JP2022082474A - Abnormality diagnosis device and abnormality diagnosis method of vibration machine - Google Patents

Abstract

To improve accuracy and convenience of diagnosis in an abnormality diagnosis device and an abnormality diagnosis method of a vibration machine.SOLUTION: An abnormality diagnosis device 10 of a motor 1 includes: a storage part 11 which stores the reference data obtained by performing Fourier transformation on the vibration data of the motor 1; a transformation part 12b which perform the Fourier transformation on an acquisition part 12a for acquiring the vibration data of the motor 1 over time and the vibration data acquired by the acquisition part 12a to be used as the diagnostic data; and a diagnosis part 12c which diagnoses abnormality when the diagnostic data differs from the reference data by a predetermined amount or more by comparing a spectral value of the reference data with a spectral value of the diagnostic data over time.SELECTED DRAWING: Figure 1

Description

The present invention relates to an abnormality diagnosing device for a vibrating machine and an abnormality diagnosing method.

In vibration machines such as rotary machines, abnormality diagnosis using vibration information is carried out. Patent Document 1 discloses a method for diagnosing a rotating machine using such vibration information. In the diagnostic method of Patent Document 1, a state evaluation index is calculated by a principal component analysis method from a plurality of dimensional vibration parameters calculated from vibration information of a rotating machine, and the quality of the rotating machine is judged based on the state evaluation index. There is.

Japanese Unexamined Patent Publication No. 2008-58191

With a diagnostic method as in Patent Document 1, it is difficult to capture changes over time in a rotating machine. In addition, it is necessary to prepare various dimensional vibration parameters, which is complicated as a diagnostic method. Further, when the vibration information is affected by the carrier frequency of the inverter or the like, it is necessary to remove the influence. Overall, there is room for improvement in the accuracy and convenience of abnormality diagnosis.

An object of the present invention is to improve the accuracy and convenience of diagnosis in an abnormality diagnosis device and an abnormality diagnosis method for a vibrating machine.

The first aspect of the present invention is
A storage unit that stores reference data obtained by Fourier transforming the vibration data of a vibrating machine,
An acquisition unit that acquires vibration data of the vibration machine over time,
A conversion unit that Fourier transforms the vibration data acquired by the acquisition unit into diagnostic data,
An abnormality diagnostic device for a vibrating machine, comprising a diagnostic unit that compares the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more. I will provide a.

According to this configuration, diagnostic data is generated from vibration data that changes with time of the vibration machine to be diagnosed, and abnormality diagnosis is performed by comparing with the reference data of the vibration machine to be diagnosed. Therefore, since the data to be compared are from the same vibration machine, it is possible to capture changes over time in the vibration data, and it is possible to perform highly accurate abnormality diagnosis. For example, in a vibration machine using an inverter, the influence of the carrier frequency of the inverter is similarly added to both the past vibration data and the current vibration data. Therefore, by taking these differences, it is possible to perform an abnormality diagnosis without being affected by the carrier frequency of the inverter, which is convenient. Further, since the vibration data is acquired over time and used as diagnostic data, it is possible to detect an abnormality when the change of the vibration data with time becomes large. The vibrating machine here broadly refers to a machine that can generate vibration, and includes, for example, a rotating machine, a reciprocating machine, and a pipe through which water flows. Further, the vibration data may be vibration acceleration, vibration velocity, or the like.

The diagnostic unit may perform the comparison with an average value, an effective value, a minimum value, or a maximum value for each predetermined frequency band.

According to this configuration, abnormality diagnosis can be performed for each predetermined frequency band. Therefore, the cause of the abnormality can be identified according to the frequency band. The predetermined frequency band may be appropriately set according to the type of the vibrating machine and the frequency range of the acquired vibration data. The effective value is a quantity representing a time-averaged magnitude (intensity), and is the square root of the average of squared values.

The diagnostic unit may compare the peak value of the reference data with the peak value of the diagnostic data.

According to this configuration, abnormality diagnosis can be easily performed. In addition, large vibrations that take peak values often lead to abnormalities, and the accuracy of abnormality diagnosis can be improved by comparing the peak values. The comparison of the peak values here may be performed at a plurality of points in order from the largest value.

The diagnostic unit may evaluate the degree of coincidence between the waveform image of the reference data and the waveform image of the diagnostic data, and may diagnose an abnormality when the degree of coincidence is equal to or less than a predetermined value.

According to this configuration, the entire waveform can be compared as an image, so that the diagnostic accuracy can be improved. For example, an abnormality may be diagnosed when the degree of agreement is 50% or less.

The storage unit stores a plurality of reference data and stores a plurality of reference data.
When the diagnostic unit compares each of the plurality of reference data with the diagnostic data and obtains a plurality of comparison results, and even one of the plurality of comparison results is not abnormal. May be diagnosed as normal as a whole.

According to this configuration, it is possible to suppress the misdiagnosis of the change in vibration due to the change in the carrier frequency of the inverter or the change in the operation mode in the normal operation of the vibrating machine as an abnormal state.

The second aspect of the present invention is
Prepare the reference data obtained by Fourier transforming the vibration data of the vibrating machine.
The vibration data of the vibration machine is acquired over time, and the vibration data is acquired over time.
The vibration data acquired over time is Fourier transformed to obtain diagnostic data.
Provided is an abnormality diagnosis method for a vibrating machine, which comprises comparing the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnosing an abnormality when the diagnostic data differs from the reference data by a predetermined value or more. do.

According to the present invention, the accuracy and convenience of abnormality diagnosis can be improved in the abnormality diagnosis device and the abnormality diagnosis method of a vibrating machine.

Schematic block diagram of the abnormality diagnosis device for vibration machines. The control block diagram of the abnormality diagnosis apparatus which concerns on 1st Embodiment of this invention. The flowchart which shows the abnormality diagnosis method of 1st Embodiment. A graph showing the waveforms of reference data and diagnostic data. The bar graph corresponding to the reference data D0 of FIG. The bar graph corresponding to the diagnostic data D4 of FIG. A line graph showing reference data in the first modification. A line graph showing diagnostic data in the first modification. A line graph showing the difference between the diagnostic data and the reference data in the second modification. The flowchart which shows the abnormality diagnosis method of 2nd Embodiment.

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

(First Embodiment)
With reference to FIG. 1, the abnormality diagnosis device 10 of the vibration machine 1 of the first embodiment acquires the vibration data of the vibration machine 1 and diagnoses the abnormality from the vibration data. The vibration machine 1 broadly refers to a machine that can generate vibration, and includes, for example, a rotating machine, a reciprocating machine, and a pipe through which a liquid flows. Hereinafter, the motor 1 of the fan 2 used in an air conditioner such as an air conditioner will be described as an example of the vibration machine 1.

The fan 2 used in the air conditioner is driven by a motor 1 and blows temperature-controlled air into the room. The rotation speed of the motor 1 is adjusted by the inverter 3. The inverter 3 is controlled by the control device 4.

A vibration sensor 5 is attached to the motor 1. The vibration sensor 5 measures the vibration data of the motor 1 and transmits it to the abnormality diagnosis device 10. The vibration data can be vibration acceleration, vibration velocity, and the like. Here, vibration acceleration will be described as an example of vibration data. The vibration sensor 5 measures the vibration acceleration of the motor 1 at predetermined intervals (for example, once a day) and transmits it to the abnormality diagnosis device 10. In this embodiment, the vibration sensor 5 measures the vibration acceleration in the frequency range of 1 to 10000 Hz. Alternatively, the vibration sensor 5 may measure vibration data in any frequency range. For example, the vibration sensor 5 may acquire vibration data up to a frequency higher than 10000 Hz.

