Motor noise detection method and system based on waveform analysis

Abstract

The invention discloses a motor noise detection method and system based on waveform analysis, and relates to the related field of motor fault diagnosis, and the method comprises the steps: carrying out the time domain collection of an original noise signal during the operation of a motor, obtaining an original waveform signal, carrying out the Fourier transform of the original waveform signal, extracting the spectrum characteristics, and obtaining the spectrum energy distribution; performing threshold segmentation on the frequency spectrum energy distribution, screening out motors which meet a preset quality standard but have preset abnormal indications, and obtaining critical qualified motors; extracting an original waveform signal of the critically qualified motor, and extracting a transient impact component from the original waveform signal to obtain a transient feature; and matching the transient feature with an early fault mode in a pre-constructed fault feature library, and outputting a potential fault detection result. The technical problem that the current qualified motor with potential faults cannot be identified in the existing motor noise detection is solved, and the technical effect of accurately identifying the qualified motor with the potential faults is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of motor fault diagnosis, in particular to a motor noise detection method and system based on waveform analysis. BACKGROUND

[0002] In the field of motor production and application, motor noise detection is a key link to ensure motor quality and reliability. Among them, it is extremely important to accurately identify the motor that seems to be qualified at the moment of detection, but actually has potential fault hidden dangers, because such motor will often fail in a short period of time after being put into use, which seriously affects the production progress and increases the maintenance cost. At present, the main method to solve the motor fault detection problem is to use the conventional noise detection technology to analyze the overall motor running noise, and to determine whether the motor is qualified according to the established qualified standard. If qualified, it is allowed to leave the factory, and if unqualified, it is returned for treatment. The current method only relies on the conventional single qualified standard for judgment, and lacks in-depth mining and analysis of the subtle characteristics in the motor noise signal that reflect early potential faults, which leads to the inability to effectively distinguish the motor that is qualified at the moment of detection but has potential fault risk, so that this part of the motor flows into the market soon after the fault, causing loss to the user.

[0003] At present, in the related technology, the motor noise detection has the technical problem of being unable to identify the motor that is qualified at the moment but has potential fault. SUMMARY

[0004] The present application provides a motor noise detection method based on waveform analysis, which adopts time domain collection of noise signals during motor operation to obtain original waveform signals, converts time domain signals into frequency domain signals through Fourier transform, extracts frequency spectrum energy distribution characteristics, performs threshold segmentation on the frequency spectrum characteristics, and divides the motor into three categories: completely qualified normal motor, unqualified motor that does not meet the standard, and critical qualified motor that meets the basic standard but has abnormal characteristics. For the critical qualified motor, further extract the transient impact component in the original waveform signal to obtain transient characteristics, match these transient characteristics with known early fault patterns, output the potential fault type if the matching is successful, and mark it as a state that needs to be monitored if there is no matching. The technical means solves the technical problem that the existing motor noise detection cannot identify the motor that is qualified at the moment but has potential fault, and achieves the technical effect of accurately identifying the qualified motor with potential fault.

[0005] The application provides a motor noise detection method based on waveform analysis, comprising: collecting original noise signals in time domain when a motor is running to obtain original waveform signals, performing Fourier transform on the original waveform signals, extracting frequency spectrum features, and obtaining frequency spectrum energy distribution; performing threshold segmentation on the frequency spectrum energy distribution to screen out motors that meet preset quality standards but have preset abnormal indications, and obtaining critical qualified motors; extracting the original waveform signals of the critical qualified motors, extracting transient impact components from the original waveform signals, and obtaining transient features; matching the transient features with early fault patterns in a pre-constructed fault feature library, and outputting potential fault detection results.

[0006] In possible implementation manners, original noise signals when a motor is running are collected in time domain to obtain original waveform signals, Fourier transform is performed on the original waveform signals, frequency spectrum features are extracted, and frequency spectrum energy distribution is obtained, and the following processing is performed: time domain noise signals are collected at a preset sampling frequency by a microphone array at a rated speed of the motor; noise reduction processing is performed on the time domain noise signals to obtain the original waveform signals; fast Fourier transform is performed on the original waveform signals to obtain a frequency spectrum graph; amplitudes of main frequencies, bearing characteristic frequencies and electromagnetic harmonic frequencies in the frequency spectrum graph are calculated to generate the frequency spectrum energy distribution.

[0007] In possible implementation manners, threshold segmentation is performed on the frequency spectrum energy distribution to screen out motors that meet preset quality standards but have preset abnormal indications, and critical qualified motors are obtained, and the following processing is performed: main frequency amplitudes of historical qualified motor samples are dynamically collected, the preset quality standards are determined according to a dynamic threshold interval of the main frequency amplitudes; bearing characteristic frequency amplitudes of the historical qualified motor samples are dynamically collected, and qualified bearing characteristic frequency indications are determined according to a dynamic threshold interval of the bearing characteristic frequency amplitudes; electromagnetic harmonic frequency amplitudes of the historical qualified motor samples are dynamically collected, and qualified electromagnetic harmonic frequency indications are determined according to a dynamic threshold interval of the electromagnetic harmonic frequency amplitudes; amplitudes that do not meet the qualified bearing characteristic frequency indications and the qualified electromagnetic harmonic frequency indications are set as preset abnormal indications; if the main frequency amplitude in the frequency spectrum energy distribution meets the preset quality standards, and the preset abnormal indications exist in the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude, the corresponding motor is marked as a critical qualified motor.

[0008] In possible implementation manners, the following processing is performed: the historical qualified motor samples are motors that are consistent with a motor model to be detected, have been in service, and have not appeared faults in an early fault exposure period.

[0009] In a possible implementation, the transient impact component is extracted from the original waveform signal to obtain a transient feature, and the following processing is performed: wavelet packet decomposition is performed on the original waveform signal to obtain a time-frequency energy distribution matrix; pulse detection is performed on the time-frequency energy distribution matrix to calculate pulse intervals and amplitudes, and the transient feature is obtained.

[0010] In a possible implementation, a fault feature library is constructed in advance, and the following processing is performed: a first historical critical qualified motor sample is dynamically collected, the first historical critical qualified motor sample being a motor of the same type as the motor to be detected, having been in service, and having a first fault mode in the early fault exposure period; a first historical transient feature of the first historical critical qualified motor sample is obtained; a first mapping relationship between the first historical transient feature and the first fault mode is established, and the fault feature library is constructed according to the first mapping relationship.

