Method and device for fault diagnosis of ac motor based on modulation bispectrum and storage medium

Abstract

The present application belongs to the technical field of motor fault diagnosis, and particularly relates to an AC motor fault diagnosis method and device based on modulation bispectrum and a storage medium. In order to solve the problem of signal distortion in motor fault diagnosis, the present application comprises: collecting current signals and voltage signals of an AC motor in real time, performing short-time Fourier transform on the current signals, and extracting a unimodal ridge R1 of a time-frequency spectrum TFS1; performing time-frequency analysis on the current signals based on modulation bispectrum, and extracting a unimodal ridge R2 of a time-frequency spectrum TFS2; obtaining phase average power PI according to the current signals, the voltage signals and the unimodal ridge R1; determining a characteristic frequency threshold H1 for judging whether a fault occurs according to a ratio of the phase average power PI to rated power PR of the AC motor; obtaining a fault characteristic frequency amplitude HF according to the current signals I1 and the unimodal ridge R2; and judging whether the AC motor has a motor broken bar fault and the severity of the fault according to a ratio of the fault characteristic frequency amplitude HF to the characteristic frequency threshold H1.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of motor fault diagnosis, and particularly relates to an AC motor fault diagnosis method and device based on modulation bispectrum and a storage medium. BACKGROUND

[0002] Existing motor fault diagnosis usually judges the motor state by monitoring vibration, current, voltage and other signals, among which vibration signal analysis is the most widely used, and the fault diagnosis thereof depends on high-quality sensor data, but in actual working conditions, signal distortion is often caused by noise and transmission path interference.

[0003] At present, modulation bispectrum is used in gear and bearing fault diagnosis, but it has not been applied to motor fault diagnosis, and the current fundamental frequency accounts for the main component in the current signal, so the online fault diagnosis method based on modulation bispectrum can effectively remove the fundamental frequency component and highlight the fault frequency characteristics. SUMMARY

[0004] The application provides an AC motor fault diagnosis method and device based on modulation bispectrum and a storage medium to solve the above problems.

[0005] To achieve the above purpose, the application adopts the following technical scheme:

[0006] In a first aspect, the application provides an AC motor fault diagnosis method based on modulation bispectrum, including the following steps:

[0007] Real-time acquisition of phase current I0 and phase voltage U0 of the AC motor, pretreatment of the phase current I0 and phase voltage U0 to obtain current signal I1 and voltage signal U1;

[0008] Short-time Fourier transform of the current signal I1 to obtain time-frequency spectrum TFS1 of the current signal I1, extraction of unimodal ridge R1 of the time-frequency spectrum TFS1;

[0009] Time-frequency analysis of the current signal I1 based on modulation bispectrum, and obtaining time-frequency spectrum TFS2 of the current signal I1 based on the unimodal ridge R1, extraction of unimodal ridge R2 of the time-frequency spectrum TFS2;

[0010] Obtaining phase average power PI according to the current signal I1 and voltage signal U1 and unimodal ridge R1;

[0011] Determination of a characteristic frequency threshold H1 for judging whether a fault occurs according to the ratio of the phase average power PI to the rated power PR of the AC motor;

[0012] Obtaining fault characteristic frequency amplitude HF according to the current signal I1 and unimodal ridge R2;

[0013] determining whether the AC motor has a broken bar fault and a fault severity level by a ratio of the fault characteristic frequency amplitude HF and the characteristic frequency threshold H1.

[0014] Further, the current signal I1 is subjected to modulation bispectrum-based time-frequency analysis, and a time-frequency spectrum TFS2 of the current signal I1 is obtained based on the single-peak ridge R1, and a single-peak ridge R2 of the time-frequency spectrum TFS2 is extracted, specifically as follows.

[0015] The current signal I1 is divided into X segments according to a preset segmentation method, and X current signals IN1 are obtained.

[0016] The X current signals IN1 are subjected to windowing processing respectively, and X time-domain signals IN2 are obtained.

[0017] The X time-domain signals IN2 are subjected to modulation bispectrum analysis respectively, and X three-dimensional spectra SD1 are obtained.