With reference to FIG. 2, the abnormality diagnosis device 10 is composed of hardware such as a CPU (Central Processing Unit), a RAM (Random Access Memory), and a ROM (Read Only Memory), and software mounted therein. ing. The abnormality diagnosis device 10 may be composed of an information processing device such as a desktop personal computer, a laptop computer, a workstation, or a tablet terminal, for example.

The abnormality diagnosis device 10 has a storage unit 11, a processing unit 12, and a display unit 13. The processing unit 12 includes an acquisition unit 12a, a conversion unit 12b, and a diagnosis unit 12c as a functional configuration. These are realized by the cooperation between the processor, which is a hardware resource, and the program, which is software.

The storage unit 11 stores reference data obtained by Fourier transforming the past vibration data of the motor 1. The past vibration data is the one when the normal operation state is maintained. Further, the storage unit 11 also records a program operated by the processing unit 12. The storage unit 11 may be a portable recording medium, or may be integrally configured with the processing unit 12.

The acquisition unit 12a receives vibration data from the vibration sensor 5 over time. The mode of exchanging data between the acquisition unit 12a and the vibration sensor 5 may be wired or wireless.

The conversion unit 12b Fourier transforms the vibration data acquired by the acquisition unit 12a into diagnostic data. Here, a fast Fourier transform (FFT) is used as an example of the Fourier transform. By this Fourier transform, the vibration data can be frequency-analyzed. Further, the conversion in the conversion unit 12b is executed every time the vibration data is acquired by the acquisition unit 12a. Therefore, it is possible to confirm the change over time in the frequency analysis result.

The diagnostic unit 12c compares the spectral value of the reference data with the spectral value of the diagnostic data over time, and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more. The specific method of comparison will be described later.

The display unit 13 is a part that displays the diagnosis result by the processing unit 12, and is composed of, for example, a liquid crystal display, an organic EL display, a plasma display, or the like. In the present embodiment, the display unit 13 displays the diagnosis result when the diagnosis unit 12c gives a diagnosis result that there is an abnormality. However, the display unit 13 may display the diagnosis result even when the diagnosis unit 12c gives a diagnosis result that there is no abnormality.

The abnormality diagnosis method executed by the abnormality diagnosis apparatus 10 of the present embodiment will be described with reference to FIG.

When the abnormality diagnosis method of the present embodiment is started (step S3-1), vibration data is acquired by the acquisition unit 12a (step S3-2). Next, the vibration data acquired by the acquisition unit 12a is Fourier transformed by the conversion unit 12b to obtain diagnostic data (step S3-3). Then, the diagnostic unit 12c compares the spectral value of the reference data stored in the storage unit 11 with the spectral value of the diagnostic data, and diagnoses as abnormal when the diagnostic data differs from the reference data by a predetermined value or more (step S3-4). ). If it is not diagnosed that there is an abnormality (N: step S3-4), the process returns to the process of acquiring vibration data (step S3-2), and the subsequent processes are repeated. When an abnormality is diagnosed by the diagnosis unit 12c (Y: step S3-4), the display unit 13 displays that there is an abnormality as a diagnosis result (step S3-5), and the series of processes is terminated (step). S3-6).

With reference to FIG. 4, the data to be compared in this embodiment will be illustrated. FIG. 4 is a three-dimensional graph having three axes (spectral values, frequencies, and dates). In this graph, the reference data and the diagnostic data are shown in 1 Hz increments, respectively.

The waveform data D0 located on the leftmost side of FIG. 4 is reference data and is stored in the storage unit 11 as normal data. The second and subsequent waveform data D1 to D9 in FIG. 4 are diagnostic data, and the vibration data acquired at intervals of once every two weeks (for example, the acceleration data of the vibration of the motor 1) are Fourier transformed. ..

In the graph of FIG. 4, the diagnostic data D4 on May 12, 2019 is superior to the reference data D0 in the vicinity of F1 and F2 (Hz) (see the part A in the graph). Specifically, no such predominant spectrum was identified in the diagnostic data D1 to D3 prior to May 12, 2019, and near the same frequency in the diagnostic data D5 to D9 after May 12, 2019. You can see the excellent spectrum with.

FIG. 5 shows a bar graph corresponding to the reference data D0 of FIG. FIG. 6 shows a bar graph corresponding to the diagnostic data D4 of FIG. In FIGS. 5 and 6, the horizontal axis indicates the frequency. The frequency is shown by dividing 1 to 10000 Hz by a predetermined division width. The vertical axis shows the average value of the spectrum in each divided frequency range. Regarding the predetermined division width, 1 to 10000 Hz may be equally divided, or the division width may be arbitrarily changed.

Comparing FIGS. 5 and 6, in the frequency range near F1 (Hz), which is the 14th frequency range, the average value of the spectra of the diagnostic data D4 is significantly larger than the average value of the spectra of the reference data D0. Can be confirmed. Similarly, in the frequency range near F2 (Hz), which is the 16th frequency range, it can be confirmed that the average value of the spectrum of the diagnostic data D4 is significantly larger than the average value of the spectrum of the reference data D0.

In the present embodiment, the diagnostic unit 12c has the average value of the spectra of any specific frequency range of the diagnostic data D1 to D9 larger than the average value of the spectra of the same frequency range of the reference data, for example, by 30% or more. If it is diagnosed as abnormal. In the examples of FIGS. 5 and 6, in the frequency range near F1 (Hz) and F2 (Hz), the average value of the spectrum of the diagnostic data D4 is more than twice as large as the average value of the spectrum of the reference data D0, respectively. ing. Therefore, the diagnosis unit 12c diagnoses the abnormality. The threshold value for comparison for diagnosing an abnormality may be other than 30%, and may be arbitrarily set according to the type of the vibrating machine to be diagnosed.

According to the present embodiment, diagnostic data is generated from vibration data of the motor 1 that changes with time, and abnormality diagnosis is performed by comparing with the reference data of the same motor 1 acquired in the past. Therefore, since the data to be compared is from the same motor 1, it is possible to capture the change over time in the vibration data, and it is possible to perform highly accurate abnormality diagnosis. Further, in the present embodiment, since the rotation speed of the motor 1 is controlled by the inverter 3, noise can be added to the vibration data at the carrier frequency of the inverter 3. However, the influence of the carrier frequency of the inverter 3 is similarly added to both the past vibration data and the current vibration data. Therefore, by taking these differences, the influence of the noise of the carrier frequency of the inverter 3 can be suppressed and the abnormality diagnosis can be easily performed. Further, since the vibration data is acquired over time and used as diagnostic data, it is possible to detect an abnormality when the change of the vibration data with time becomes large.

Further, according to the present embodiment, abnormality diagnosis can be performed for each predetermined frequency band. Therefore, the cause of the abnormality can be identified according to the frequency band. For example, when an abnormality is diagnosed in a low frequency region (for example, 1000 Hz or less), eccentricity (unbalance) or eccentricity (misalignment) of the motor 1 and rattling of the installation of the motor 1 are assumed. In addition, when the target is a pipe through which a liquid flows as a vibration machine, an excessive flow rate in the pipe is assumed. Further, when an abnormality is diagnosed in a high frequency region (for example, 5000 Hz or higher), it is assumed that the bearing portion supporting the shaft member that mechanically connects the motor 1 and the fan 2 is damaged or the grease is cut off.