[0011] In a possible implementation, the fault feature library is constructed in advance, and the following processing is further performed: a second historical critical qualified motor sample is dynamically collected, the second historical critical qualified motor sample being a motor of a different type but the same fault-related parameters as the motor to be detected, having been in service, and having a second fault mode in the early fault exposure period, the fault-related parameters including a bearing type, a rated speed, and a pole number; a second historical transient feature of the second historical critical qualified motor sample is obtained; a first historical qualified motor sample and a second historical qualified motor sample corresponding one-to-one to the first historical critical qualified motor sample and the second historical critical qualified motor sample are collected, a conversion coefficient is calculated according to a frequency offset ratio of the two; the second historical transient feature is converted according to the conversion coefficient to obtain a second converted historical transient feature, a second mapping relationship between the second converted historical transient feature and the second fault mode is established, and the second mapping relationship is added to the fault feature library.

[0012] In a possible implementation, the fault feature library is constructed according to the first mapping relationship, and the following processing is performed: the first mapping relationship is grouped according to different fault modes to obtain a plurality of first mapping relationship groups; all historical transient features are extracted from each first mapping relationship group to form a plurality of historical transient feature libraries; the plurality of historical transient feature libraries are marked with corresponding fault modes to obtain a plurality of early fault modes, and the plurality of early fault modes are integrated to obtain the fault feature library.

[0013] In a possible implementation, the transient feature is matched with an early failure mode in a pre-constructed failure feature library, and a potential failure detection result is output, and the following processing is performed: traversing the failure feature library, extracting a first early failure mode; calculating the similarity between the transient feature and historical transient features contained in the first early failure mode, and if the calculation result is greater than or equal to a preset similarity threshold, the first early failure mode is output as a potential failure detection result.

[0014] The application also provides a motor noise detection system based on waveform analysis, comprising: a spectrum energy distribution acquisition module, configured to collect original noise signals in a time domain when a motor is running to obtain original waveform signals, perform Fourier transform on the original waveform signals, extract spectrum features, and obtain spectrum energy distribution; a critical qualified motor acquisition module, configured to perform threshold segmentation on the spectrum energy distribution, filter out motors that meet preset quality standards but have preset abnormal indications, and obtain critical qualified motors; a transient feature acquisition module, configured to extract the original waveform signals of the critical qualified motors, extract transient impact components from the original waveform signals, and obtain transient features; and a potential failure detection module, configured to match the transient features with early failure modes in a pre-constructed failure feature library, and output potential failure detection results.

[0015] The motor noise detection method and system based on waveform analysis provided in the application first collect original noise signals in a time domain when a motor is running to obtain original waveform signals, perform Fourier transform on the original waveform signals, extract spectrum features, and obtain spectrum energy distribution, then perform threshold segmentation on the spectrum energy distribution, filter out motors that meet preset quality standards but have preset abnormal indications, obtain critical qualified motors, extract the original waveform signals of the critical qualified motors, extract transient impact components from the original waveform signals, obtain transient features, and finally match the transient features with early failure modes in a pre-constructed failure feature library, and output potential failure detection results. The technical effect of accurately identifying qualified motors with potential failures is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings of the embodiments of the application will be briefly introduced below, and the flowcharts are used to illustrate the operations performed by the system according to the embodiments of the application in the application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously according to needs. Meanwhile, other operations can be added to these processes, or one or more steps of operations can be removed from these processes.

[0017] Figure 1A flowchart of a motor noise detection method based on waveform analysis provided by an embodiment of the present application is shown.

[0018] Figure 2 A structural diagram of a motor noise detection system based on waveform analysis provided by an embodiment of the present application is shown.

[0019] Legend: spectrum energy distribution acquisition module 10, critical qualified motor acquisition module 20, transient characteristic acquisition module 30, potential fault detection module 40. DETAILED DESCRIPTION

[0020] The above description is only a summary of the technical solutions of the present application. In order to make the technical means of the present application more clear, the following specific embodiments of the present application are described in accordance with the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0021] In order to make the purposes, technical solutions and advantages of the present application more clear, the following will combine the drawings to make a further detailed description of the present application. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of protection of the present application.

[0022] In the following description, "some embodiments" are described, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict. The term "first\second" is only to distinguish similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0023] The embodiments of the present application provide a motor noise detection method based on waveform analysis, as shown in Figure 1 The method comprises the following steps:

[0024] In step S100, the original noise signal of the motor during operation is collected in time domain to obtain an original waveform signal. Fourier transform is performed on the original waveform signal to extract frequency spectrum characteristics and obtain a spectrum energy distribution.

[0025] Specifically, a high-precision acoustic sensor such as a condenser microphone or an acceleration sensor is placed at a specific position around the motor to collect the original noise signal in time domain at a certain sampling frequency when the motor is running, and an original waveform signal is obtained, which is presented in a digital form, that is, a series of discrete amplitude values and corresponding time point data sets. Among them, time domain acquisition refers to the process of continuously acquiring signal samples on the time axis at a certain frequency, which directly reflects the amplitude of the signal changing with time and can present the overall trend and periodicity of the signal.

[0026] Then the original waveform signal is processed using the Fast Fourier Transform (FFT) algorithm. FFT is a mathematical tool for converting time domain signals to frequency domain signals. By converting it to the frequency domain, the frequency spectrum characteristics are extracted, and the frequency spectrum energy distribution is obtained. The frequency spectrum energy distribution contains information about different frequency components and their corresponding energy sizes.

[0027] For example, assume that the sensor collects a time domain waveform signal when the motor is running, and a total of M data points are collected. The horizontal axis is time t (unit: seconds), and the vertical axis is amplitude A (unit: pascal or other sound pressure unit). After FFT operation, the frequency spectrum diagram is obtained, the horizontal axis becomes frequency f (unit: hertz), and the vertical axis is the amplitude value corresponding to each frequency (reflecting the energy size of the frequency component).

[0028] In one possible implementation, the original noise signal when the motor is running is collected in time domain to obtain an original waveform signal, and the original waveform signal is subjected to Fourier transform to extract frequency spectrum characteristics and obtain a frequency spectrum energy distribution. Step S100 further includes step S110, collecting time domain noise signals at a preset sampling frequency through a microphone array at a rated speed of the motor. Specifically, the microphone array is composed of multiple microphones arranged in a certain geometric shape, such as a linear array, a circular array, etc., which can receive motor noise signals from different directions and positions. The preset sampling frequency is determined according to the frequency range of the motor noise, and follows the Nyquist sampling theorem, with a sampling frequency at least twice the highest frequency of the signal, such as 44.1 kHz or 48 kHz when the highest frequency of the motor noise is 20 kHz. The signals collected by the array are subjected to analog-to-digital conversion (ADC) to obtain digital time domain noise signals.

[0029] For example, a circular array of 8 microphones is used, which are evenly distributed around the motor with a distance of 10 cm between each microphone. The sampling frequency is set to 48 kHz, and the noise of the motor running at the rated speed is collected for 10 seconds to obtain 8 sets of time domain noise signal data, each containing 480,000 sampling points. These data are stored in matrix form.