[0018] According to the middle time TN2 of the X time-domain signals IN2 and the frequency extreme FN2 corresponding to the single-peak ridge R1, the corresponding three-dimensional spectrum SD1 is sliced along the frequency extreme FN2, and X two-dimensional spectra TDS1 are obtained.

[0019] The X two-dimensional spectra TDS1 are arranged according to the order of the middle time, and a time-frequency spectrum TFS2 of the current signal I1 is obtained.

[0020] The single-peak ridge R2 of the time-frequency spectrum TFS2 of the current signal I1 is extracted.

[0021] Further, the current signal I1 is divided into X segments according to a preset segmentation method, specifically as follows.

[0022] The current signal I1 is segmented according to a length N from the first data, and the starting data of each data segment is spaced by M data in turn, and when the remaining data length is less than N, the segmentation is stopped, and X current signals are obtained.

[0023] Further, the phase average power PI is obtained according to the current signal I1, the voltage signal U1, and the single-peak ridge R1, specifically as follows.

[0024] The time Tt corresponding to a certain point It in the current signal I1 is obtained according to the single-peak ridge R1, and the frequency Pt corresponding to Tt is obtained.

[0025] The phase average power PI is obtained by the frequency Pt, the current signal I1, and the voltage signal U1.

[0026] Further, the feature frequency threshold H1 for determining whether a fault occurs is determined according to a ratio of the phase average power PI and a rated power PR of the AC motor, and specifically:

[0027] The rated power PR is obtained through the self parameters of the AC motor, an initial threshold H0 is set, and the feature frequency threshold H1 for determining whether a fault occurs is obtained according to a ratio of the phase average power PI and the rated power PR and the initial threshold H0.

[0028] Further, the fault feature frequency amplitude HF is obtained according to the current signal I1 and the single peak ridge R2, and specifically:

[0029] The time Tt corresponding to any point It in the current signal I1 is determined, and the fault feature frequency amplitude HF corresponding to Tt is obtained according to the single peak ridge R2.

[0030] In a second aspect, the present application provides an AC motor fault diagnosis device based on modulation bispectrum, which comprises:

[0031] A data acquisition module is configured to acquire a phase current I0 and a phase voltage U0 of an AC motor in real time, and to preprocess the phase current I0 and the phase voltage U0 to obtain a current signal I1 and a phase voltage U1.

[0032] A single peak ridge R1 extraction module is configured to perform short-time Fourier transform on the current signal I1 to obtain a time-frequency spectrum TFS1 of the current signal I1, and to extract a single peak ridge R1 of the time-frequency spectrum TFS1.

[0033] A single peak ridge R2 extraction module is configured to perform time-frequency analysis on the current signal I1 based on modulation bispectrum, to obtain a time-frequency spectrum TFS2 of the current signal I1 based on the single peak ridge R1, and to extract a single peak ridge R2 of the time-frequency spectrum TFS2.

[0034] A phase average power acquisition module is configured to obtain a phase average power PI according to the current signal I1, the voltage signal U1, and the single peak ridge R1.

[0035] A feature frequency threshold acquisition module is configured to determine a feature frequency threshold H1 for determining whether a fault occurs according to a ratio of the phase average power PI and a rated power PR of the AC motor.

[0036] A feature frequency amplitude acquisition module is configured to obtain a fault feature frequency amplitude HF according to the current signal I1 and the single peak ridge R2.

[0037] A fault determination module is configured to determine whether an AC motor fault occurs and the severity of the fault by comparing the fault feature frequency amplitude HF and the feature frequency threshold H1.

[0038] In a third aspect, the present application provides a storage medium, which stores a computer program, when the computer program is executed by a host, each step of the method for diagnosing faults of an AC motor based on modulation bispectrum is realized.

