object to be measured or whether a steady state continues. That is, as shown in FIG. 7, a histogram is displayed on the component device CRT. The horizontal axis in FIG. 7 is the attenuation ratio, and the vertical axis is the frequency of occurrence. Next, the reason why damage can be detected by the above means will be explained in detail.

It is not possible to use the method described above to determine for each signal that, for example, signals with an attenuation ratio in the range of 1.0 to 0.9 are AE signals, and signals other than that are noise. Have difficulty. It can simply be said that an attenuation ratio closer to 1.0 is more likely to be an AE signal, and an attenuation ratio closer to 0.0 is more likely to be noise. Therefore, the present inventors did not judge the risk of destruction of the object to be measured based on whether each signal is an AE signal, but based on statistical data of a large number of signals or statistical data. This invention employs a method of determining the degree of risk based on changes in trends, and this is a feature of the present invention.

As already mentioned, mechanical noise signals have a large attenuation ratio, for example, an attenuation ratio of 1.0 to 0.3 (expressed in dB).
0 dB to -10 dB), the smaller the value of the attenuation ratio, the more occurrences occur, as illustrated by the solid line 27 in Fig. 8, and the histogram increases monotonically upward to the right. This has been confirmed in research by researchers. On the other hand, since the value of the attenuation ratio of the AE signal is generally about 1.0 to 0.5, a histogram with a normal distribution as shown in the dotted line 28 in Figure 8 is obtained, and the central value is the final stage of a crack, etc. The higher the value, the further to the left on the graph (the larger the value of the attenuation ratio), that is, the histogram distribution becomes as shown by the solid line 29. From the histogram thus obtained, the evaluation index R is calculated using the following formula.

Evaluation index R = total frequency between attenuation ratios 0 and α/total frequency between attenuation ratios α and β The inventors have determined the evaluation index for α and β, which are appropriately set empirically based on the measurement results of a large number of equipment. It has been confirmed that harmful cracks and the like are found when R exceeds a certain threshold value L. α and β of the specific damping ratio are empirical values, and α>β, and according to the experience of the present inventors based on this example, the damping ratio is 0.9.
It was in the range of ~0.5.

In addition, the threshold L is insensitive to the shape and material of the equipment and the member being measured, the load conditions, etc., and if the characteristics of the AE sensor, the cutoff frequency of the high-pass filter, the attenuation ratio values α and β, etc. are constant, it is approximately The present inventors found that it is constant. As described above, if the evaluation index R is less than the threshold L, it can be determined that harmful cracks do not exist, or even if they exist, they are not progressing. The further away from the threshold L, the more dangerous the state can be determined. In the present invention, this is referred to as determining the size.

In the example, when detecting damage to the spindle shaft of a steel hot rolling mill, the evaluation index R is 0.6 to 0.8,
Very good detection results were obtained with a threshold value of L0.2 to 0.3, and we succeeded in preventing serious breakage accidents. Note that the evaluation index R and threshold L are not limited to these.

Also, as is clear from the above explanation, the display device
By simply reading the changes in the histogram displayed on the CRT, the operator can determine the presence or absence of damage to the object to be measured and the degree of damage progress.

Furthermore, it is also possible to connect the X-Y recorder 22 to the output circuit 20 in FIG. 4 to display and record a graph as shown in FIG. 8.

Now, as described above, in the present invention, the AE signal is divided into two circuits with different attenuation effects, each consisting of a high-pass filter, an envelope detector, and a peak hold circuit, and the attenuation amounts of the signals obtained from each circuit are compared. By classifying and accumulating the signals according to the attenuation ratio, it is possible to detect the presence or absence of damage to the object to be measured or the degree of progress of the damage from the accumulation tendency (for example, histogram distribution). To summarize the above, the classification criteria that have been set, i.e., the attenuation ratio is in the range of 1.0 to 0.5 (as explained in detail earlier, this is a range that can be easily determined by experience or actual measurement). Then, calculate the ratio (evaluation index R) of the frequency of occurrence for each high and low damping ratio (0 to α, α to β), compare the ratio with the set threshold value (empirical value), and determine the magnitude, that is, the difference. More accurate and precise detection becomes possible by determining the magnitude of and the change tendency of the difference.

Furthermore, in the embodiment shown in FIG. When an object moves or rotates, the wireless signal transmission and reception of the present invention has many advantages.