[0030] Step S120, the time domain noise signal is denoised to obtain the original waveform signal. Specifically, the time domain noise signal is denoised by using spectral subtraction, wavelet threshold denoising and the like. Taking spectral subtraction as an example, the background noise spectrum when the motor is stationary is first obtained, then the time domain noise signal collected when the motor is running is subjected to short-time Fourier transform (STFT) to obtain a frequency spectrum, the amplitude of each frequency component is subtracted from the amplitude of the background noise spectrum (if the result is negative, zero is taken), and inverse short-time Fourier transform (ISTFT) is performed to obtain the original waveform signal after denoising; wavelet threshold denoising is to decompose the time domain signal by wavelet, quantize the wavelet coefficients by soft threshold or hard threshold, remove the small coefficients representing noise, and then reconstruct the denoised signal. The original waveform signal reduces the interference of environmental noise and can more purely reflect the noise characteristics of the motor itself.

[0031] Step S130, performing fast Fourier transform on the original waveform signal to obtain a frequency spectrum. Specifically, the FFT algorithm can quickly convert the time domain signal into the frequency domain signal, the number of points of FFT is determined according to the sampling frequency and the number of data points of the original waveform signal, and the power of 2 can be selected to improve the calculation efficiency. After FFT operation, the frequency spectrum in complex number form is obtained, and the amplitude (module) part is taken to obtain the frequency spectrum, the horizontal axis is the frequency, from 0 to half of the sampling frequency (according to the sampling theorem), and the vertical axis is the amplitude of each frequency component, reflecting the energy size of the frequency component in the signal.

[0032] Step S140, calculating the amplitudes of the main frequency, bearing characteristic frequency and electromagnetic harmonic frequency in the frequency spectrum to generate the frequency spectrum energy distribution. Specifically, the main frequency corresponds to the rotating speed frequency of the motor, which can be calculated by the number of poles, rated rotating speed and the like of the motor; the bearing characteristic frequency includes the characteristic frequencies of the inner ring, outer ring and rolling body, and its calculation formula is related to the bearing model, motor rotating speed and the like; the electromagnetic harmonic frequency is related to the number of poles, winding structure and the like of the motor. By finding these specific frequency points in the frequency spectrum, the amplitudes are read and sorted into a frequency spectrum energy distribution data structure, for example, stored in the form of a dictionary, with the frequency name (main frequency, bearing characteristic frequency and the like) as the key and the corresponding amplitude as the value.

[0033] This implementation adopts a microphone array to collect noise signals, can more comprehensively receive noise from different directions and positions of the motor, increase the coverage of the signal, improve the sampling accuracy, effectively remove background environmental noise and some random noise interference through noise reduction processing, improve the signal-to-noise ratio, and make the original waveform signal more clearly present the characteristics of the motor noise. The FFT is used to quickly obtain the frequency spectrum, and the amplitude of the key frequency components such as the main frequency, bearing characteristic frequency and electromagnetic harmonic frequency is calculated. The frequency spectrum characteristics corresponding to different components of the motor (such as the motor rotor, bearing, stator, etc.) can be accurately extracted, the running state of each key component in the motor is monitored, and potential faults of the motor, such as bearing wear and electromagnetic imbalance, can be found early.

[0034] In step S200, the frequency spectrum energy distribution is threshold segmented to screen out the motor satisfying the preset quality standard but having the preset abnormal indication, and a critical qualified motor is obtained.

[0035] Specifically, threshold segmentation refers to a processing method of setting one or more threshold values to divide data samples into different categories or intervals, which is used to distinguish the normal and abnormal states of signals in signal processing. According to the pre-set quality standard, that is, the determined energy threshold range of each frequency band, the frequency spectrum energy distribution is compared and judged point by point or in sections. With the help of conditional judgment statements and array operation functions in programming languages, the motor satisfying the preset quality standard (the spectrum energy is in the qualified range) but having the preset abnormal indication (such as small energy fluctuations of certain specific frequencies) is screened out, so as to determine the critical qualified motor. The critical qualified motor refers to the motor whose basic running state is normal but contains early fault signs and is in the edge state of qualified and unqualified. Such a motor may deteriorate and fail later, and is the motor that needs to be focused on and monitored.

[0036] In a possible implementation, the spectrum energy distribution is threshold segmented, and a motor that meets a preset quality standard but has a preset abnormal indication is screened out to obtain a critical qualified motor. Step S200 further includes step S210 of dynamically collecting a main frequency amplitude of a historical qualified motor sample, determining the preset quality standard according to a dynamic threshold interval of the main frequency amplitude, and the historical qualified motor sample being a motor that is consistent with a to-be-detected motor model, has been in service, and has no failure in an early failure exposure period. Specifically, through a data acquisition system, a main frequency amplitude data of a motor that is consistent with a to-be-detected motor model (is completely same as the to-be-detected motor in design, structure, performance parameter, and the like), has been in service (has run in an actual working condition, and has accumulated real operation data and failure information), and has no failure in an early failure exposure period is acquired in real time. The early failure exposure period refers to a specific time period (for example, 6 months) after the motor starts to run. If the motor has no failure in the period, the motor is a qualified motor. If the motor has a failure in the period, the motor is a critical qualified motor. These motors are historical qualified samples, and the data of the motors are stored in a database. In order to maintain the timeliness and representativeness of the samples, a sliding window mechanism is used, and a fixed sample capacity, for example, 1000 samples, is set. When new qualified motor sample data enters, the earliest collected old sample is removed from the window according to the time sequence, so that the samples are always the latest and representative data set. According to the main frequency amplitude data, statistical characteristics (for example, mean, standard deviation, and the like) are calculated, and a dynamic threshold interval of the main frequency amplitude, that is, the preset quality standard, is determined. For example, the dynamic threshold interval can be set as mean ± 3 times standard deviation, so as to cover the main frequency amplitude range of most normal running motors.

[0037] For example, assuming that the main frequency amplitude data of the historical qualified motor sample is statistically 80 dB and the standard deviation is 2 dB. The dynamic threshold interval is set as 80 dB ± 3 × 2 dB, that is, 74 dB-86 dB. Only the motor with the main frequency amplitude falling in the interval is regarded as meeting the preset quality standard.