[0039] Compared with the prior art, the present application has the following advantages:

[0040] The present application acquires the current signal and the voltage signal of the AC motor in real time, uses short-time Fourier transform to perform time-frequency analysis on the current signal, and extracts the single-peak ridge R1 of the time-frequency spectrum TFS1; performs time-frequency analysis on the current signal based on the time-frequency analysis method of modulation bispectrum, and extracts the single-peak ridge R2 of the time-frequency spectrum TFS2; obtains the phase average power PI according to the current signal, the voltage signal and the single-peak ridge R1; determines the characteristic frequency threshold H1 for judging whether a fault occurs according to the ratio of the phase average power PI and the rated power PR of the AC motor; obtains the fault characteristic frequency amplitude HF according to the current signal I1 and the single-peak ridge R2; judges whether the motor bar breakage fault of the AC motor occurs and the severity of the fault through the ratio of the fault characteristic frequency amplitude HF and the characteristic frequency threshold H1; the modulation bispectrum is applied to the motor fault diagnosis through the above-mentioned scheme of the present application, which can effectively remove the fundamental frequency component and highlight the fault frequency characteristics, and at the same time, the modulation bispectrum can effectively filter the modulation component of the frequency, which is conducive to the judgment of the fault characteristics. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flowchart of the method for diagnosing faults of an AC motor based on modulation bispectrum;

[0042] Figure 2 is a structural diagram of the device for diagnosing faults of an AC motor based on modulation bispectrum. DETAILED DESCRIPTION

[0043] In order to further illustrate the technical scheme of the present application, the present application will be further described through examples. Example 1

[0044] As shown in the figure, the method for diagnosing faults of an AC motor based on modulation bispectrum of the present embodiment comprises the following steps: Figure 1

[0045] S1, the phase current I0 and the phase voltage U0 of the AC motor are acquired in real time, the phase current I0 and the phase voltage U0 are preprocessed to obtain the current signal I1 and the voltage signal U1;

[0046] S2, the current signal I1 is subjected to short-time Fourier transform to obtain the time-frequency spectrum TFS1 of the current signal I1, and the single-peak ridge R1 of the time-frequency spectrum TFS1 is extracted; ​

[0047] S3, performing modulation bispectrum-based time-frequency analysis on the current signal I1, and obtaining a time-frequency spectrum TFS2 of the current signal I1 based on the single-peak ridge R1, and extracting a single-peak ridge R2 of the time-frequency spectrum TFS2;

[0048] S4, obtaining a phase average power PI according to the current signal I1 and the voltage signal U1 and the single-peak ridge R1;

[0049] S5, determining a characteristic frequency threshold H1 for judging whether a fault occurs according to a ratio of the phase average power PI and a rated power PR of the AC motor;

[0050] S6, obtaining a fault characteristic frequency amplitude HF according to the current signal I1 and the single-peak ridge R2;

[0051] S7, judging whether the AC motor has a motor broken bar fault and a fault severity according to a ratio of the fault characteristic frequency amplitude HF and the characteristic frequency threshold H1.

[0052] It can be understood that the embodiment uses a current transformer and a voltage transformer to collect a phase current I0 and a phase voltage U0 of the AC motor, uses a Chebyshev low-pass filter to filter high-frequency components in the phase current I0 and the phase voltage U0, and uses a sliding window method to remove abnormal data in the filtered current, to obtain a current signal I1 and a voltage signal U1.

[0053] Performing short-time Fourier transform on the current signal I1 to obtain a time-frequency spectrum TFS1 ([time, frequency, amplitude]) of the current signal I1, and extracting a single-peak ridge R1 ([time, frequency value, maximum amplitude]) of the time-frequency spectrum TFS1.

[0054] Performing modulation bispectrum-based time-frequency analysis on the current signal I1, and obtaining a time-frequency spectrum TFS2 of the current signal I1 based on the single-peak ridge R1, and extracting a single-peak ridge R2 of the time-frequency spectrum TFS2, specifically:

[0055] Dividing the current signal I1 into X segments according to a preset segmentation method, and intercepting a data segment from the first data according to a length N, and the starting data is sequentially spaced by M, and when the data length is insufficient to guarantee N, the segmentation is terminated, to obtain X data segments, i.e., X current signal segments IN1; simply speaking, assuming that the total length of the current signal I1 is 20 current data, the length of each data segment is 5, and the starting data of each data segment is spaced by 2, then the first data segment is 1-5, the second data segment is 3-7, and the third data segment is 5-9.

