JP5321646B2 - Abnormality inspection method and abnormality inspection device - Google Patents

Description

  The present invention relates to an abnormality inspection method and an abnormality inspection apparatus for inspecting whether a rotating body is abnormal.

  Under various influences such as friction and stress load, an abnormality such as damage may occur in a rotating body such as a bearing that rotates. The rotating body in which the abnormality has occurred becomes unstable and performs an abnormal operation different from the normal operation. When the rotating body performs an abnormal operation, an abnormal sound that does not occur during normal operation is often generated. For this reason, if an abnormal sound can be detected from the rotating body, it is possible to inspect whether there is an abnormality such as damage to the rotating body.

  Patent Document 1 describes a conventional abnormality inspection apparatus using an AE (acoustic emission) sensor as means for detecting abnormal sound from a rotating body.

  The operation of the conventional abnormality inspection apparatus 100 described in Patent Document 1 will be described with reference to FIG. The AE sensor 102 attached to the bearing 101 outputs an AE signal generated by the rotating body 103 to the abnormality inspection unit 104. The abnormality inspection unit 104 performs frequency analysis on the input AE signal and calculates the occurrence frequency of the AE signal.

  When the rotator 103 is damaged, an AE signal is generated at a frequency having a correlation with the rotation speed of the rotator 103. By using this, the abnormality inspection means 104 determines whether or not the calculated occurrence frequency of the AE signal has a correlation with the rotational speed of the rotating body 103. Diagnose that there is. As described above, the conventional abnormality inspection apparatus 100 inspects whether the rotating body 103 is abnormal.

JP 2002-181038 A

  By the way, there is a manufacturing site where an inspector detects an abnormal sound from a rotating body to inspect the rotating body for an abnormality. Such an inspection by an inspector is called a sensory inspection, and an inspection by an AE sensor as described in Patent Document 1 is called a physical inspection.

  In recent manufacturing sites, automation has been promoted, and sensory testing is expected to be converted to physical testing. The converted physical inspection may be required to be able to be inspected with the same level of accuracy as the sensory inspection.

  However, in the sensory test, an abnormality of the rotator 103 is inspected based on human senses. Therefore, just because an AE signal correlated with the rotational speed of the rotator 103 is detected, the rotator 103 It is not always diagnosed that there is an abnormality. For this reason, when an AE signal having a correlation with the number of rotations of the rotating body 103 is detected, the physical inspection described in Patent Document 1 for diagnosing the rotating body 103 as having an abnormality has the same level of accuracy as the sensory inspection. There is a problem that inspection cannot be performed.

  Therefore, an object of the present invention is to solve the above-described problems and provide an abnormality inspection method and an abnormality inspection apparatus that can perform a physical inspection with the same level of accuracy as a sensory inspection.

In order to achieve the above object, the abnormality inspection method of the present invention acquires an AE signal from a rotating body, calculates an occurrence frequency of each of a plurality of frequency components in the acquired AE signal from the rotating body, select the frequency components of the AE signal appearing at preset frequency from the AE signals from the rotary member calculated frequencies, earthenware pots Chi 4 0 kHz following components of a temporal change in the frequency components of the selected AE signal The presence or absence of abnormality of the rotating body is inspected based only on the amplitude of the rotation.

The abnormality inspection apparatus of the present invention includes an AE acquisition unit that acquires an AE signal from a rotating body, a calculation unit that calculates the occurrence frequency of each of a plurality of frequency components in the acquired AE signal from the rotating body, selection means for selecting the frequency component of the AE signal appearing at preset frequency from the AE signals from the rotating body calculating the occurrence frequency, Chi sales of time variation of the frequency components of the selected AE signal 4 Inspection means for inspecting whether or not the rotating body is abnormal based only on the amplitude of a component of 0 kHz or less.

  As described above, according to the present invention, physical inspection can be performed with the same level of accuracy as sensory inspection.