[0038] Step S220, dynamically collect the bearing characteristic frequency amplitude of the historical qualified motor sample, and determine the qualified bearing characteristic frequency indication according to the dynamic threshold interval of the bearing characteristic frequency amplitude. Specifically, the bearing characteristic frequency amplitude of the historical qualified motor sample is also acquired by using the dynamic collection method. The sample number is kept constant and updated constantly by using the sliding window. According to the collected data, the statistical characteristics are calculated, the dynamic threshold interval of the bearing characteristic frequency amplitude is determined, and the qualified bearing characteristic frequency indication is determined. For example, by using the same mean ± 3 times standard deviation method, if the mean is 60 dB and the standard deviation is 1.5 dB, the dynamic threshold interval of the qualified bearing characteristic frequency amplitude is 60 dB ± 3 × 1.5 dB, that is, 55.5 dB-64.5 dB, and the bearing frequency characteristic with the amplitude in this interval is normal.

[0039] Step S230, dynamically collect the electromagnetic harmonic frequency amplitude of the historical qualified motor sample, and determine the qualified electromagnetic harmonic frequency indication according to the dynamic threshold interval of the electromagnetic harmonic frequency amplitude. Specifically, similar to steps S210 and S220, the electromagnetic harmonic frequency amplitude of the historical qualified motor sample is dynamically collected, and the dynamic update of the sample is maintained. The dynamic threshold interval of the electromagnetic harmonic frequency amplitude is determined based on the statistical characteristics, and is used as the qualified electromagnetic harmonic frequency indication.

[0040] Step S240, set the amplitudes not meeting the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication as preset abnormal indications. Specifically, the amplitudes not meeting the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication, that is, the amplitudes exceeding the range of the corresponding dynamic threshold interval, are defined as the preset abnormal indications. The program code can be written to set the condition judgment rule, and the collected frequency spectrum energy distribution data is compared one by one to filter out the characteristic amplitudes not meeting the normal range.

[0041] Step S250, if the main frequency amplitude in the frequency spectrum energy distribution meets the preset quality standard, and the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude has the preset abnormal indication, the corresponding motor is marked as a critical qualified motor. Specifically, the frequency spectrum energy distribution is comprehensively judged. The program logic is written to first check whether the main frequency amplitude meets the preset quality standard (in the dynamic threshold interval), and then check whether the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude has the preset abnormal indication (exceeds the corresponding dynamic threshold interval). If both conditions are met, that is, the main frequency is normal but the bearing or electromagnetic harmonic frequency is abnormal, the corresponding motor is marked as a critical qualified motor. The marking can be performed by adding a specific label in the database or setting a corresponding identification bit in the data structure.

[0042] This implementation can timely reflect the normal characteristics of the motor changing over time in the actual running environment by dynamically collecting historical qualified motor sample data and continuously updating the sample set, making the determination of the preset quality standard and qualified indication more accurate. The dynamic threshold interval is determined based on statistical characteristics, which follows the data distribution law, can better contain the reasonable fluctuations in normal operation, and avoids misjudging normal motors as abnormal due to unreasonable fixed threshold setting. By comprehensively considering the characteristics of multiple key frequency components such as main frequency, bearing characteristic frequency and electromagnetic harmonic frequency, the critical qualified motor that seems qualified but may have early signs of failure can be effectively screened out.

[0043] Step S300, extracting the original waveform signal of the critical qualified motor, extracting the transient impact component from the original waveform signal to obtain the transient feature.

[0044] Specifically, for the original waveform signal of the screened critical qualified motor, the wavelet transform method is used to extract the transient impact component. The transient impact component is a part of the signal that changes temporarily and sharply, which is caused by mechanical failure (such as collision, sudden friction, etc.), and is characterized by sharp peaks or pulses in the waveform. Wavelet transform has good time-frequency localization characteristics and can focus on the transient change part of the signal. By selecting appropriate wavelet basis functions (such as db4, sym5, etc.) and decomposition scales, the original waveform signal is decomposed at multiple scales to separate the detail signal part containing the transient impact feature, and then the transient feature is obtained. These features include the time point, amplitude, duration, and other key parameters of the impact.

[0045] For example, the original waveform signal is decomposed by 3 layers using the db4 wavelet basis function to obtain the detail coefficients at different scales, which correspond to the transient impact information in different frequency bands. By reconstructing the detail signal, the transient impact component waveform is obtained, and the information such as the time point and size of the impact peak is counted as the transient feature.

[0046] In one possible implementation, the transient impact component is extracted from the original waveform signal to obtain the transient feature, and step S300 further includes step S310 of performing wavelet packet decomposition on the original waveform signal to obtain a time-frequency energy distribution matrix. Specifically, wavelet packet decomposition is a multi-resolution analysis method that can decompose a signal into sub-band signals of different frequency bands. Specifically, by selecting appropriate wavelet basis functions (such as db4, sym5, etc.) and decomposition levels (such as 3 layers), the original waveform signal is decomposed at multiple scales to obtain a time-frequency energy distribution matrix. The time-frequency energy distribution matrix contains energy information at different time points and different frequency bands.

[0047] For example, assuming that the sampling frequency of the original waveform signal is 48 kHz and the data length is 1024 points. A db4 wavelet base function is selected for 3-layer wavelet packet decomposition, and 8 sub-band signals (frequency bands) are obtained. Each sub-band signal corresponds to a different frequency band, and the energy value of each sub-band signal is calculated to form a time-frequency energy distribution matrix. The rows of the matrix represent time points, the columns represent frequency bands, and the matrix elements represent the energy values of the corresponding time points and frequency bands.

[0048] At step S320, pulse detection is performed on the time-frequency energy distribution matrix, and the pulse interval and amplitude are calculated to obtain the transient feature. Specifically, pulse detection is used to identify transient impact components in the signal and calculate the pulse interval and amplitude. Specifically, an energy threshold is set according to the background noise energy level, and energy values higher than the threshold are considered as potential transient impact components. The time-frequency energy distribution matrix is traversed, and when the energy value of a certain time point and frequency band exceeds the set threshold, it is marked as a pulse. For each marked pulse, its amplitude (energy value) and interval (time difference between adjacent pulses) are calculated.

[0049] For example, in the time-frequency energy distribution matrix, the energy threshold is set to be 3 times the standard deviation of the background noise energy. When traversing the matrix, it is found that the energy value of a certain time point in the high frequency band exceeds the threshold, and it is marked as a pulse. The amplitude and interval of the pulse are recorded. Repeat this process to obtain the amplitudes and intervals of all pulses to form a transient feature vector.

[0050] This implementation uses wavelet packet decomposition, which can decompose the signal into multiple sub-band signals of different frequency bands, capture the subtle changes of the signal at different time points and frequency bands, and more accurately extract the transient impact components than traditional Fourier transform methods.

[0051] At step S400, the transient feature is matched with early fault patterns in a pre-constructed fault feature library, and a potential fault detection result is output.