[0056] Performing windowing processing on the X current signal segments IN1 to obtain X time domain signals IN2, and the calculation formula is as follows:

[0057] ; that is, 1-5 for data segment 2, 3-7 for data segment 3, 5-9 for data segment 4, and so on.

[0058] The X time-domain signals IN2 are obtained by windowing the X current signals IN1, and the calculation formula is as follows:

[0059] wherein,

[0060] ;

[0061] In the formula, [index of current sample], n is the window length. N

[0062] The three-dimensional spectrum SD1 ([frequency, frequency, amplitude]) is obtained by performing bispectrum analysis on the X time-domain signals IN2, and the calculation formula is as follows:

[0063] ;

[0064] wherein,

[0065] ;

[0066] In the formula, SD1 ( f 1, f 2) represents the amplitude corresponding to the frequency coordinates ( f 1, f 2); X ( f 2) represents the amplitude corresponding to the frequency f 2 of the Fourier transform of the time-domain signal IN2; X ( f 2+ f 1) represents the amplitude corresponding to the frequency f 2+ f 1 of the Fourier transform of the time-domain signal IN2; X ( f 2- f 1) represents the amplitude corresponding to the frequency f 2- f 1 of the Fourier transform of the time-domain signal IN2; E {.} represents mathematical expectation.

[0067] ​According to the middle time TN2 of each time domain signal IN2 and the frequency extreme FN2 corresponding to the unimodal ridge line R1 ([time, frequency extreme, amplitude]), the corresponding three-dimensional spectrum diagram SD1 ([frequency, frequency, amplitude]) is sliced along the frequency extreme FN2 to obtain X two-dimensional spectrum diagrams TDS1 ([frequency, amplitude]).

[0068] The X two-dimensional spectrum diagrams TDS1 are arranged according to the order of the middle times to obtain a time-frequency spectrum diagram TFS2 ([time, frequency, amplitude]) of the current signal I1.

[0069] The unimodal ridge line R2 ([time, frequency value, maximum amplitude]) of the time-frequency spectrum diagram TFS2 ([time, frequency, amplitude]) is extracted, that is, the frequency corresponding to the maximum amplitude of each time is obtained.

[0070] The phase average power PI is obtained according to the current signal I1 and the voltage signal U1 and the unimodal ridge line R1, and specifically:

[0071] The time corresponding to a point It in the current signal I1 is Tt, and according to the unimodal ridge line R1 ([time, frequency value, maximum amplitude]), the frequency Pt corresponding to Tt is obtained.

[0072] The phase average power PI is obtained through the frequency Pt, the current signal I1 and the voltage signal U1, and the calculation formula is as follows:

[0073] ;

[0074] In the formula, .

[0075] The threshold value H1 of the characteristic frequency for judging whether a fault occurs is determined according to the ratio of the phase average power PI to the rated power PR of the alternating current motor, and specifically:

[0076] The rated power PR is obtained through the parameters of the alternating current motor, an initial threshold value H0 is set, and the threshold value H1 of the characteristic frequency for judging whether a fault occurs is obtained according to the ratio of the phase average power PI to the rated power PR and the initial threshold value H0, and the calculation formula is as follows:

[0077] ;

[0078] In the formula, the rated power PR is obtained from the parameters of the alternating current motor, and H0 is the initial threshold value.

[0079] The fault characteristic frequency amplitude HF is obtained according to the current signal I1 and the unimodal ridge line R2, and specifically:

[0080] The time corresponding to a point It in the current signal I1 is Tt, and the fault characteristic frequency amplitude HF corresponding to Tt is obtained according to the unimodal ridge line R2 ([time, frequency value, maximum amplitude]).

[0081] The ratio of the fault characteristic frequency amplitude HF to the characteristic frequency threshold H1 is calculated, and the severity of the fault is determined according to Table 1.