FIG. 9 is a schematic perspective view of an example in which damage to the spindle shaft 30 of a steel hot rolling mill was actually measured, and 6 is an AE sensor fixed to the joint 30 of the spindle shaft, and 31 is the same as shown in FIG. 32 is a battery as a power source, 33 and 34 are mounting bands for fixing them to the shaft, 35 is a conductive wire, 36 is a transmitting antenna, 37 is a ring-shaped receiving antenna supported by a support device (not shown), and 38 is a ring-shaped receiving antenna that is connected from the FM receiving section 16 to the display device CRT 21 and X in FIG.
-Y Shows the cabinet containing up to the recorder 22. The results of detection using the method shown in this example were extremely good, and the presence or absence of a crack and its degree of progression could be clearly detected.

The advantage of the present invention is that equipment can be diagnosed in a short time by attaching only one AE sensor to a member to be measured.Furthermore, visual inspection is not easily possible from the outside and disassembly takes a lot of effort. It is possible to discover whether cracks have occurred in a member, which requires time and effort, and it is also possible to determine whether or not cracks that have already been discovered are actually dangerous. Since this type of detection can be performed while the machine is actually operating, there is no need to stop the equipment, and dangerous cracks can be detected with high accuracy under actual load conditions. In this way, the economic benefits of the present invention are significant because dangerous damage can be detected before the equipment component breaks.

[Brief explanation of drawings]

Figure 1 is a schematic explanatory diagram of the AE signal waveform, Figure 2 is the AE signal waveform diagram.
Graphs of frequency distribution of signals and mechanical noise signals,
Figure 3 is a schematic diagram of the waveform obtained when the AE signal and mechanical noise signal are passed through a high-pass filter.
FIG. 4 is a schematic block diagram of an embodiment of the apparatus according to the method of the present invention, FIG. 5 is a schematic block diagram explaining the ratio measuring circuit according to the present invention, FIG. 6 is a graph explaining the output principle, and FIG. The figure shows a histogram showing the classification accumulation by attenuation ratio obtained by the present invention.
Figure 8 is a graph showing the attenuation ratio and the frequency of occurrence.
FIG. 9 is a schematic perspective view of an embodiment of the method of the present invention. 1a to 1c...AE signal waveform, 5a, 5b...
Mechanical noise signal waveform, 6...AE sensor, 7...
Preamplifier, 8a, 8b...High pass filter, 9a-9b...Envelope detector, 10a, 10
b...Peak hold circuit, 11...Ratio measurement circuit, 12...Modulation circuit, 13...FM oscillator, 1
4...Transmission antenna, 15...Reception antenna, 1
6...FM receiver, 17...pulse width demodulation circuit,
18... Pulse width side constant circuit, 19a to 19n...
Counting/memory circuit, 20...Output circuit, 21...Display device CRT, 22...X-Y recorder, 24
... Integrator, 25 ... Adder, 26 ... Comparator,
30...Joint part, 32...Battery, 31...Box (with built-in transmitter), 33, 34...Mounting band, 36
...Transmission antenna, 37 ...Reception antenna, 38
... Cabinet.

Claims (1)

[Claims] 1. After amplifying the signal obtained from the AE sensor attached to the object to be measured, the signal is divided into two circuits each having a different attenuation effect depending on the frequency band, and then Compare the amount of attenuation for the signals obtained from the circuits, select one of the multiple counting circuits depending on the magnitude of the attenuation ratio, count each time a signal occurs, and count the signals in multiple counting circuits depending on the number of signals. A method for detecting damage to an object, characterized by detecting damage from a trend in numerical values. 2 After amplifying the signal obtained from the AE sensor attached to the object to be measured, the signal is divided into two connected circuits with different attenuation effects depending on the frequency band, and then the signal obtained from each circuit is Compare the amount of attenuation, select one of the multiple counting circuits depending on the magnitude of the attenuation ratio, count each time a signal occurs, measure a large number of signals, and then count the signals in a specific part of the multiple counting circuits. Damage is detected by comparing the ratio of the number of signals detected and the number counted in a separate specific part with a predetermined damage discrimination threshold and determining the size of the number. A method for detecting damage to an object, characterized by: 3. The method according to claim 1 or 2, wherein the attenuation ratio is converted into a code signal and the code signal is transmitted and received via a wireless circuit.