[0052] Specifically, a fault feature library is established, and the transient features of a large number of known fault types of the motor are analyzed and classified in advance to form standard feature templates of different early fault patterns. Among them, the early fault pattern is a relatively hidden and unobvious state pattern of the fault in the initial stage, which contains the typical signs and characteristic performances of the early stage of a specific fault type. When detecting, the extracted transient feature vector (such as a vector containing impact peak value, interval time, energy distribution, etc.) is matched with each template in the fault feature library one by one for similarity measurement, and similarity measurement indexes such as Euclidean distance and correlation coefficient are used. When the similarity exceeds the set threshold, it is determined that the matching is successful, and the corresponding potential fault type is output; if no template with high enough similarity is found, it is marked for monitoring.

[0053] In a possible implementation, the fault feature library is constructed in advance, and step S400 further includes step S410 of dynamically collecting a first historical critical qualified motor sample, the first historical critical qualified motor sample being a motor consistent with the motor model to be detected, having served and having appeared a first fault mode in the early fault exposure period. Specifically, a motor consistent with the motor model to be detected, having served and having appeared a first fault mode in the early fault exposure period is dynamically collected as a sample (that is, the first historical critical qualified motor sample). Here, the dynamic collection is also implemented by using the sliding window mechanism, the sample capacity is kept unchanged, and when a new sample comes in, a corresponding number of old samples are removed in chronological order, to ensure the timeliness and representativeness of the sample. The first fault mode refers to a specific fault type of the motor in the early fault exposure period, such as bearing wear.

[0054] In step S420, a first historical transient feature of the first historical critical qualified motor sample is acquired. Specifically, the original waveform signal of the collected first historical critical qualified motor sample is processed by using the method in step S300 (such as the wavelet packet decomposition technique), to extract the transient feature. That is, the original waveform signal is subjected to wavelet packet decomposition to obtain a time-frequency energy distribution matrix, and then the transient features such as pulse interval and amplitude are calculated through pulse detection. These transient features are the characteristic performances corresponding to the first fault mode of the motor, and can reflect specific physical phenomena of the fault. For example, bearing wear can cause periodic impact pulses, and the corresponding transient features are specific pulse intervals and amplitudes.

[0055] In step S430, a first mapping relationship between the first historical transient feature and the first fault mode is established, and the fault feature library is constructed according to the first mapping relationship. Specifically, the acquired first historical transient feature and the corresponding first fault mode are associated by using a data structure such as a dictionary or a database table, to establish the first mapping relationship, in which the key is the transient feature (such as a combination of parameters such as pulse interval and amplitude), and the value is the corresponding fault mode. Based on the above mapping relationship, the fault feature library is constructed. The library is dynamically updated, and with the continuous addition of new samples and the removal of old samples, the latest and most representative fault feature information is reflected in time. Each entry in the fault feature library contains a specific fault mode and the corresponding transient feature, to provide a benchmark and a reference for fault detection.

[0056] The implementation has strong pertinence and practicality by using the first historical critical qualified motor sample consistent with the motor model to be detected and having faults in the early fault exposure period. The mapping relationship based on actual fault data can more accurately identify the potential fault type and reduce misjudgment.

[0057] In a possible implementation, the fault feature library is constructed in advance, and step S400 further includes step S440 of dynamically collecting a second historical critical qualified motor sample. The second historical critical qualified motor sample is a motor inconsistent with the motor model to be detected but consistent in fault-related parameters, having served and having a second fault mode in the early fault exposure period. The fault-related parameters include bearing model, rated speed and pole number. Specifically, the second historical critical qualified motor sample set is continuously updated through a sliding window mechanism. The second historical critical qualified motor sample is different from the motor model to be detected, but the bearing model, rated speed and pole number of the sample motor are the same as those of the motor to be detected. Although the second historical critical qualified motor sample is different from the overall design of the motor to be detected, the key components (such as bearings) and basic operating parameters (such as rated speed and pole number) are consistent with the motor to be detected, and a specific fault mode occurs in the early fault exposure period. The second historical critical qualified motor sample is introduced to supplement the fault feature library. Since the fault-related parameters of the second historical critical qualified motor sample are consistent with those of the motor to be detected, the fault features of the second historical critical qualified motor sample can be used for fault detection of the motor to be detected after conversion, and the first historical critical qualified motor sample is complementary.

[0058] In step S450, a second historical transient feature of the second historical critical qualified motor sample is acquired. Specifically, the original waveform signal of the collected second historical critical qualified motor sample is processed by using the method in step S300 (such as wavelet packet decomposition and other technical means) to extract the transient feature. The transient feature reflects the characteristic performance of the sample motor when the second fault mode occurs.

[0059] At step S460, the first historical qualified motor sample and the second historical qualified motor sample corresponding to the first historical critical qualified motor sample and the second historical critical qualified motor sample are collected, and a conversion coefficient is calculated according to the frequency offset ratio of the two. Specifically, the first historical qualified motor sample and the second historical qualified motor sample corresponding to the first historical critical qualified motor sample and the second historical critical qualified motor sample are collected, that is, the first historical qualified motor sample and the first historical critical qualified motor sample are of the same type, and the second historical qualified motor sample and the second historical critical qualified motor sample are of the same type. Moreover, these qualified motor samples do not appear to have failed in the early failure exposure period. According to the frequency offset ratio of the first historical qualified motor sample and the second historical qualified motor sample, the conversion coefficient is calculated. The frequency offset ratio refers to the difference ratio between the characteristic frequencies (such as the main frequency, bearing characteristic frequency, etc.) of the motor to be detected and the second historical qualified motor sample under the same operating conditions, which reflects the frequency difference caused by different motor types. The frequency offset ratio is calculated by comparing the characteristic frequencies of the first historical qualified motor sample and the second historical qualified motor sample. The conversion coefficient is the frequency offset ratio, which is used to convert the time parameter of the second historical transient feature from the frequency range of the second historical qualified motor sample to the frequency range of the motor to be detected.

[0060] At step S470, the second historical transient feature is converted according to the conversion coefficient to obtain a second converted historical transient feature, a second mapping relationship between the second converted historical transient feature and the second failure mode is established, and the second mapping relationship is added to the failure feature library. Specifically, the second historical transient feature is adjusted by using the conversion coefficient to obtain the second converted historical transient feature, so as to ensure that the second historical transient feature is consistent with the frequency range of the motor to be detected, and the feature has comparability. Through a data structure (such as a dictionary or a database table), the second converted historical transient feature is associated with the second failure mode to establish the second mapping relationship, wherein the key is the converted transient feature and the value is the corresponding failure mode. The second mapping relationship is added to the failure feature library. In this way, the failure feature library contains failure features from different types of motors (but with the same failure associated parameters), which enriches the content of the feature library.