[0082] Table 1 Fault Severity Table

[0083] HF to H1 ratio Motor rotor condition Recommendation ≤0.6 Excellent None 0.6~0.85 Good None 0.85~1 Risk exists Monitor trend 1~1.75 1-2 broken bars Shorten monitoring interval, observe trend ≥1.75 Multiple broken bars Repair or replace as soon as possible . Embodiment 2

[0084] As Figure 2 shown, the alternating current motor fault diagnosis device based on modulation bispectrum includes:

[0085] A data acquisition module 1 for real-time acquisition of phase current I0 and phase voltage U0 of an alternating current motor, and pre-processing of the phase current I0 and phase voltage U0 to obtain current signal I1 and voltage signal U1;

[0086] A unimodal ridge line R1 extraction module 2 for performing short-time Fourier transform on the current signal I1 to obtain a time-frequency spectrum TFS1 of the current signal I1, and extracting a unimodal ridge line R1 of the time-frequency spectrum TFS1;

[0087] A unimodal ridge line R2 extraction module 3 for performing modulation bispectrum-based time-frequency analysis on the current signal I1, and obtaining a time-frequency spectrum TFS2 of the current signal I1 based on the unimodal ridge line R1, and extracting a unimodal ridge line R2 of the time-frequency spectrum TFS2;

[0088] A phase average power acquisition module 4 for acquiring phase average power PI according to the current signal I1 and voltage signal U1 and unimodal ridge line R1;

[0089] A characteristic frequency threshold acquisition module 5 for determining a characteristic frequency threshold H1 for judging whether a fault occurs according to the ratio of the phase average power PI to the rated power PR of the alternating current motor;

[0090] A characteristic frequency amplitude acquisition module 6 for acquiring a fault characteristic frequency amplitude HF according to the current signal I1 and unimodal ridge line R2;

[0091] A fault judgment module 7 for judging whether the alternating current motor has a motor broken bar fault and the severity of the fault by the ratio of the fault characteristic frequency amplitude HF to the characteristic frequency threshold H1.

[0092] Embodiment 3

[0093] The embodiment provides a storage medium, which stores a computer program, and the computer program is executed by a host to realize each step in the method.

[0094] It can be understood that the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk.

[0095] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A fault diagnosis method for AC motors based on modulated bispectral modulation, characterized in that, Includes the following steps: The phase current I0 and phase voltage U0 of the AC motor are acquired in real time, and the phase current I0 and phase voltage U0 are preprocessed to obtain the current signal I1 and the voltage signal U1. Perform a short-time Fourier transform on the current signal I1 to obtain the time spectrum diagram TFS1 of the current signal I1, and extract the single-peak ridge line R1 of the time spectrum diagram TFS1. Time-frequency analysis based on modulation bispectrum is performed on the current signal I1, and the time spectrum diagram TFS2 of the current signal I1 is obtained based on the single peak ridge line R1. The single peak ridge line R2 of the time spectrum diagram TFS2 is extracted. The phase average power PI is obtained based on the current signal I1, the voltage signal U1, and the single-peak ridge line R1. The characteristic frequency threshold H1 for determining whether a fault has occurred is determined based on the ratio of the phase average power PI to the rated power PR of the AC motor. Specifically, the rated power PR is obtained through the AC motor's own parameters, an initial threshold H0 is set, and the characteristic frequency threshold H1 for determining whether a fault has occurred is obtained based on the ratio of the phase average power PI and the rated power PR, as well as the initial threshold H0. The calculation formula is as follows: ； In the formula, the rated power PR is obtained from the AC motor's own parameters, and H0 is the initial threshold. The fault characteristic frequency amplitude HF is obtained based on the current signal I1 and the single-peak ridge line R2. The ratio of the fault characteristic frequency amplitude HF to the characteristic frequency threshold H1 is used to determine whether the AC motor has a broken bar fault and the severity of the fault.