In this embodiment, the mean value of the spectrum in a specific frequency range is confirmed and the reference data and the diagnostic data are compared, but the value to be compared is the effective value, the minimum value, or the maximum value of the spectrum in the specific frequency range. It may be a value or the like.

(First modification)
With reference to FIGS. 7 and 8, the peak value may be used as a method for comparing the diagnostic data and the reference data by the diagnostic unit 12c.

In the first modification, the diagnostic unit 12c compares the peak value of the reference data and the peak value of the diagnostic data in descending order, and diagnoses as abnormal when they differ by a predetermined value or more. Hereinafter, specific examples will be described.

FIG. 7 shows reference data and FIG. 8 shows diagnostic data. The horizontal axis of FIGS. 7 and 8 shows the frequency, and the vertical axis shows the spectrum.

In the graph of FIG. 7, the largest first peak is indicated by reference numeral P7-1 and the second largest peak is indicated by reference numeral P7-2. Similarly, in FIG. 8, the largest first peak is indicated by reference numeral P8-1 and the second largest peak is indicated by reference numeral P8-2.

In the examples of FIGS. 7 and 8, the first peak value of about 5.1 (P8-1) of the diagnostic data is about 13% larger than the first peak value of about 4.5 (P7-1) of the reference data. Further, the second peak value of about 4.2 (P8-1) of the diagnostic data is about 17% larger than the second peak value of about 3.6 (P7-1) of the reference data.

In the diagnostic unit 12c, as described above, the peak values are compared in descending order. When the threshold value for diagnosing an abnormality is set to 30% as in the above embodiment, in the examples of FIGS. 7 and 8, the comparison result exceeding the 30% cannot be confirmed, so that the abnormality is not diagnosed. However, the number of peak values to be compared may be one or three or more, and the threshold value is not limited to 30%.

According to the first modification, since the peak values are simply compared in descending order regardless of the frequency range, the abnormality diagnosis can be easily performed. Specifically, as shown in the examples of FIGS. 7 and 8, the frequency of the first peak P7-1 of the reference data and the frequency of the first peak P8-1 of the diagnostic data are different from each other. Therefore, in such a case, the peak values of different frequencies are compared with each other. However, a specific frequency range may be set as in the above embodiment, and peak values in the specific frequency range may be compared with each other.

In addition, large vibrations that take peak values often lead to abnormalities, and the accuracy of abnormality diagnosis can be improved by extracting and comparing peak values.

(Second modification)
With reference to FIG. 9, as a method of comparing the diagnostic data and the reference data by the diagnostic unit 12c, the difference waveform between the diagnostic data and the reference data may be used.

In the second modification, the diagnostic unit 12c diagnoses an abnormality when the waveform (difference waveform) obtained by subtracting the spectral value of the reference data from the spectral value of the diagnostic data becomes a slope waveform. Hereinafter, specific examples will be described.

The horizontal axis of FIG. 9 shows the frequency, and the vertical axis shows the spectrum. FIG. 9 shows a difference value obtained by subtracting the waveform of the reference data from the waveform of the diagnostic data.

In the graph of FIG. 9, the slope waveform can be confirmed in the vicinity of the frequency of 10000 Hz (see the reference numeral P9 portion). Here, the slope waveform means a waveform in which the FFT spectrum rises in a slope shape from around 2000 Hz to a lower frequency side, or from around 9000 Hz to a higher frequency side than that.

In the graph of FIG. 9, the spectral value increases about 8 times from about 0.2 to about 1.6 as the frequency increases from 9001 to 10000 Hz. Therefore, since the slope waveform is confirmed in the frequency range, the diagnosis unit 12c diagnoses that there is an abnormality in the frequency range.

According to this modification, an abnormality is likely to occur in the slope waveform, and by detecting the occurrence of the waveform, it is possible to perform an abnormality diagnosis with higher accuracy.

(Third modification example)
In the third modification, the comparison in the diagnostic unit 12c is performed based on the image recognition.

When the diagnostic data is input, the diagnostic unit 12c generates a waveform image of the diagnostic data and calculates the degree of coincidence with the waveform image of the reference data stored in the storage unit 11. The diagnosis unit 12c diagnoses an abnormality when the degree of agreement is equal to or less than a predetermined value.

The above diagnosis uses a trained model trained with reference data as teacher data by a neural network such as a convolutional neural network (CNN). The trained model is included in the diagnosis unit 12c, and when the degree of agreement is equal to or less than a predetermined value, it is diagnosed that there is an abnormality. The trained model may be diagnosed as abnormal, for example, if the degree of agreement is less than 50%.

According to the third modification, the entire waveform can be compared as an image, so that the diagnostic accuracy can be improved.

(Second Embodiment)
The abnormality diagnosis device 10 of the motor (vibration machine) 1 of the second embodiment is different from the first embodiment in the abnormality diagnosis method. Other than this, it is substantially the same as the first embodiment. Therefore, the description of the portion shown in the first embodiment may be omitted.

In the present embodiment, the motor 1 is controlled by the control device 4 and the inverter 3, and the motor 1 is driven in a plurality of operation modes. For example, the plurality of operating modes include a cooling mode, a heating mode, and a blowing mode.

The storage unit 11 stores a plurality of reference data corresponding to the plurality of operation modes. Specifically, the storage unit 11 stores at least one reference data for each of the above operation modes. For example, the storage unit 11 may store a total of 15 reference data corresponding to the cooling mode, the heating mode, and the ventilation mode, 5 each.

The diagnostic unit 12c compares each of the plurality of reference data with the diagnostic data, acquires a plurality of comparison results, and if any one of the plurality of comparison results is not abnormal, the whole is obtained. As a result, it is diagnosed that it is not abnormal.

With reference to FIG. 10, the abnormality diagnosis method executed by the abnormality diagnosis device 10 of the present embodiment will be specifically described.

When the abnormality diagnosis method of the present embodiment is started (step S10-1), vibration data is acquired by the acquisition unit 12a (step S10-2). Next, the vibration data acquired by the acquisition unit 12a is Fourier transformed by the conversion unit 12b to obtain diagnostic data (step S10-3). Then, the diagnostic unit 12c compares the spectral value of one reference data stored in the storage unit 11 with the spectral value of the diagnostic data, and determines whether or not the diagnostic data differs from the reference data by a predetermined value or more (step). S10-4).

If the difference is not larger than a predetermined value in the determination process (step S10-4) (N: step S10-4), the process returns to the process of acquiring vibration data (step S10-2), and the subsequent processes are repeated. When the difference is larger than a predetermined value (Y: step S10-4), the spectral value of another reference data stored in the storage unit 11 is compared with the spectral value of the diagnostic data, and the diagnostic data differs from the reference data by a predetermined value or more. It is determined whether or not it is present (step S10-5).

If the difference is not larger than a predetermined value in the determination process (step S10-5) (N: step S10-5), the process returns to the process of acquiring vibration data (step S10-2), and the subsequent processes are repeated. When the difference is larger than a predetermined value (Y: step S10-5), the spectral value of another reference data stored in the storage unit 11 is compared with the spectral value of the diagnostic data, and the diagnostic data from the reference data is equal to or larger than the predetermined value. It is determined whether or not they are different (step S10-6).