[0061] This implementation reduces the dependence on samples completely consistent with the type of the motor to be detected by introducing motor samples consistent with the failure associated parameters of the motor to be detected but of different types. The transient features of different types of motors are adjusted by using the conversion coefficient, so that the failure feature library can still perform effective failure detection in the case of lack of sufficient samples of the same type.

[0062] In a possible implementation, the step S430 of constructing the fault feature library according to the first mapping relationship further includes the step S431 of grouping the first mapping relationship according to different fault modes to obtain a plurality of first mapping relationship groups. Specifically, the first mapping relationship is grouped according to different fault modes, for example, by traversing the first mapping relationship set, and the mapping relationship with the same fault mode is divided into the same group according to the identification (such as the fault mode name or number) of the fault mode. The mapping relationship in each group corresponds to the same fault mode, and the fault modes are different between groups, thereby obtaining a plurality of first mapping relationship groups.

[0063] The step S432 extracts all historical transient features from each first mapping relationship group to establish a plurality of historical transient feature libraries. Specifically, each first mapping relationship group contains a plurality of mapping relationships with the same fault mode, and each mapping relationship is associated with a specific historical transient feature. These features are collected to form a historical transient feature library corresponding to a specific fault mode. If there are n first mapping relationship groups, n historical transient feature libraries can be obtained.

[0064] The step S433 labels the plurality of historical transient feature libraries according to corresponding fault modes to obtain a plurality of early fault modes, and integrates the plurality of early fault modes to obtain the fault feature library. Specifically, a corresponding fault mode label is added to each historical transient feature library, which explicitly indicates the specific fault type associated with the feature library. Then, the historical transient feature libraries with labels are integrated together to construct the fault feature library. This library classifies different fault modes according to the fault mode as the classification basis, and systematically organizes and stores the transient features corresponding to different fault modes.

[0065] This implementation reduces the number of samples that need to be compared one by one during fault detection by combining samples with the same fault mode. By using the grouped feature library, only a few fault modes need to be compared during detection, which speeds up the detection and improves the detection efficiency.

[0066] In a possible implementation, the step S400 further includes the step S480 of traversing the fault feature library to extract a first early fault mode. Specifically, each early fault mode in the fault feature library is accessed in sequence. The fault feature library contains a plurality of early fault modes, and each mode is associated with a corresponding historical transient feature library. During the traversal process, the first early fault mode and its corresponding historical transient feature library are first extracted. For example, the first early fault mode in the fault feature library may be a bearing wear fault mode, and the corresponding historical transient feature library contains a plurality of transient features related to bearing wear.

[0067] Step S490: Calculate the similarity between the transient feature and the historical transient features contained in the first early fault mode. If the calculation result is greater than or equal to a preset similarity threshold, the first early fault mode is output as a potential fault detection result. Specifically, the extracted transient features are compared with the historical transient features contained in the first early fault mode. The similarity calculation method can use Euclidean distance, correlation coefficient, etc. A preset similarity threshold (e.g., 0.8) is set. If the calculation result is greater than or equal to the threshold, the transient feature is considered to match the early fault mode. The matched early fault mode is output as a potential fault detection result. If the similarity is lower than the threshold, the next early fault mode is traversed until a matching mode is found or all modes are traversed.

[0068] This implementation reduces the complexity of each calculation and improves overall detection efficiency by extracting and matching early failure modes one by one. Using similarity calculation to quantify and match transient features allows for an objective assessment of the similarity between features, effectively reducing false positives and improving detection accuracy.

[0069] This application employs time-domain acquisition of noise signals during motor operation to obtain the original waveform signal. Fourier transform is used to convert the time-domain signal into a frequency-domain signal, extracting spectral energy distribution features. Threshold segmentation is then applied to these spectral features to classify motors into three categories: fully qualified normal motors, substandard unqualified motors, and critically qualified motors that meet basic standards but exhibit abnormal characteristics. For critically qualified motors, transient impact components are further extracted from their original waveform signals to obtain transient features. These transient features are then matched with known early fault modes. If a match is successful, the potential fault type is output; otherwise, it is marked as a state requiring monitoring. This approach solves the technical problem of existing motor noise detection methods that cannot identify motors that are currently qualified but have potential faults, achieving the technical effect of accurately identifying qualified motors with potential faults.

[0070] In the above text, refer to Figure 1 A method for detecting motor noise based on waveform analysis according to an embodiment of the present invention has been described in detail. Next, reference will be made to... Figure 2 A waveform analysis-based motor noise detection system according to an embodiment of the present invention is described.

[0071] The motor noise detection system based on waveform analysis according to the embodiment of the present application is used to solve the technical problem that the existing motor noise detection cannot identify the motor with potential faults, and achieves the technical effect of accurately identifying the qualified motor with potential faults. The motor noise detection system based on waveform analysis comprises a spectrum energy distribution acquisition module 10, a critical qualified motor acquisition module 20, a transient feature acquisition module 30, and a potential fault detection module 40.

[0072] The spectrum energy distribution acquisition module 10 is configured to collect the original noise signal of the motor in time domain to obtain an original waveform signal, perform Fourier transform on the original waveform signal to extract a frequency spectrum feature, and obtain a spectrum energy distribution. The critical qualified motor acquisition module 20 is configured to perform threshold segmentation on the spectrum energy distribution to screen out the motor that meets the preset quality standard but has a preset abnormal indication, and obtain a critical qualified motor. The transient feature acquisition module 30 is configured to extract the original waveform signal of the critical qualified motor, extract a transient impact component from the original waveform signal, and obtain a transient feature. The potential fault detection module 40 is configured to match the transient feature with an early fault mode in a pre-constructed fault feature library, and output a potential fault detection result.

[0073] In the following, the specific configuration of the spectrum energy distribution acquisition module 10 will be described in detail. As described above, the original noise signal of the motor is collected in time domain to obtain an original waveform signal, Fourier transform is performed on the original waveform signal to extract a frequency spectrum feature, and a spectrum energy distribution is obtained. The spectrum energy distribution acquisition module 10 can further comprise a time domain noise signal acquisition unit configured to acquire a time domain noise signal at a rated speed of the motor by a microphone array at a preset sampling frequency; a noise reduction processing unit configured to perform noise reduction processing on the time domain noise signal to obtain the original waveform signal; a fast Fourier transform unit configured to perform fast Fourier transform on the original waveform signal to obtain a frequency spectrum diagram; and a spectrum energy distribution generation unit configured to calculate the amplitude of the main frequency, bearing characteristic frequency and electromagnetic harmonic frequency in the frequency spectrum diagram, and generate the spectrum energy distribution.

