A motor fault detection method and device based on a voiceprint classification model

Abstract

The application provides a motor fault detection method and device based on a voiceprint classification model, comprising the following steps: collecting an audio signal of a motor to be tested; preprocessing the audio signal of the motor to be tested; setting a custom mel filter bank, calling the custom mel filter bank to process each frame of signal, extracting the mel frequency cepstral coefficient feature of each frame, and combining the mel frequency cepstral coefficient features of all frames into a first feature matrix; performing standardization processing on the extracted first feature matrix; using multiple single-class support vector machine models based on a three-level classification strategy to determine whether there is a model matched with the first feature matrix, if yes, determining that the motor is running normally, and if no, determining that the motor has an obstacle. The application has the advantages of strong working condition adaptability and low misjudgment rate, uses a non-contact audio collection and a custom filter bank to improve feature recognition and abnormal identification accuracy, and has good generalization.

Description

Technical Field

[0001] This application relates to the field of industrial equipment condition monitoring and fault diagnosis technology, and more specifically, to a method and device for motor fault detection based on a voiceprint classification model. Background Technology

[0002] Electric motors are key power equipment in industrial production, and their operating status directly affects production safety and efficiency. Currently, motor condition monitoring mainly employs the following methods:

[0003] 1. Contact-based detection using vibration sensors requires mounting contact vibration sensors on the motor surface to determine the equipment status by collecting vibration signals. The disadvantages of this method are: complex sensor installation, difficult wiring, high cost, and it is not suitable for enclosed or high-speed rotating motor structures.

[0004] 2. Fault identification methods based on general audio features employ general audio features such as Mel-scale frequency cepstral coefficients (MFCC), spectral centroid, and zero-crossing rate for fault judgment. The drawbacks of this method are: it does not differentiate between power, speed, and cylinder block variations; sound signatures vary significantly under different loads and structures; and a uniform model is prone to misjudgment.

[0005] 3. Based on traditional thresholds or simple classification, the motor's state is judged by setting thresholds for one or more physical quantities. The disadvantages of this method are: it can only identify obvious faults, cannot provide early warnings, cannot distinguish between load fluctuations and physical faults, and has a high false alarm rate and poor localization capability.

[0006] 4. Monitoring based on traditional multi-class support vector machines. The disadvantage of this method is that it requires collecting training samples for all types of faults, but obtaining various fault samples in industrial sites is difficult and costly, resulting in weak model generalization. Retraining is required after changing the motor model or operating conditions.

[0007] In summary, there is a need in this field to provide a motor fault detection method and device based on a voiceprint classification model to overcome the shortcomings of the prior art. Summary of the Invention

[0008] This application provides a method and apparatus for motor fault detection based on a voiceprint classification model, which can solve the problems existing in the prior art. The objective of this application is achieved through the following technical solution.

[0009] In a first aspect, one embodiment of this application provides a motor fault detection method based on a voiceprint classification model, which includes several steps:

[0010] Step S1: Acquire the audio signal of the motor under test;

[0011] Step S2: Preprocess the audio signal of the motor under test;

[0012] Step S3: Set up a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into the first feature matrix; where, the custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for motor audio characteristics.

[0013] Step S4: Standardize the extracted first feature matrix; the standardization process uses Z-score standardization, Min-Max normalization, or MaxAbs normalization.

[0014] Step S5: Use multiple single-class support vector machine models based on a three-level classification strategy to determine whether there is a model that matches the first feature matrix. If yes, proceed to step S6; otherwise, proceed to step S7. The three-level classifications are arranged in the order of judgment as power, rotor frequency, and cylinder block.

[0015] Step S6: Determine that the motor is operating normally;

[0016] Step S7: Determine if a motor malfunction has occurred and output a malfunction report.

[0017] According to the above-described embodiment of the motor fault detection method provided in this application, the training method of the single-class support vector machine model includes the following steps:

[0018] Step S41: Collect the audio signal of the training motor operating normally;

[0019] Step S42: Preprocess the audio signal; wherein, the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation.

[0020] Step S43: Set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix.

[0021] Step S44: Standardize the extracted second feature matrix;

[0022] Step S45: Using the second feature matrix as input, train independent single-class support vector machine models according to the type of training motor; wherein, the type of motor corresponds to the three-level classification.

[0023] According to the motor fault detection method provided in one embodiment of this application, the method includes the following steps: setting a custom Mel filter bank, calling the custom Mel filter bank to process each frame of signal, and extracting the Mel frequency cepstral coefficient features of each frame.

[0024] Step S31: Initialize the custom Mel filter bank, set the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate;

[0025] Step S32: Linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points;

[0026] Step S33: Calculate the frequency resolution based on the number of FFT points, and convert the frequency points into FFT frequency domain index values;

[0027] Step S34: Generate the response curves of multiple filters in the form of a triangular window function to form a filter bank matrix.