2. The AC motor fault diagnosis method based on modulation bispectrum according to claim 1, characterized in that, The step of performing time-frequency analysis based on modulation bispectrum on the current signal I1, obtaining the time-frequency spectrum TFS2 of the current signal I1 based on the single-peak ridge R1, and extracting the single-peak ridge R2 of the time-frequency spectrum TFS2 is specifically as follows: The current signal I1 is divided into X segments according to a preset segmentation method to obtain X segments of current signal IN1; Windowing is applied to the X segments of current signal IN1 to obtain X time-domain signals IN2; Modulation bispectral analysis was performed on X time-domain signals IN2 to obtain X three-dimensional spectra SD1; Based on the midpoint TN2 of X time-domain signals IN2 and the frequency extreme value FN2 corresponding to the single-peak ridge line R1, the corresponding three-dimensional spectrum SD1 is sliced ​​along the frequency extreme value FN2 to obtain X two-dimensional spectrum TDS1. Arrange the X two-dimensional spectrum diagrams TDS1 according to the order of their intermediate times to obtain the time spectrum diagram TFS2 of the current signal I1; Extract the single-peak ridge R2 from the time-frequency spectrum TFS2 of the current signal I1.

3. The AC motor fault diagnosis method based on modulation bispectrum according to claim 2, characterized in that, The process of dividing the current signal I1 into X segments according to a preset segmentation method is as follows: The current signal I1 is segmented into data segments of length N, starting from the first data. The starting data of each segment is spaced M data points apart. When the remaining data length is less than N, the segmentation is stopped, resulting in X segments of current signal.

4. The AC motor fault diagnosis method based on modulation bispectrum according to claim 3, characterized in that, The step of obtaining the phase average power PI based on the current signal I1, the voltage signal U1, and the single-peak ridge R1 is specifically as follows: The time corresponding to a certain point It in the current signal I1 is Tt. The frequency Pt corresponding to Tt is obtained according to the single-peak ridge line R1. The phase average power PI is obtained by using the frequency Pt, the current signal I1, and the voltage signal U1.

5. The AC motor fault diagnosis method based on modulated bispectrum according to claim 1, characterized in that, The process of obtaining the fault characteristic frequency amplitude HF based on the current signal I1 and the single-peak ridge line R2 is specifically as follows: The time corresponding to any point It in the current signal I1 is Tt. The fault characteristic frequency amplitude HF corresponding to Tt is obtained according to the single-peak ridge line R2.

6. A fault diagnosis device for AC motors based on modulated bispectral modulation, characterized in that, include: Data acquisition module: used to acquire the phase current I0 and phase voltage U0 of AC motor in real time, and preprocess the phase current I0 and phase voltage U0 to obtain current signal I1 and phase voltage U1; Single-peak ridge R1 extraction module: used to perform short-time Fourier transform on the current signal I1 to obtain the time spectrum diagram TFS1 of the current signal I1, and extract the single-peak ridge R1 of the time spectrum diagram TFS1; Single-peak ridge R2 extraction module: used to perform time-frequency analysis based on modulation bispectrum on the current signal I1, and obtain the time spectrum diagram TFS2 of the current signal I1 based on the single-peak ridge R1, and extract the single-peak ridge R2 of the time spectrum diagram TFS2; Phase average power acquisition module: used to acquire phase average power PI based on the current signal I1, voltage signal U1, and single-peak ridge R1; The characteristic frequency threshold acquisition module is used to determine the characteristic frequency threshold H1 for judging whether a fault has occurred based on the ratio of the phase average power PI and the rated power PR of the AC motor. Specifically, it obtains the rated power PR through the AC motor's own parameters, sets an initial threshold H0, and obtains the characteristic frequency threshold H1 for judging whether a fault has occurred based on the ratio of the phase average power PI and the rated power PR and the initial threshold H0. The calculation formula is as follows: ； In the formula, the rated power PR is obtained from the AC motor's own parameters, and H0 is the initial threshold. Characteristic frequency amplitude acquisition module: used to acquire the fault characteristic frequency amplitude HF based on the current signal I1 and the single-peak ridge R2; Fault diagnosis module: used to determine whether the AC motor has a broken bar fault and the severity of the fault by the ratio of the fault characteristic frequency amplitude HF to the characteristic frequency threshold H1.

7. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by the main controller, implements each step of the AC motor fault diagnosis method based on modulation bispectrum as described in any one of claims 1-5.

Images