Configuration diagram of an abnormality inspection apparatus according to an embodiment It is the figure which showed the AE signal acquired with the AE sensor from the rotating body, (a) The figure which shows the AE signal acquired from the rotating body which the inspector diagnosed as normal, (b) From the rotating body which the inspector diagnosed as abnormal The figure which shows the AE signal which is acquired The figure which showed the relationship between the frequency which AE signal generate | occur | produces, and the frequency The figure which showed the relation between the occurrence frequency and the frequency in the AE signal from the rotating body diagnosed as normal by the inspector The figure which showed the relationship between the frequency and the amplitude in the AE signal from the rotating body diagnosed as normal by the inspector The flowchart which shows operation | movement of the abnormality inspection apparatus of embodiment Schematic diagram showing a conventional abnormality inspection device

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a schematic diagram showing an abnormality inspection apparatus 1 according to the present embodiment and a rotating body 2 that is a test object.

  First, the rotating body 2 will be described using a bearing including a rolling element 2a, an inner ring 2b, and an outer ring 2c. A rotating shaft 3 is rotatably supported by the rotating body 2. If the rotating body 2 includes an abnormality, AE occurs at a specific frequency due to the abnormality. The abnormality inspection device 1 inspects the presence or absence of an abnormality in the rotating body 2 by detecting an AE signal (hereinafter referred to as an AE signal) generated from the rotating body 2.

  Next, the configuration of the abnormality inspection apparatus 1 will be described. The abnormality inspection apparatus 1 includes an AE sensor 4 that acquires an AE signal from the rotating body 2, an amplifier 5 that amplifies the acquired AE signal, and an A / D conversion that converts the input AE signal into a digital signal and outputs the digital signal. And an analysis device 7 for analyzing the AE signal, an inspection device 8 for inspecting whether the rotating body 2 is abnormal based on the analysis result, and a display 9 for displaying the inspection result. An AE sensor 4 as an AE acquisition unit of the abnormality inspection apparatus 1 is attached to the rotating body 2 with an epoxy adhesive. Further, the AE sensor 4 acquires the AE signal generated by the rotating body 2 as an AE signal represented by a voltage value.

  Next, details of the AE signal analysis processing by the analysis device 7 will be described. In the description, a small bearing having a diameter of about 10 mm is used as the rotating body 2.

  First, the function of the analysis device 7 will be described. Some AE signals acquired from the rotating body 2 are acquired even when the rotating body 2 is in a normal state. For this reason, it is necessary to select only an AE signal (AE signal having a specific frequency of occurrence) that occurs at a specific frequency as a result of the abnormality of the rotating body 2 from the acquired AE signals. Therefore, the analysis device 7 has a function as a calculation unit for calculating the occurrence frequency of the acquired AE signal, and further, among the AE signals for which the occurrence frequency is calculated, only the AE signal having the occurrence frequency due to the abnormality is included. It also has a function as selection means for selecting.

  Next, FIG. 2 (a) shows an AE signal acquired by the AE sensor 4 from the rotating body 2 diagnosed by the inspector as normal, and FIG. 2 (b) shows from the rotating body 2 diagnosed by the inspector as abnormal. The acquired AE signal is shown. 2A and 2B, the horizontal axis represents time (seconds), and the vertical axis represents voltage (V). As shown in FIGS. 2A and 2B, the intensity of the AE signal acquired by the AE sensor 4 varies depending on the acquisition time. Such an AE signal that changes with time is referred to as a time change signal. Moreover, as shown in FIGS. 2A and 2B, the appearances of both time-varying signals are very similar, and only the AE signal having a specific frequency that is caused by an abnormality is selected from the appearances of the waveforms. It is difficult to do. For this reason, generally used techniques such as envelope detection cannot be used.

  Therefore, in the present embodiment, only the AE signal resulting from the abnormality of the rotating body 2 in FIG. 1 is acquired in the following three steps.