[0074] In the following, the specific configuration of the critical qualified motor obtaining module 20 will be described in detail. As described above, the threshold segmentation is performed on the spectrum energy distribution to screen out the motors that meet the preset quality standard but have the preset abnormal indication, to obtain the critical qualified motor. The critical qualified motor obtaining module 20 can further include: a preset quality standard determining unit configured to dynamically collect the main frequency amplitude of the historical qualified motor sample, and determine the preset quality standard according to the dynamic threshold interval of the main frequency amplitude; a qualified bearing characteristic frequency indication determining unit configured to dynamically collect the bearing characteristic frequency amplitude of the historical qualified motor sample, and determine the qualified bearing characteristic frequency indication according to the dynamic threshold interval of the bearing characteristic frequency amplitude; a qualified electromagnetic harmonic frequency indication determining unit configured to dynamically collect the electromagnetic harmonic frequency amplitude of the historical qualified motor sample, and determine the qualified electromagnetic harmonic frequency indication according to the dynamic threshold interval of the electromagnetic harmonic frequency amplitude; a preset abnormal indication setting unit configured to set the amplitudes that do not meet the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication as the preset abnormal indication; and a critical qualified motor marking unit configured to mark the corresponding motor as the critical qualified motor if the main frequency amplitude in the spectrum energy distribution meets the preset quality standard, and the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude has the preset abnormal indication.

[0075] In the following, the specific configuration of the critical qualified motor obtaining module 20 will be described in detail. As described above, the threshold segmentation is performed on the spectrum energy distribution to screen out the motors that meet the preset quality standard but have the preset abnormal indication, to obtain the critical qualified motor. The critical qualified motor obtaining module 20 can further include: a preset quality standard determining unit configured to dynamically collect the main frequency amplitude of the historical qualified motor sample, and determine the preset quality standard according to the dynamic threshold interval of the main frequency amplitude; a qualified bearing characteristic frequency indication determining unit configured to dynamically collect the bearing characteristic frequency amplitude of the historical qualified motor sample, and determine the qualified bearing characteristic frequency indication according to the dynamic threshold interval of the bearing characteristic frequency amplitude; a qualified electromagnetic harmonic frequency indication determining unit configured to dynamically collect the electromagnetic harmonic frequency amplitude of the historical qualified motor sample, and determine the qualified electromagnetic harmonic frequency indication according to the dynamic threshold interval of the electromagnetic harmonic frequency amplitude; a preset abnormal indication setting unit configured to set the amplitudes that do not meet the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication as the preset abnormal indication; and a critical qualified motor marking unit configured to mark the corresponding motor as the critical qualified motor if the main frequency amplitude in the spectrum energy distribution meets the preset quality standard, and the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude has the preset abnormal indication.

[0076] In the following, the specific configuration of the critical qualified motor obtaining module 20 will be described in detail. As described above, the threshold segmentation is performed on the spectrum energy distribution to screen out the motors that meet the preset quality standard but have the preset abnormal indication, to obtain the critical qualified motor. The critical qualified motor obtaining module 20 can further include: a preset quality standard determining unit configured to dynamically collect the main frequency amplitude of the historical qualified motor sample, and determine the preset quality standard according to the dynamic threshold interval of the main frequency amplitude; a qualified bearing characteristic frequency indication determining unit configured to dynamically collect the bearing characteristic frequency amplitude of the historical qualified motor sample, and determine the qualified bearing characteristic frequency indication according to the dynamic threshold interval of the bearing characteristic frequency amplitude; a qualified electromagnetic harmonic frequency indication determining unit configured to dynamically collect the electromagnetic harmonic frequency amplitude of the historical qualified motor sample, and determine the qualified electromagnetic harmonic frequency indication according to the dynamic threshold interval of the electromagnetic harmonic frequency amplitude; a preset abnormal indication setting unit configured to set the amplitudes that do not meet the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication as the preset abnormal indication; and a critical qualified motor marking unit configured to mark the corresponding motor as the critical qualified motor if the main frequency amplitude in the spectrum energy distribution meets the preset quality standard, and the bearing characteristic frequency amplitude or the electromagnetic harmonic frequency amplitude has the preset abnormal indication.

[0077] The specific configuration of the potential fault detection module 40 will be described in detail below. As described above, the fault feature library is pre-constructed, and the potential fault detection module 40 can further include: a first historical critical qualified motor sample dynamic acquisition unit for dynamically acquiring a first historical critical qualified motor sample, the first historical critical qualified motor sample being a motor of the same type as the motor to be detected, having served, and having appeared a first fault mode within the early fault exposure period; a first historical transient feature acquisition unit for acquiring a first historical transient feature of the first historical critical qualified motor sample; and a fault feature library construction unit for establishing a first mapping relationship between the first historical transient feature and the first fault mode, and constructing the fault feature library according to the first mapping relationship.

[0078] Wherein, the fault feature library is pre-constructed, the potential fault detection module 40 can further include: a second historical critical qualified motor sample dynamic acquisition unit for dynamically acquiring a second historical critical qualified motor sample, the second historical critical qualified motor sample being a motor of the same type as the motor to be detected, having served, and having appeared a second fault mode within the early fault exposure period, the fault correlation parameters including bearing type, rated speed and pole number; a second historical transient feature acquisition unit for acquiring a second historical transient feature of the second historical critical qualified motor sample; a conversion coefficient calculation unit for acquiring a first historical qualified motor sample and a second historical qualified motor sample corresponding one-to-one to the first historical critical qualified motor sample and the second historical critical qualified motor sample, and calculating a conversion coefficient according to the frequency offset ratio of the two; and a fault feature library expansion unit for converting the second historical transient feature according to the conversion coefficient to obtain a second converted historical transient feature, establishing a second mapping relationship between the second converted historical transient feature and the second fault mode, and adding the second mapping relationship into the fault feature library.

[0079] Wherein, the fault feature library is constructed according to the first mapping relationship, and the fault feature library construction unit can further include: a grouping subunit for grouping the first mapping relationship according to different fault modes to obtain a plurality of first mapping relationship groups; a historical transient feature library building subunit for extracting all historical transient features from each first mapping relationship group to build a plurality of historical transient feature libraries; and an early fault mode integration subunit for marking the plurality of historical transient feature libraries with corresponding fault modes to obtain a plurality of early fault modes, and integrating the plurality of early fault modes to obtain the fault feature library.

[0080] The potential fault detection module 40 can further comprise: a fault feature library traversal unit configured to traverse the fault feature library and extract a first early fault pattern; and a potential fault detection result output unit configured to calculate a similarity between the transient feature and historical transient features contained in the first early fault pattern, and output the first early fault pattern as a potential fault detection result if the calculation result is greater than or equal to a preset similarity threshold.

[0081] The motor noise detection system based on waveform analysis provided by the embodiment of the application can execute the motor noise detection method based on waveform analysis provided by any embodiment of the application, and has the corresponding function modules and beneficial effects of the execution method.

