JPS5948437B2 - Abnormal signal detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an abnormality signal detection device for detecting an abnormality signal in an abnormality detection device for a machine or the like that performs periodic motion such as rotation or reciprocating motion.

Conventionally, in machines that perform periodic motion, sounds and vibrations are generated when abnormalities occur, and conversely, it is possible to detect abnormal conditions in machines etc. from signals such as sounds and vibrations. One thing is well known.

By the way, the most important point in this abnormality detection is to faithfully detect only the signal originating from the occurrence of the abnormality from other noise and unnecessary components. There are various methods for removing this noise and unnecessary components, but one of the most effective methods is to use a bandpass filter to detect only abnormal signals and block all other components. It is. In fact, the abnormal vibrations and abnormal sounds that are detected when an abnormality occurs are the natural vibrations of any part of the machine that is the target of abnormality detection, so if the natural vibration frequency that appears in the abnormal signal is known, In other words, this method can be said to be very effective. However, in reality, it is generally difficult to predict in advance the type of abnormality that will occur, the location where it will occur, and the natural vibration frequency determined by it. be. The purpose of this invention is to
It is an object of the present invention to provide a device for effectively detecting abnormal signals, which compensates for the above-mentioned drawbacks. This countermeasure is as follows.
Consists of points.

That is, at least the signal frequency components unrelated to the abnormal signal are detected and blocked, and signals below a certain level are removed as unnecessary signals. By the way, in general, machines that perform periodic motion generate some kind of vibration or noise during operation, regardless of whether there is an abnormality or not, but the signal components that account for most of the power of these signals These are the synchronized component and its harmonics. Regarding the relationship between the periodic motion that generates the periodic signal and the abnormal signal, it can be said that it is the periodic motion that induces the abnormal signal from the mechanism of generation of the abnormal signal. for example,
In the meshing of gears with scratched teeth, an impact force is generated due to a sudden change in the tooth profile when the scratched parts mesh, and this induces an abnormal signal. This abnormal signal is the natural vibration of the gear. In addition, in hydraulic pumps,
If galling occurs on the side wall, an abnormality signal is generated each time the vane slides over that area. The abnormal signal at this time is the natural vibration of the vane or side wall. For gears, the period is the meshing period of the teeth, and for hydraulic pumps, the period is the reciprocal of the product of the number of revolutions and the number of vanes, and this periodic motion induces an abnormal signal. Therefore, the periodic signal that occupies most of the power serves as a trigger for inducing an abnormal signal, but at least it is not an abnormal signal. From this point of view, unnecessary components can be largely removed by removing periodic signal components and their harmonics. Moreover, since the periodic signal component is determined from rotation or reciprocating motion that can be easily detected, this technique is easy to apply. The above method is a method of removing at least high-power unnecessary components unrelated to the abnormal signal in a frequency band considerably distant from the frequency band of the abnormal signal, and is highly effective. This method is second best to the method of extraction using a filter.

Therefore, some noise components remain in the detected signal.
Generally, this noise component always exists regardless of the occurrence of an abnormal signal, and its level is lower than that of the abnormal signal. Since the power of the abnormal signal generated by the occurrence of an abnormality has at least a positive correlation with the degree of abnormality, the degree of abnormality can be detected by measuring the power of the abnormal signal. Low-level noise components that are always present are integrated and are included as a large error, which must be removed. This noise removal can be achieved by adding a threshold circuit that can freely select a threshold value before integration. Next, an explanation will be given with reference to drawings showing examples. FIG. 1 is a block diagram of an abnormal signal detection device including the above-mentioned functions. Signals such as sounds and vibrations detected by a signal detector 1 are sent to a high-pass filter 2. The lower limit frequency of this high-pass filter 2 is determined by the motion period of a machine or the like detected by the period detector 3 and the order value given to the lower limit frequency setter 4. This order value is the order of the harmonic components below which they are blocked. and,
The signal that has passed through the high-pass filter 2 is subjected to envelope detection by an envelope detection circuit 5, and is further averaged an appropriate number of times by an averaging processing circuit 6 to remove noise. The output from this averaging processing circuit 6 is input to a threshold circuit 7, where low-level noise components are removed, and then integrated for a certain period of time by an integrating circuit 8 and output to a display 9. FIG. 2a shows an example in which gear meshing noise is processed by a high-frequency filter, and FIG. 2b shows the time axis of a expanded.

In the figure, the meshing noise is represented by S1, the meshing synchronization signal is represented by P, the gear shaft rotation synchronization signal is represented by R, and the meshing noise after passing through the high-pass filter is represented by C. The signal S fluctuates in the same period as the signal P,
It is a periodic signal with about 4 times higher harmonics or superimposed. b signal S
The high-frequency signal component seen in the signal is an abnormal signal originating from the scratch, and when the signal S is processed to remove the component synchronized with the meshing synchronization signal and its harmonic components, the signal C is obtained. It is confirmed that an abnormal signal synchronized with signal R clearly appears. FIG. 3 shows a signal after envelope detection has been performed in the processing block of FIG. The small signals that exist in the vicinity of are noise components.

By measuring the power of the abnormal signal, it is possible to estimate the degree of abnormality, but the estimation accuracy is
As already mentioned, it depends on the noise removal rate. Further, although the noise component existing near the O level in FIG. 3 is at a low level, it continues to exist and may cause a large error in the next integration step. In order to remove this noise component, a threshold level is set as shown in FIG. 3, and all signals below the level are removed. The threshold circuit 7 performs this function, and this processing can be expected to significantly improve the S/N ratio. As described above, according to the abnormal signal detection device of the present invention, abnormal signals from machines that perform periodic motion such as rotation or reciprocating motion can be effectively detected.

[Brief explanation of drawings]

Figure 1 is a block diagram of one embodiment of the abnormal signal detection device of the present invention, Figure 2a is a diagram of abnormal signal detection by a high-pass filter, Figure 2b is an enlarged view of a, and Figure 3 is an envelope. It is an abnormal signal diagram after detection. 1...Signal detector, 2...High-pass filter, 3...Period detector, 4...1. Lower limit frequency setter, 5... ...Envelope detection circuit, 6...
...Averaging processing circuit, T...Threshold circuit, 8.
...Integrator circuit, 9...Display device.

Claims (1)

[Claims] 1. A signal detector that detects signals such as sounds and vibrations of machines that move periodically, a high-pass filter that receives the output of the signal detector and passes high-frequency components, and detects the motion of the machines, etc. a period detector that detects the period; a lower limit frequency setter that outputs a setting signal for the lower limit frequency of the high-pass filter based on the output of the period detector; and an envelope detection circuit that performs envelope detection of the high-frequency component. an averaging processing circuit that averages the output of the envelope detection circuit; a threshold circuit that removes low-level noise components from the output of the averaging processing circuit; and the averaging processing circuit that passes through the threshold circuit. An abnormal signal detection device comprising: an integrating circuit that integrates the output of the integrating circuit; and a display that displays the output of the integrating circuit.