  First, as a first step, a time change signal for each frequency is calculated from the AE signal acquired by the AE sensor 4 of FIG. Since the time change signal of the AE signal acquired by the AE sensor 4 includes AE signals of a plurality of frequencies, the time change for each frequency is obtained using the short-time frequency analysis for the acquired AE signals of the plurality of frequencies. Calculate the signal. Specifically, the acquired AE signal is cut out for each short section, time-frequency conversion is performed by Fourier transform for each cut section, and the sections after Fourier transform are combined along time. Then, a time change signal for each frequency is calculated. The first step is performed by the time change signal calculation unit 7a which is a time change signal calculation means of the analysis device 7. In the first step, short-time Fourier transform is performed as short-time frequency analysis. Incidentally, the short-time Fourier transform is a technique in which an AE signal is cut out using a window function having a certain size, and the cut-out AE signal is Fourier transformed.

Next, as a second step, the frequency of occurrence of the AE signal is calculated for each frequency from the calculated time change signal for each frequency. The occurrence frequency of the AE signal at each frequency can be calculated by time-frequency conversion by Fourier transform. By this second step, an AE occurrence frequency map as shown in FIG. 3 can be created. The horizontal axis of FIG. 3 represents the AE signal generation frequency (s ⁻¹ ), and the vertical axis represents the frequency (Hz) of the generated AE signal. For example, FIG. 3 shows that in the region A, an AE signal having a frequency of about 100 (kHz) is generated about 160 times per second, that is, at a frequency of about 160 (s ⁻¹ ). Indicates. In the region B, it is shown that an AE signal having a frequency of about 220 (kHz) is generated at a frequency of about 290 (s ⁻¹ ). The second step is performed by the occurrence frequency calculation unit 7b which is an occurrence frequency calculation means provided in the analysis device 7 of FIG. The unit of the frequency of the AE signal and the frequency of occurrence is substantially the same because any value is a value obtained by Fourier transform of the time signal.

  Subsequently, as a third step, an AE signal having an occurrence frequency set in advance as an occurrence frequency caused by an abnormality is selected from the AE signals for which the occurrence frequency has been calculated in the second step. Since the occurrence frequency of the AE signal acquired by the AE sensor 4 can be calculated through the first step and the second step, only the AE signal generated due to the abnormality can be selected in the third step. It is. Further, since the frequency of the AE signal generated due to the abnormality is determined by the rotating body 2, the value of the frequency of occurrence of the AE signal due to the abnormality can be set in advance. Here, the value of the natural frequency of the rotating body 2 is used as the value of the occurrence frequency set in advance. The third step is performed by the selection unit 7c that is a selection unit provided in the analysis device 7.

Here, the following four are given as specific examples of the natural frequency of the rotating body 2 in FIG. 1 and are shown in (Equation 1) to (Equation 4). As the first natural frequency, the shaft rotation frequency f _r (s ⁻¹ ) depending on the rotation frequency of the rotating shaft 3 is _{expressed by} (Equation ¹ ), and as the second, the rolling body rotation frequency depends on the rotation frequency of the rolling element 2a. f _b (s ⁻¹ ) is represented by (Equation 2), and third, the inner ring rolling element passing frequency f _i (s ⁻¹ ) depending on the frequency at which the rolling element 2 a passes the inner ring 2 b is represented by (Equation 3), Fourth, the outer ring rolling element passing frequency f _o (s ⁻¹ ) depending on the frequency at which the rolling element 2 a passes the outer ring 2 c is shown in (Equation 4).

In these (Equation 1) to (Equation 4), the rotational speed of the rotary shaft 3 is N (rpm), the number of rolling elements 2a is Z (pieces), the diameter of the rolling elements 2a is D _a (mm), pitch circle diameter is d _m (mm).

Here, when N = 3200 (rpm), Z = 8 (pieces), D _a = 2 (mm), and d _m = 9 (mm), f _r = 53.3 (s ⁻¹ ), f _b = 114.0 (s ⁻¹ ), f _i = 260.7 (s ⁻¹ ), f _o = 165.9 (s ⁻¹ ).