When the comparison process is executed for all the reference data as in steps S10-4 to S10-6 and the diagnostic data is larger than a predetermined value for all the reference data (Y: step S10-6), the diagnosis is made. As a result, an abnormality is displayed on the display unit 13 (step S10-7), and a series of processes is completed (step S10-8). In other words, if any one of the plurality of comparison results is not abnormal (N: steps S10-4, S10-5, or S10-6), the overall result is diagnosed as normal. ..

According to the present embodiment, it is possible to suppress misdiagnosis of a change in vibration due to a change in the operation mode in the normal operation of the motor 1 as an abnormality. If, unlike this embodiment, only one reference data is prepared, the state of vibration changes when the operation mode is switched from the cooling mode to the heating mode, and normal operation accompanied by a mere change in the operation mode is performed. There is a risk of misdiagnosis as an abnormal condition. However, in the present embodiment, since the reference data corresponding to each of the cooling mode and the heating mode are prepared, the above misdiagnosis can be suppressed and the diagnosis accuracy can be improved.

Further, instead of comparing the reference data and the diagnostic data for each operation mode, it is possible to compare the reference data of all the other operation modes with the obtained diagnostic data. This makes it possible to suppress misdiagnosis by limiting the operation mode and performing diagnosis.

For example, it is generally expected to be used in the heating mode in winter. Therefore, based on this assumption, if the diagnosis is limited to the reference data of the heating mode in winter, there is a risk of misdiagnosis when the cooling mode is used in winter. However, in the present embodiment, since the diagnosis can be performed using the reference data corresponding to the cooling mode even in the winter season, such misdiagnosis can be suppressed. Alternatively, a plurality of reference data when only the carrier frequency of the inverter 3 is changed for one operation mode may be set. This makes it possible to suppress misdiagnosis when the carrier frequency of the inverter 3 is changed in the same operation mode. It should also be noted that, as shown in the first modification below, it may be useful to set the reference data according to the time.

(First modification)
In the first modification, the reference data corresponding to the operation mode different depending on the time is stored in the storage unit 11 as normal data, and the diagnosis unit 12c compares the reference data with the diagnosis data according to the time. For example, in the summer, the reference data corresponding to the cooling mode is stored in the storage unit 11 as normal data, and in the winter, the reference data corresponding to the heating mode is stored in the storage unit 11 as normal data. The reference data and the diagnostic data are compared according to the above.

In practice, for example, a year is from January 1st to March 31st, April 1st to June 30th, July 1st to September 30th, and October 1st to December 31st. The reference data corresponding to each period may be stored in the storage unit 11 as normal data, and the diagnosis unit 12c may compare each reference data with the diagnosis data according to each period.

The diagnostic unit 12c can properly use the reference data to be compared with the diagnostic data according to the above four timings. As a result, the number of reference data for comparative diagnosis can be reduced, so that the calculation cost can be reduced.

(Second modification)
In the second modification, a plurality of reference data corresponding to the automatic operation mode are stored in the storage unit 11 as normal data, and when the automatic operation mode is executed, the diagnosis unit 12c performs an abnormality diagnosis based on the reference data.

In the automatic operation mode, the operation mode may be switched automatically. In this modification, a plurality of reference data that can be switched for each automatic operation mode are stored in the storage unit 11, and the diagnostic unit 12c compares each of the plurality of reference data with the diagnostic data. If any one of the plurality of comparison results is not abnormal, the diagnostic unit 12c diagnoses that the overall result is not abnormal.

According to the second modification, even when the operation mode is automatically switched as in the automatic operation mode. Since the reference data of each operation mode can be set as normal data for diagnosis, accurate abnormality diagnosis can be performed.

(Third modification example)
In the third modification, the reference data is stored in the storage unit 11 for each operation mode. The diagnosis unit 12c grasps the current operation mode of the motor 1 and compares the reference data corresponding to the operation mode with the diagnosis data to perform an abnormality diagnosis.

According to the third modification, since the abnormality diagnosis is performed using the reference data corresponding to the current operation mode, it is possible to perform an accurate abnormality diagnosis according to the actual situation.

(Fourth modification)
In the fourth modification, the data after maintenance of the motor 1 is stored in the storage unit 11 for each device model number as reference data, and the diagnosis unit 12c performs diagnosis using the reference data for each device model number.

According to the fourth modification, since the data having a high possibility of normal operation after maintenance can be used as the reference data, the accuracy of the abnormality diagnosis can be improved.

Although the specific embodiment of the present invention and the modification thereof have been described above, the present invention is not limited to the above-described embodiment and can be variously modified and carried out within the scope of the present invention. For example, an embodiment of the present invention may be obtained by appropriately combining the contents of individual embodiments and modifications.

Further, as a comparison method in the diagnostic unit 12c, even if it is diagnosed as abnormal when the spectrum of the waveform (difference waveform) obtained by subtracting the waveform of the reference data from the waveform of the diagnostic data increases at a rate of a predetermined value or more as the frequency increases. good. Further, the abnormality diagnosis may be performed based on each index value such as the total value (integral value), the average value, the maximum value, the minimum value, or the correlation of the spectrum of the difference waveform in a predetermined frequency range.

Further, an existing statistical method may be used as a comparison method in the diagnostic unit 12c. Examples of the statistical method include a method using Spearman's rank correlation coefficient, Pearson product moment correlation coefficient, covariance, histogram, tolerance method, KL divergence, JS divergence, and the like. In any of the statistical methods, the diagnostic unit 12c diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more.

1 Motor (vibration machine)
2 Fan 3 Inverter 4 Control device 5 Vibration sensor 10 Abnormality diagnosis device 11 Storage unit 12 Processing unit 12a Acquisition unit 12b Conversion unit 12c Diagnosis unit 13 Display unit

The present invention relates to an abnormality diagnosing device for a vibrating machine and an abnormality diagnosing method.

In vibration machines such as rotary machines, abnormality diagnosis using vibration information is carried out. Patent Document 1 discloses a method for diagnosing a rotating machine using such vibration information. In the diagnostic method of Patent Document 1, a state evaluation index is calculated by a principal component analysis method from a plurality of dimensional vibration parameters calculated from vibration information of a rotating machine, and the quality of the rotating machine is judged based on the state evaluation index. There is.

Japanese Unexamined Patent Publication No. 2008-58191

With a diagnostic method as in Patent Document 1, it is difficult to capture changes over time in a rotating machine. In addition, it is necessary to prepare various dimensional vibration parameters, which is complicated as a diagnostic method. Further, when the vibration information is affected by the carrier frequency of the inverter or the like, it is necessary to remove the influence. Overall, there is room for improvement in the accuracy and convenience of abnormality diagnosis.

An object of the present invention is to improve the accuracy and convenience of diagnosis in an abnormality diagnosis device and an abnormality diagnosis method for a vibrating machine.

The first aspect of the present invention is
A storage unit that stores reference data obtained by Fourier transforming the vibration data of a vibrating machine,
An acquisition unit that acquires vibration data of the vibration machine over time,
A conversion unit that Fourier transforms the vibration data acquired by the acquisition unit into diagnostic data,
It is provided with a diagnostic unit that compares the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more .
The storage unit stores a plurality of reference data and stores a plurality of reference data.
When the diagnostic unit compares each of the plurality of reference data with the diagnostic data and obtains a plurality of comparison results, and even one of the plurality of comparison results is not abnormal. Provides an abnormality diagnostic device for a vibrating machine that diagnoses as a non-abnormal result as a whole .