[0082] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module included is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific name of each functional unit is only for easy mutual differentiation, and does not limit the protection scope of the present application.

[0083] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps described in the present application can be executed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A method of detecting motor noise based on waveform analysis, characterized by, The method comprises: Original noise signals during motor operation are collected in time domain to obtain original waveform signals, Fourier transform is performed on the original waveform signals to extract frequency spectrum characteristics, and a frequency spectrum energy distribution is obtained; Threshold segmentation is performed on the frequency spectrum energy distribution to screen out motors that meet a preset quality standard but have a preset abnormal indication, and a critical qualified motor is obtained; The original waveform signals of the critical qualified motor are extracted, transient impact components are extracted from the original waveform signals, and transient characteristics are obtained; The transient characteristics are matched with early fault modes in a pre-constructed fault feature library, and a potential fault detection result is output.

2. The waveform analysis-based motor noise detection method of claim 1, wherein, Original noise signals during motor operation are collected in time domain to obtain original waveform signals, Fourier transform is performed on the original waveform signals to extract frequency spectrum characteristics, and a frequency spectrum energy distribution is obtained, comprising: At a rated speed of the motor, time domain noise signals are collected at a preset sampling frequency through a microphone array; Noise reduction is performed on the time domain noise signals to obtain the original waveform signals; Fast Fourier transform is performed on the original waveform signals to obtain a frequency spectrum graph; The amplitudes of main frequencies, bearing characteristic frequencies and electromagnetic harmonic frequencies in the frequency spectrum graph are calculated to generate the frequency spectrum energy distribution.

3. The waveform analysis-based motor noise detection method of claim 2, wherein, Threshold segmentation is performed on the frequency spectrum energy distribution to screen out motors that meet a preset quality standard but have a preset abnormal indication, and a critical qualified motor is obtained, comprising: The amplitudes of main frequencies of historical qualified motor samples are dynamically collected, and the preset quality standard is determined according to a dynamic threshold interval of the amplitudes of main frequencies; The amplitudes of bearing characteristic frequencies of the historical qualified motor samples are dynamically collected, and a qualified bearing characteristic frequency indication is determined according to a dynamic threshold interval of the amplitudes of bearing characteristic frequencies; The amplitudes of electromagnetic harmonic frequencies of the historical qualified motor samples are dynamically collected, and a qualified electromagnetic harmonic frequency indication is determined according to a dynamic threshold interval of the amplitudes of electromagnetic harmonic frequencies; Amplitudes that do not meet the qualified bearing characteristic frequency indication and the qualified electromagnetic harmonic frequency indication are set as the preset abnormal indication; If the amplitudes of main frequencies in the frequency spectrum energy distribution meet the preset quality standard, and the preset abnormal indication exists in the amplitudes of bearing characteristic frequencies or electromagnetic harmonic frequencies, the corresponding motor is marked as a critical qualified motor.

4. The waveform analysis-based motor noise detection method of claim 3, wherein, The historical qualified motor samples are motors that are consistent with the type of the motor to be detected, have been in service, and have not appeared a fault within an early fault exposure period.

5. The waveform analysis-based motor noise detection method of claim 1, wherein, Transient impact components are extracted from the original waveform signals to obtain transient characteristics, comprising: Wavelet packet decomposition is performed on the original waveform signals to obtain a time-frequency energy distribution matrix; Pulse detection is performed on the time-frequency energy distribution matrix to calculate pulse intervals and amplitudes, and the transient characteristics are obtained.

6. The waveform analysis-based motor noise detection method of claim 4, wherein, A fault feature library is pre-constructed, comprising: A first historical critical qualified motor sample is dynamically collected, the first historical critical qualified motor sample is a motor that is consistent with the type of the motor to be detected, has been in service, and has appeared a first fault mode within the early fault exposure period; First historical transient characteristics of the first historical critical qualified motor sample are obtained; establish a first mapping relationship between the first historical transient feature and the first failure mode, and construct the failure feature library according to the first mapping relationship.

7. The waveform analysis-based motor noise detection method of claim 6, wherein, The pre-constructed failure feature library further comprises: dynamically collecting a second historical critical qualified motor sample, the second historical critical qualified motor sample being a motor that is inconsistent with the to-be-detected motor model but consistent in failure-related parameters, has been in service, and has a second failure mode in the early failure exposure period, the failure-related parameters including a bearing model, a rated rotating speed, and a pole number; obtaining a second historical transient feature of the second historical critical qualified motor sample; collecting a first historical qualified motor sample and a second historical qualified motor sample corresponding to the first historical critical qualified motor sample and the second historical critical qualified motor sample, respectively, and calculating a conversion coefficient according to a frequency offset ratio of the two; converting the second historical transient feature according to the conversion coefficient to obtain a second converted historical transient feature, establishing a second mapping relationship between the second converted historical transient feature and the second failure mode, and adding the second mapping relationship into the failure feature library.

8. The waveform analysis-based motor noise detection method of claim 6, wherein, constructing the failure feature library according to the first mapping relationship comprises: grouping the first mapping relationships according to different failure modes to obtain a plurality of first mapping relationship groups; extracting all historical transient features from each first mapping relationship group to form a plurality of historical transient feature libraries; labeling the plurality of historical transient feature libraries according to corresponding failure modes to obtain a plurality of early failure modes, and integrating the plurality of early failure modes to obtain the failure feature library.

9. The waveform analysis based motor noise detection method of claim 1, wherein, matching the transient feature with early failure modes in the pre-constructed failure feature library, and outputting a potential failure detection result, comprising: traversing the failure feature library to extract a first early failure mode; calculating a similarity between the transient feature and a historical transient feature contained in the first early failure mode, and if the calculation result is greater than or equal to a preset similarity threshold, outputting the first early failure mode as a potential failure detection result.

10. A motor noise detection system based on waveform analysis, characterized by, The system is used to implement the motor noise detection method based on waveform analysis according to any one of claims 1-9, and the system comprises: a spectrum energy distribution acquisition module, configured to collect an original noise signal of a motor in time domain to obtain an original waveform signal, perform Fourier transform on the original waveform signal, extract a frequency spectrum feature, and obtain a spectrum energy distribution; a critical qualified motor acquisition module, configured to perform threshold segmentation on the spectrum energy distribution, filter out a motor that meets a preset quality standard but has a preset abnormal indication, and obtain a critical qualified motor; a transient feature acquisition module, configured to extract the original waveform signal of the critical qualified motor, extract a transient impact component from the original waveform signal, and obtain a transient feature; a potential failure detection module, configured to match the transient feature with early failure modes in a pre-constructed failure feature library, and output a potential failure detection result.

Images