As shown in FIG. 3, AE signals are detected by the outer ring element passing frequency ^{_{f o (165.9 (s -1)}} ) from the rotary member 2 in FIG. 1. From this, it can be seen that the rotating body 2 has many AE signals generated when the rolling element 2a passes through the outer ring 2c. As described above, in the third step, an AE signal having an occurrence frequency corresponding to these natural frequencies is selected from the AE signals whose occurrence frequencies are calculated in the second step.

  Next, a process of inspecting whether or not the rotating body 2 is abnormal based on the selected AE signal of the occurrence frequency, that is, an abnormality inspection process will be described. In the abnormality inspection process, the rotor 2 is inspected for abnormality by performing threshold processing on the AE signal having the frequency of occurrence selected by the selection unit 7c in the third process. The threshold processing is processing for diagnosing that there is an abnormality in the rotating body 2 when a threshold is provided for the amplitude of the detected AE signal and an AE signal having an amplitude exceeding the threshold is detected. is there. In the present embodiment, the amplitude as the threshold is 20 (V).

  Note that the abnormality inspection step is performed by the inspection apparatus 8 of FIG. When there is no AE signal having the selected occurrence frequency, the threshold value processing is performed with the amplitude value of the detected AE signal set to 0.

  Subsequently, a process performed in the present embodiment will be described in order to obtain an inspection accuracy equivalent to the sensory inspection performed by the inspector.

FIG. 4 shows a graph obtained by performing up to the second step described above on the rotating body 2 of FIG. 1 diagnosed by the inspector as having no abnormality. The horizontal axis in FIG. 4 indicates the AE signal generation frequency (s ⁻¹ ), and the vertical axis indicates the AE signal frequency (Hz). Next, FIG. 5 shows a graph showing the relationship between the frequency of the AE signal and the amplitude obtained by integrating the graph of FIG. 4 in the horizontal axis direction. The horizontal axis in FIG. 5 indicates the frequency (Hz) of the AE signal, and the vertical axis indicates the amplitude (V) of the AE signal. At this time, if threshold processing is performed in the abnormality inspection process without performing special processing, many AE signals exceeding the threshold are detected. In this case, the inspection device 8 diagnoses that the rotating body 2 diagnosed by the inspector as having no abnormality is abnormal. That is, physical inspection cannot be performed with the same level of accuracy as sensory inspection.

  Thus, as a result of intensive studies, the inventors have performed an abnormality inspection process by performing threshold processing only for AE signals having a frequency of 40 (kHz) or less, and the level is equivalent to the accuracy of sensory inspection by an inspector. We found that physical inspection can be performed with high accuracy. For this reason, in the present embodiment, threshold processing is performed in the abnormality inspection process only for AE signals having a frequency of 40 (kHz) or less. In FIG. 5, since there is no AE signal of 40 (kHz) or more exceeding the threshold value, the inspection device 8 diagnoses the rotating body 2 from which this AE signal is acquired as having no abnormality. That is, the physical inspection by the inspection device 8 shows the same inspection result as the sensory inspection by the inspector.

  Next, the verification result of the effect of performing the threshold processing in the abnormality inspection process only for the AE signal of 40 (kHz) or less will be described. Here, for the sake of explanation, performing threshold processing in the abnormality inspection process only for AE signals of 40 (kHz) or less will be referred to as an approximation process. In addition, the case where the abnormality inspection process is performed for an AE signal having a frequency higher than 40 (kHz) is a case where threshold processing is not performed. 300 specimens were used for the verification, and the inspection results of these specimens when the approximation process was not performed and when the approximation process was performed were compared with the inspection results of the sensory test by the inspector. .

  First, verification results when the approximation process is not performed will be described. For 13 specimens out of 300 specimens, the inspection results obtained by the physical inspection when the approximation process was not performed and the inspection results obtained by the sensory inspection were different. More specifically, 13 specimens were diagnosed as having no abnormality in the sensory test and diagnosed as having an abnormality in the physical test when the approximation process was not performed.

  On the other hand, for all 300 specimens, the inspection results of the physical inspection and the inspection result of the sensory inspection coincided with each other when the approximation process was performed.

  From this verification result, it has become clear that the physical inspection of the test object can be performed with the accuracy equivalent to the accuracy of the sensory inspection by the inspector by performing the approximation process.