According to this configuration, diagnostic data is generated from vibration data that changes with time of the vibration machine to be diagnosed, and abnormality diagnosis is performed by comparing with the reference data of the vibration machine to be diagnosed. Therefore, since the data to be compared are from the same vibration machine, it is possible to capture changes over time in the vibration data, and it is possible to perform highly accurate abnormality diagnosis. For example, in a vibration machine using an inverter, the influence of the carrier frequency of the inverter is similarly added to both the past vibration data and the current vibration data. Therefore, by taking these differences, it is possible to perform an abnormality diagnosis without being affected by the carrier frequency of the inverter, which is convenient. Further, since the vibration data is acquired over time and used as diagnostic data, it is possible to detect an abnormality when the change of the vibration data with time becomes large. The vibrating machine here broadly refers to a machine that can generate vibration, and includes, for example, a rotating machine, a reciprocating machine, and a pipe through which water flows. Further, the vibration data may be vibration acceleration, vibration velocity, or the like. In addition, it is possible to suppress a misdiagnosis of a change in vibration due to a change in the carrier frequency of the inverter or a change in the operation mode in the normal operation of the vibrating machine as an abnormal state.

The diagnostic unit may perform the comparison with an average value, an effective value, a minimum value, or a maximum value for each predetermined frequency band.

According to this configuration, abnormality diagnosis can be performed for each predetermined frequency band. Therefore, the cause of the abnormality can be identified according to the frequency band. The predetermined frequency band may be appropriately set according to the type of the vibrating machine and the frequency range of the acquired vibration data. The effective value is a quantity representing a time-averaged magnitude (intensity), and is the square root of the average of squared values.

The diagnostic unit may compare the peak value of the reference data with the peak value of the diagnostic data.

According to this configuration, abnormality diagnosis can be easily performed. In addition, large vibrations that take peak values often lead to abnormalities, and the accuracy of abnormality diagnosis can be improved by comparing the peak values. The comparison of the peak values here may be performed at a plurality of points in order from the largest value.

The second aspect of the present invention is
A storage unit that stores reference data obtained by Fourier transforming the vibration data of a vibrating machine,
An acquisition unit that acquires vibration data of the vibration machine over time,
A conversion unit that Fourier transforms the vibration data acquired by the acquisition unit into diagnostic data,
With a diagnostic unit that compares the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more.
Equipped with
The diagnostic unit provides an abnormality diagnosis device for a vibrating machine that evaluates the degree of coincidence between the waveform image of the reference data and the waveform image of the diagnostic data and diagnoses an abnormality when the degree of coincidence is equal to or less than a predetermined value. ..

According to this configuration, diagnostic data is generated from vibration data that changes with time of the vibration machine to be diagnosed, and abnormality diagnosis is performed by comparing with the reference data of the vibration machine to be diagnosed. Therefore, since the data to be compared are from the same vibration machine, it is possible to capture changes over time in the vibration data, and it is possible to perform highly accurate abnormality diagnosis. For example, in a vibration machine using an inverter, the influence of the carrier frequency of the inverter is similarly added to both the past vibration data and the current vibration data. Therefore, by taking these differences, it is possible to perform an abnormality diagnosis without being affected by the carrier frequency of the inverter, which is convenient. Further, since the vibration data is acquired over time and used as diagnostic data, it is possible to detect an abnormality when the change of the vibration data with time becomes large. The vibrating machine here broadly refers to a machine that can generate vibration, and includes, for example, a rotating machine, a reciprocating machine, and a pipe through which water flows. Further, the vibration data may be vibration acceleration, vibration velocity, or the like. Moreover, since the entire waveform can be compared as an image, the diagnostic accuracy can be improved. For example, an abnormality may be diagnosed when the degree of agreement is 50% or less.

A third aspect of the present invention is
Prepare the reference data obtained by Fourier transforming the vibration data of the vibrating machine.
The vibration data of the vibration machine is acquired over time, and the vibration data is acquired over time.
The vibration data acquired over time is Fourier transformed to obtain diagnostic data.
It includes comparing the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnosing an abnormality when the diagnostic data differs from the reference data by a predetermined value or more .
The diagnosis provides a method for diagnosing an abnormality of a vibrating machine, which evaluates the degree of coincidence between the waveform image of the reference data and the waveform image of the diagnostic data, and diagnoses an abnormality when the degree of coincidence is equal to or less than a predetermined value .

A fourth aspect of the present invention is
Prepare the reference data obtained by Fourier transforming the vibration data of the vibrating machine.
The vibration data of the vibration machine is acquired over time, and the vibration data is acquired over time.
The vibration data acquired over time is Fourier transformed to obtain diagnostic data.
The spectral value of the reference data and the spectral value of the diagnostic data are compared over time, and when the diagnostic data differs from the reference data by a predetermined value or more, an abnormality is diagnosed.
Including that
In the preparation, a plurality of the reference data are prepared, and
In the diagnosis, each of the plurality of reference data is compared with the diagnostic data to obtain a plurality of comparison results, and when even one of the plurality of comparison results is not abnormal. Provides a method for diagnosing abnormalities in vibrating machines, which diagnoses as a non-abnormal result as a whole.

According to the present invention, the accuracy and convenience of abnormality diagnosis can be improved in the abnormality diagnosis device and the abnormality diagnosis method of a vibrating machine.

Schematic block diagram of the abnormality diagnosis device for vibration machines. The control block diagram of the abnormality diagnosis apparatus which concerns on 1st Embodiment of this invention. The flowchart which shows the abnormality diagnosis method of 1st Embodiment. A graph showing the waveforms of reference data and diagnostic data. The bar graph corresponding to the reference data D0 of FIG. The bar graph corresponding to the diagnostic data D4 of FIG. A line graph showing reference data in the first modification. A line graph showing diagnostic data in the first modification. A line graph showing the difference between the diagnostic data and the reference data in the second modification. The flowchart which shows the abnormality diagnosis method of 2nd Embodiment.

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

(First Embodiment)
With reference to FIG. 1, the abnormality diagnosis device 10 of the vibration machine 1 of the first embodiment acquires the vibration data of the vibration machine 1 and diagnoses the abnormality from the vibration data. The vibration machine 1 broadly refers to a machine that can generate vibration, and includes, for example, a rotating machine, a reciprocating machine, and a pipe through which a liquid flows. Hereinafter, the motor 1 of the fan 2 used in an air conditioner such as an air conditioner will be described as an example of the vibration machine 1.

The fan 2 used in the air conditioner is driven by a motor 1 and blows temperature-controlled air into the room. The rotation speed of the motor 1 is adjusted by the inverter 3. The inverter 3 is controlled by the control device 4.

A vibration sensor 5 is attached to the motor 1. The vibration sensor 5 measures the vibration data of the motor 1 and transmits it to the abnormality diagnosis device 10. The vibration data can be vibration acceleration, vibration velocity, and the like. Here, vibration acceleration will be described as an example of vibration data. The vibration sensor 5 measures the vibration acceleration of the motor 1 at predetermined intervals (for example, once a day) and transmits it to the abnormality diagnosis device 10. In this embodiment, the vibration sensor 5 measures the vibration acceleration in the frequency range of 1 to 10000 Hz. Alternatively, the vibration sensor 5 may measure vibration data in any frequency range. For example, the vibration sensor 5 may acquire vibration data up to a frequency higher than 10000 Hz.

With reference to FIG. 2, the abnormality diagnosis device 10 is composed of hardware such as a CPU (Central Processing Unit), a RAM (Random Access Memory), and a ROM (Read Only Memory), and software mounted therein. ing. The abnormality diagnosis device 10 may be composed of an information processing device such as a desktop personal computer, a laptop computer, a workstation, or a tablet terminal, for example.

The abnormality diagnosis device 10 has a storage unit 11, a processing unit 12, and a display unit 13. The processing unit 12 includes an acquisition unit 12a, a conversion unit 12b, and a diagnosis unit 12c as a functional configuration. These are realized by the cooperation between the processor, which is a hardware resource, and the program, which is software.

The storage unit 11 stores reference data obtained by Fourier transforming the past vibration data of the motor 1. The past vibration data is the one when the normal operation state is maintained. Further, the storage unit 11 also records a program operated by the processing unit 12. The storage unit 11 may be a portable recording medium, or may be integrally configured with the processing unit 12.

The acquisition unit 12a receives vibration data from the vibration sensor 5 over time. The mode of exchanging data between the acquisition unit 12a and the vibration sensor 5 may be wired or wireless.

The conversion unit 12b Fourier transforms the vibration data acquired by the acquisition unit 12a into diagnostic data. Here, a fast Fourier transform (FFT) is used as an example of the Fourier transform. By this Fourier transform, the vibration data can be frequency-analyzed. Further, the conversion in the conversion unit 12b is executed every time the vibration data is acquired by the acquisition unit 12a. Therefore, it is possible to confirm the change over time in the frequency analysis result.

The diagnostic unit 12c compares the spectral value of the reference data with the spectral value of the diagnostic data over time, and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more. The specific method of comparison will be described later.

The display unit 13 is a part that displays the diagnosis result by the processing unit 12, and is composed of, for example, a liquid crystal display, an organic EL display, a plasma display, or the like. In the present embodiment, the display unit 13 displays the diagnosis result when the diagnosis unit 12c gives a diagnosis result that there is an abnormality. However, the display unit 13 may display the diagnosis result even when the diagnosis unit 12c gives a diagnosis result that there is no abnormality.

The abnormality diagnosis method executed by the abnormality diagnosis apparatus 10 of the present embodiment will be described with reference to FIG.

When the abnormality diagnosis method of the present embodiment is started (step S3-1), vibration data is acquired by the acquisition unit 12a (step S3-2). Next, the vibration data acquired by the acquisition unit 12a is Fourier transformed by the conversion unit 12b to obtain diagnostic data (step S3-3). Then, the diagnostic unit 12c compares the spectral value of the reference data stored in the storage unit 11 with the spectral value of the diagnostic data, and diagnoses as abnormal when the diagnostic data differs from the reference data by a predetermined value or more (step S3-4). ). If it is not diagnosed that there is an abnormality (N: step S3-4), the process returns to the process of acquiring vibration data (step S3-2), and the subsequent processes are repeated. When an abnormality is diagnosed by the diagnosis unit 12c (Y: step S3-4), the display unit 13 displays that there is an abnormality as a diagnosis result (step S3-5), and the series of processes is terminated (step). S3-6).

With reference to FIG. 4, the data to be compared in this embodiment will be illustrated. FIG. 4 is a three-dimensional graph having three axes (spectral values, frequencies, and dates). In this graph, the reference data and the diagnostic data are shown in 1 Hz increments, respectively.

The waveform data D0 located on the leftmost side of FIG. 4 is reference data and is stored in the storage unit 11 as normal data. The second and subsequent waveform data D1 to D9 in FIG. 4 are diagnostic data, and the vibration data acquired at intervals of once every two weeks (for example, the acceleration data of the vibration of the motor 1) are Fourier transformed. ..

In the graph of FIG. 4, the diagnostic data D4 on May 12, 2019 is superior to the reference data D0 in the vicinity of F1 and F2 (Hz) (see the part A in the graph). Specifically, no such predominant spectrum was identified in the diagnostic data D1 to D3 prior to May 12, 2019, and near the same frequency in the diagnostic data D5 to D9 after May 12, 2019. You can see the excellent spectrum with.

FIG. 5 shows a bar graph corresponding to the reference data D0 of FIG. FIG. 6 shows a bar graph corresponding to the diagnostic data D4 of FIG. In FIGS. 5 and 6, the horizontal axis indicates the frequency. The frequency is shown by dividing 1 to 10000 Hz by a predetermined division width. The vertical axis shows the average value of the spectrum in each divided frequency range. Regarding the predetermined division width, 1 to 10000 Hz may be equally divided, or the division width may be arbitrarily changed.

Comparing FIGS. 5 and 6, in the frequency range near F1 (Hz), which is the 14th frequency range, the average value of the spectra of the diagnostic data D4 is significantly larger than the average value of the spectra of the reference data D0. Can be confirmed. Similarly, in the frequency range near F2 (Hz), which is the 16th frequency range, it can be confirmed that the average value of the spectrum of the diagnostic data D4 is significantly larger than the average value of the spectrum of the reference data D0.

In the present embodiment, the diagnostic unit 12c has the average value of the spectra of any specific frequency range of the diagnostic data D1 to D9 larger than the average value of the spectra of the same frequency range of the reference data, for example, by 30% or more. If it is diagnosed as abnormal. In the examples of FIGS. 5 and 6, in the frequency range near F1 (Hz) and F2 (Hz), the average value of the spectrum of the diagnostic data D4 is more than twice as large as the average value of the spectrum of the reference data D0, respectively. ing. Therefore, the diagnosis unit 12c diagnoses the abnormality. The threshold value for comparison for diagnosing an abnormality may be other than 30%, and may be arbitrarily set according to the type of the vibrating machine to be diagnosed.

According to the present embodiment, diagnostic data is generated from vibration data of the motor 1 that changes with time, and abnormality diagnosis is performed by comparing with the reference data of the same motor 1 acquired in the past. Therefore, since the data to be compared is from the same motor 1, it is possible to capture the change over time in the vibration data, and it is possible to perform highly accurate abnormality diagnosis. Further, in the present embodiment, since the rotation speed of the motor 1 is controlled by the inverter 3, noise can be added to the vibration data at the carrier frequency of the inverter 3. However, the influence of the carrier frequency of the inverter 3 is similarly added to both the past vibration data and the current vibration data. Therefore, by taking these differences, the influence of the noise of the carrier frequency of the inverter 3 can be suppressed and the abnormality diagnosis can be easily performed. Further, since the vibration data is acquired over time and used as diagnostic data, it is possible to detect an abnormality when the change of the vibration data with time becomes large.

Further, according to the present embodiment, abnormality diagnosis can be performed for each predetermined frequency band. Therefore, the cause of the abnormality can be identified according to the frequency band. For example, when an abnormality is diagnosed in a low frequency region (for example, 1000 Hz or less), eccentricity (unbalance) or eccentricity (misalignment) of the motor 1 and rattling of the installation of the motor 1 are assumed. In addition, when the target is a pipe through which a liquid flows as a vibration machine, an excessive flow rate in the pipe is assumed. Further, when an abnormality is diagnosed in a high frequency region (for example, 5000 Hz or higher), it is assumed that the bearing portion supporting the shaft member that mechanically connects the motor 1 and the fan 2 is damaged or the grease is cut off.