  In addition, it was predicted that only the AE signal of the frequency within the audible range should be focused on to bring the accuracy of the physical test close to the accuracy of the sensory test. The inventors have found that this has to be done. That is, among the AE signals having a frequency of 40 (kHz) or less, the frequencies of 20 (kHz) or more and 40 (kHz) or less are outside the human audible range, but the AE signal having a frequency outside the audible range is physically included. The inventors have found that by performing the inspection, a physical inspection equivalent to the accuracy of the sensory inspection can be performed.

  Finally, the operation of the abnormality inspection by the abnormality inspection apparatus 1 of FIG. 1 will be described using the flowchart of FIG.

  In step S <b> 1 of FIG. 6, a time change signal for each frequency is calculated from the AE signal acquired by the AE sensor 4. This step S1 is performed by the time change signal calculation unit 7a. In addition, the above-mentioned 1st process corresponds to this step S1.

  In step S2, the occurrence frequency of the AE signal is calculated for each frequency from the time change signal for each frequency calculated in step S1. This step S2 is performed by the occurrence frequency calculation unit 7b. In addition, the above-mentioned 2nd process corresponds to this step S2.

  In step S3, an AE signal having a preset occurrence frequency is selected from the AE signals for which the occurrence frequency has been calculated in step S2. This step S3 is performed by the selection unit 7c, and the preset occurrence frequency is stored in advance in the selection unit 7c according to the natural frequency of the rotating body 2. The above-described third step corresponds to this step S3.

  In step S4, threshold processing is performed only for the AE signal whose frequency is 40 (kHz) or less in the AE signal having the frequency of occurrence selected in step S3. The approximation process described above corresponds to this step S4.

  In step S5, it is determined whether the rotating body 2 is normal or abnormal based on the result of the threshold processing. If the rotating body 2 is normal, that is, if no signal exceeding the threshold value is detected in step S4, the process proceeds to step S6. On the other hand, if the rotating body 2 has an abnormality, that is, if a signal exceeding the threshold value is detected in step S4, the process proceeds to step S7. Steps S5 and S6 are performed by the inspection device 8.

  In step S6, the display 9 displays that the rotating body 2 is normal.

  In step S7, a message indicating that the rotating body 2 is abnormal is displayed on the display 9.

  As described above, the abnormality inspection apparatus 1 inspects whether the rotating body 2 is abnormal by the steps S1 to S7.

  As described above, the abnormality inspection apparatus 1 can perform a physical inspection with the same level of accuracy as a sensory inspection by an inspector.

  The present invention can be applied to abnormality inspection of products having a rotating body such as a ventilation fan, an air purifier, and a motor.

DESCRIPTION OF SYMBOLS 1 Abnormality inspection apparatus 2 Rotating body 4 AE sensor 7 Analysis apparatus 7a Time change signal calculation part 7b Occurrence frequency calculation part 7c Selection part 8 Inspection apparatus

Claims (2)

Acquire AE signal from the rotating body,
Calculating the frequency of occurrence of each of a plurality of frequency components in the acquired AE signal from the rotating body;
Select the frequency component of the AE signal that appears at a preset occurrence frequency from the AE signal from the rotating body that has calculated the occurrence frequency,
Abnormal inspection method for inspecting the presence or absence of abnormality of the rotating body based only on amplitude Urn Chi 4 0 kHz following components of a temporal change in the frequency components of the selected AE signals. AE acquisition means for acquiring an AE signal from the rotating body;
Calculating means for calculating the frequency of occurrence of each of a plurality of frequency components in the acquired AE signal from the rotating body;
Selecting means for selecting a frequency component of the AE signal that appears at a preset occurrence frequency from among the AE signals from the rotating body for which the occurrence frequency is calculated;
And inspecting means for inspecting the presence or absence of abnormality of the rotating body based only on amplitude Urn Chi 4 0 kHz following components of a temporal change in the frequency components of the selected AE signals,
An abnormality inspection apparatus comprising:

Images