In this embodiment, the mean value of the spectrum in a specific frequency range is confirmed and the reference data and the diagnostic data are compared, but the value to be compared is the effective value, the minimum value, or the maximum value of the spectrum in the specific frequency range. It may be a value or the like.

(First modification)
With reference to FIGS. 7 and 8, the peak value may be used as a method for comparing the diagnostic data and the reference data by the diagnostic unit 12c.

In the first modification, the diagnostic unit 12c compares the peak value of the reference data and the peak value of the diagnostic data in descending order, and diagnoses as abnormal when they differ by a predetermined value or more. Hereinafter, specific examples will be described.

FIG. 7 shows reference data and FIG. 8 shows diagnostic data. The horizontal axis of FIGS. 7 and 8 shows the frequency, and the vertical axis shows the spectrum.

In the graph of FIG. 7, the largest first peak is indicated by reference numeral P7-1 and the second largest peak is indicated by reference numeral P7-2. Similarly, in FIG. 8, the largest first peak is indicated by reference numeral P8-1 and the second largest peak is indicated by reference numeral P8-2.

In the examples of FIGS. 7 and 8, the first peak value of about 5.1 (P8-1) of the diagnostic data is about 13% larger than the first peak value of about 4.5 (P7-1) of the reference data. Further, the second peak value of about 4.2 (P8-1) of the diagnostic data is about 17% larger than the second peak value of about 3.6 (P7-1) of the reference data.

In the diagnostic unit 12c, as described above, the peak values are compared in descending order. When the threshold value for diagnosing an abnormality is set to 30% as in the above embodiment, in the examples of FIGS. 7 and 8, the comparison result exceeding the 30% cannot be confirmed, so that the abnormality is not diagnosed. However, the number of peak values to be compared may be one or three or more, and the threshold value is not limited to 30%.

According to the first modification, since the peak values are simply compared in descending order regardless of the frequency range, the abnormality diagnosis can be easily performed. Specifically, as shown in the examples of FIGS. 7 and 8, the frequency of the first peak P7-1 of the reference data and the frequency of the first peak P8-1 of the diagnostic data are different from each other. Therefore, in such a case, the peak values of different frequencies are compared with each other. However, a specific frequency range may be set as in the above embodiment, and peak values in the specific frequency range may be compared with each other.

In addition, large vibrations that take peak values often lead to abnormalities, and the accuracy of abnormality diagnosis can be improved by extracting and comparing peak values.

(Second modification)
With reference to FIG. 9, as a method of comparing the diagnostic data and the reference data by the diagnostic unit 12c, the difference waveform between the diagnostic data and the reference data may be used.

In the second modification, the diagnostic unit 12c diagnoses an abnormality when the waveform (difference waveform) obtained by subtracting the spectral value of the reference data from the spectral value of the diagnostic data becomes a slope waveform. Hereinafter, specific examples will be described.

The horizontal axis of FIG. 9 shows the frequency, and the vertical axis shows the spectrum. FIG. 9 shows a difference value obtained by subtracting the waveform of the reference data from the waveform of the diagnostic data.

In the graph of FIG. 9, the slope waveform can be confirmed in the vicinity of the frequency of 10000 Hz (see the reference numeral P9 portion). Here, the slope waveform means a waveform in which the FFT spectrum rises in a slope shape from around 2000 Hz to a lower frequency side, or from around 9000 Hz to a higher frequency side than that.

In the graph of FIG. 9, the spectral value increases about 8 times from about 0.2 to about 1.6 as the frequency increases from 9001 to 10000 Hz. Therefore, since the slope waveform is confirmed in the frequency range, the diagnosis unit 12c diagnoses that there is an abnormality in the frequency range.

According to this modification, an abnormality is likely to occur in the slope waveform, and by detecting the occurrence of the waveform, it is possible to perform an abnormality diagnosis with higher accuracy.

(Third modification example)
In the third modification, the comparison in the diagnostic unit 12c is performed based on the image recognition.

When the diagnostic data is input, the diagnostic unit 12c generates a waveform image of the diagnostic data and calculates the degree of coincidence with the waveform image of the reference data stored in the storage unit 11. The diagnosis unit 12c diagnoses an abnormality when the degree of agreement is equal to or less than a predetermined value.

The above diagnosis uses a trained model trained with reference data as teacher data by a neural network such as a convolutional neural network (CNN). The trained model is included in the diagnosis unit 12c, and when the degree of agreement is equal to or less than a predetermined value, it is diagnosed that there is an abnormality. The trained model may be diagnosed as abnormal, for example, if the degree of agreement is less than 50%.

According to the third modification, the entire waveform can be compared as an image, so that the diagnostic accuracy can be improved.

(Second Embodiment)
The abnormality diagnosis device 10 of the motor (vibration machine) 1 of the second embodiment is different from the first embodiment in the abnormality diagnosis method. Other than this, it is substantially the same as the first embodiment. Therefore, the description of the portion shown in the first embodiment may be omitted.

In the present embodiment, the motor 1 is controlled by the control device 4 and the inverter 3, and the motor 1 is driven in a plurality of operation modes. For example, the plurality of operating modes include a cooling mode, a heating mode, and a blowing mode.

The storage unit 11 stores a plurality of reference data corresponding to the plurality of operation modes. Specifically, the storage unit 11 stores at least one reference data for each of the above operation modes. For example, the storage unit 11 may store a total of 15 reference data corresponding to the cooling mode, the heating mode, and the ventilation mode, 5 each.

The diagnostic unit 12c compares each of the plurality of reference data with the diagnostic data, acquires a plurality of comparison results, and if any one of the plurality of comparison results is not abnormal, the whole is obtained. As a result, it is diagnosed that it is not abnormal.

With reference to FIG. 10, the abnormality diagnosis method executed by the abnormality diagnosis device 10 of the present embodiment will be specifically described.

When the abnormality diagnosis method of the present embodiment is started (step S10-1), vibration data is acquired by the acquisition unit 12a (step S10-2). Next, the vibration data acquired by the acquisition unit 12a is Fourier transformed by the conversion unit 12b to obtain diagnostic data (step S10-3). Then, the diagnostic unit 12c compares the spectral value of one reference data stored in the storage unit 11 with the spectral value of the diagnostic data, and determines whether or not the diagnostic data differs from the reference data by a predetermined value or more (step). S10-4).

If the difference is not larger than a predetermined value in the determination process (step S10-4) (N: step S10-4), the process returns to the process of acquiring vibration data (step S10-2), and the subsequent processes are repeated. When the difference is larger than a predetermined value (Y: step S10-4), the spectral value of another reference data stored in the storage unit 11 is compared with the spectral value of the diagnostic data, and the diagnostic data differs from the reference data by a predetermined value or more. It is determined whether or not it is present (step S10-5).

If the difference is not larger than a predetermined value in the determination process (step S10-5) (N: step S10-5), the process returns to the process of acquiring vibration data (step S10-2), and the subsequent processes are repeated. When the difference is larger than a predetermined value (Y: step S10-5), the spectral value of another reference data stored in the storage unit 11 is compared with the spectral value of the diagnostic data, and the diagnostic data from the reference data is equal to or larger than the predetermined value. It is determined whether or not they are different (step S10-6).

When the comparison process is executed for all the reference data as in steps S10-4 to S10-6 and the diagnostic data is larger than a predetermined value for all the reference data (Y: step S10-6), the diagnosis is made. As a result, an abnormality is displayed on the display unit 13 (step S10-7), and a series of processes is completed (step S10-8). In other words, if any one of the plurality of comparison results is not abnormal (N: steps S10-4, S10-5, or S10-6), the overall result is diagnosed as normal. ..

According to the present embodiment, it is possible to suppress misdiagnosis of a change in vibration due to a change in the operation mode in the normal operation of the motor 1 as an abnormality. If, unlike this embodiment, only one reference data is prepared, the state of vibration changes when the operation mode is switched from the cooling mode to the heating mode, and normal operation accompanied by a mere change in the operation mode is performed. There is a risk of misdiagnosis as an abnormal condition. However, in the present embodiment, since the reference data corresponding to each of the cooling mode and the heating mode are prepared, the above misdiagnosis can be suppressed and the diagnosis accuracy can be improved.

Further, instead of comparing the reference data and the diagnostic data for each operation mode, it is possible to compare the reference data of all the other operation modes with the obtained diagnostic data. This makes it possible to suppress misdiagnosis by limiting the operation mode and performing diagnosis.

For example, it is generally expected to be used in the heating mode in winter. Therefore, based on this assumption, if the diagnosis is limited to the reference data of the heating mode in winter, there is a risk of misdiagnosis when the cooling mode is used in winter. However, in the present embodiment, since the diagnosis can be performed using the reference data corresponding to the cooling mode even in the winter season, such misdiagnosis can be suppressed. Alternatively, a plurality of reference data when only the carrier frequency of the inverter 3 is changed for one operation mode may be set. This makes it possible to suppress misdiagnosis when the carrier frequency of the inverter 3 is changed in the same operation mode. It should also be noted that, as shown in the first modification below, it may be useful to set the reference data according to the time.

(First modification)
In the first modification, the reference data corresponding to the operation mode different depending on the time is stored in the storage unit 11 as normal data, and the diagnosis unit 12c compares the reference data with the diagnosis data according to the time. For example, in the summer, the reference data corresponding to the cooling mode is stored in the storage unit 11 as normal data, and in the winter, the reference data corresponding to the heating mode is stored in the storage unit 11 as normal data. The reference data and the diagnostic data are compared according to the above.

In practice, for example, a year is from January 1st to March 31st, April 1st to June 30th, July 1st to September 30th, and October 1st to December 31st. The reference data corresponding to each period may be stored in the storage unit 11 as normal data, and the diagnosis unit 12c may compare each reference data with the diagnosis data according to each period.

The diagnostic unit 12c can properly use the reference data to be compared with the diagnostic data according to the above four timings. As a result, the number of reference data for comparative diagnosis can be reduced, so that the calculation cost can be reduced.

(Second modification)
In the second modification, a plurality of reference data corresponding to the automatic operation mode are stored in the storage unit 11 as normal data, and when the automatic operation mode is executed, the diagnosis unit 12c performs an abnormality diagnosis based on the reference data.

In the automatic operation mode, the operation mode may be switched automatically. In this modification, a plurality of reference data that can be switched for each automatic operation mode are stored in the storage unit 11, and the diagnostic unit 12c compares each of the plurality of reference data with the diagnostic data. If any one of the plurality of comparison results is not abnormal, the diagnostic unit 12c diagnoses that the overall result is not abnormal.

According to the second modification, even when the operation mode is automatically switched as in the automatic operation mode. Since the reference data of each operation mode can be set as normal data for diagnosis, accurate abnormality diagnosis can be performed.

(Third modification example)
In the third modification, the reference data is stored in the storage unit 11 for each operation mode. The diagnosis unit 12c grasps the current operation mode of the motor 1 and compares the reference data corresponding to the operation mode with the diagnosis data to perform an abnormality diagnosis.

According to the third modification, since the abnormality diagnosis is performed using the reference data corresponding to the current operation mode, it is possible to perform an accurate abnormality diagnosis according to the actual situation.

(Fourth modification)
In the fourth modification, the data after maintenance of the motor 1 is stored in the storage unit 11 for each device model number as reference data, and the diagnosis unit 12c performs diagnosis using the reference data for each device model number.

According to the fourth modification, since the data having a high possibility of normal operation after maintenance can be used as the reference data, the accuracy of the abnormality diagnosis can be improved.

Although the specific embodiment of the present invention and the modification thereof have been described above, the present invention is not limited to the above-described embodiment and can be variously modified and carried out within the scope of the present invention. For example, an embodiment of the present invention may be obtained by appropriately combining the contents of individual embodiments and modifications.

Further, as a comparison method in the diagnostic unit 12c, even if it is diagnosed as abnormal when the spectrum of the waveform (difference waveform) obtained by subtracting the waveform of the reference data from the waveform of the diagnostic data increases at a rate of a predetermined value or more as the frequency increases. good. Further, the abnormality diagnosis may be performed based on each index value such as the total value (integral value), the average value, the maximum value, the minimum value, or the correlation of the spectrum of the difference waveform in a predetermined frequency range.

Further, an existing statistical method may be used as a comparison method in the diagnostic unit 12c. Examples of the statistical method include a method using Spearman's rank correlation coefficient, Pearson product moment correlation coefficient, covariance, histogram, tolerance method, KL divergence, JS divergence, and the like. In any of the statistical methods, the diagnostic unit 12c diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more.

1 Motor (vibration machine)
2 Fan 3 Inverter 4 Control device 5 Vibration sensor 10 Abnormality diagnosis device 11 Storage unit 12 Processing unit 12a Acquisition unit 12b Conversion unit 12c Diagnosis unit 13 Display unit

Claims (6)

A storage unit that stores reference data obtained by Fourier transforming the vibration data of a vibrating machine,
An acquisition unit that acquires vibration data of the vibration machine over time,
A conversion unit that Fourier transforms the vibration data acquired by the acquisition unit into diagnostic data,
An abnormality diagnostic device for a vibrating machine, comprising a diagnostic unit that compares the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnoses an abnormality when the diagnostic data differs from the reference data by a predetermined value or more. .. The abnormality diagnosis device for a vibrating machine according to claim 1, wherein the diagnostic unit performs the comparison with an average value, an effective value, a minimum value, or a maximum value for each predetermined frequency band. The abnormality diagnosis device for a vibrating machine according to claim 1, wherein the diagnostic unit compares the peak value of the reference data with the peak value of the diagnostic data. The vibrating machine according to claim 1, wherein the diagnostic unit evaluates the degree of coincidence between the waveform image of the reference data and the waveform image of the diagnostic data, and diagnoses an abnormality when the degree of coincidence is equal to or less than a predetermined value. Abnormality diagnostic device. The storage unit stores a plurality of reference data and stores a plurality of reference data.
When the diagnostic unit compares each of the plurality of reference data with the diagnostic data and obtains a plurality of comparison results, and even one of the plurality of comparison results is not abnormal. The abnormality diagnosis device for a vibrating machine according to any one of claims 1 to 3, which is diagnosed as not abnormal as a whole. Prepare the reference data obtained by Fourier transforming the vibration data of the vibrating machine.
The vibration data of the vibration machine is acquired over time, and the vibration data is acquired over time.
The vibration data acquired over time is Fourier transformed to obtain diagnostic data.
A method for diagnosing an abnormality of a vibrating machine, which comprises comparing the spectral value of the reference data with the spectral value of the diagnostic data over time and diagnosing an abnormality when the diagnostic data differs from the reference data by a predetermined value or more.

Images