[0028] According to the above-described embodiment of the present application, the motor fault detection method is divided into three types: high power, medium power, and low power. Among them, high power is greater than 70% of the preset motor power threshold, medium power is greater than 30% of the preset motor power threshold and less than or equal to 70% of the preset motor power threshold, and low power is less than or equal to 30% of the preset motor power threshold.

[0029] According to the above-described embodiment of the present application, the type of rotor frequency refers to the type of motor classified according to the rotor speed and the number of magnetic pole pairs, and the type of cylinder refers to the type of motor classified according to the cylinder structure.

[0030] According to the above-described embodiment of the present application, the motor fault detection method includes a first feature matrix consisting of 6 to 48 Vimer frequency cepstral coefficient feature matrices.

[0031] According to the motor fault detection method provided in one embodiment of this application, the preprocessing further includes windowing each frame of signal using a Hanning window, and then performing a fast Fourier transform on each frame of signal.

[0032] Secondly, one embodiment of this disclosure provides a motor fault detection device based on a voiceprint classification model, which includes:

[0033] First audio acquisition module: used to acquire audio signals of the motor under test;

[0034] First audio preprocessing module: used to preprocess the audio signal of the motor under test;

[0035] The first audio filtering module is used to set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into the first feature matrix. The custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for the audio characteristics of motors.

[0036] The first audio normalization module is used to normalize the extracted first feature matrix; the normalization process uses Z-score normalization, Min-Max normalization, or MaxAbs normalization.

[0037] Obstacle detection module: Used to determine whether there is a model that matches the first feature matrix using multiple single-class support vector machine models based on a three-level classification strategy. If yes, the motor is determined to be operating normally; if no, the motor is determined to be obstructed and an obstacle report is output. The three-level classification is arranged in the order of judgment as power, rotor frequency and cylinder block.

[0038] According to one embodiment of the present application, the motor fault detection device further includes a training module for a single-class support vector machine model, the training module comprising:

[0039] The second audio acquisition submodule is used to acquire audio signals from the training motor during normal operation.

[0040] The second audio preprocessing submodule is used to preprocess the audio signal; the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation.

[0041] The second audio filtering submodule is used to set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix.

[0042] The second audio normalization submodule is used to normalize the extracted second feature matrix.

[0043] Model training submodule: Used to train independent single-class support vector machine models based on the type of training motor, using the second feature matrix as input; where the type of motor corresponds to the three-level classification.

[0044] According to one embodiment of this application, the motor fault detection device includes setting a custom Mel filter bank, processing each frame of signal by calling the custom Mel filter bank, and extracting the Mel frequency cepstral coefficient features of each frame by calling the following sub-modules:

[0045] Initialization submodule: Used to initialize a custom Mel filter bank, setting the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate;

[0046] Frequency division point submodule: used to linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points;

[0047] Frequency Conversion Submodule: Used to calculate frequency resolution based on the number of FFT points and convert frequency points into FFT frequency domain index values;

[0048] The filter matrix submodule is used to generate the response curves of multiple filters in the form of a triangular window function, forming a filter bank matrix.

[0049] According to one embodiment of the present application, the motor fault detection device is divided into three types of power: high power, medium power, and low power. Among them, high power is greater than 70% of the preset motor power threshold, medium power is greater than 30% of the preset motor power threshold and less than or equal to 70% of the preset motor power threshold, and low power is less than or equal to 30% of the preset motor power threshold.

[0050] According to the above-described embodiment of the present application, the type of rotor frequency refers to the type of motor classified according to the rotor speed and the number of magnetic pole pairs, and the type of cylinder refers to the type of motor classified according to the cylinder structure.

[0051] According to one embodiment of the present application, the motor fault detection device provides a first feature matrix that is a 6 to 48 Vimer frequency cepstral coefficient feature matrix.

[0052] According to one embodiment of the present application, the motor fault detection device further includes preprocessing by applying a Hanning window to each frame of signal and then performing a fast Fourier transform on each frame of signal.

[0053] The advantages of the motor fault detection method and apparatus based on the voiceprint classification model according to the embodiments of this application are as follows: It is applicable to motor audio collected through non-contact methods, adaptable to various motors, and reduces deployment costs; it classifies motors into three levels—power segment, rotor frequency, and cylinder structure—with strong adaptability to operating conditions, effectively distinguishing voiceprint differences under different loads; a custom filter bank enhances the recognizability of motor audio features and improves the accuracy of anomaly identification; it allows training the model using only normal samples, solving the problem of insufficient fault samples in industrial settings; the end-to-end process supports embedded real-time deployment, enabling online determination of motor status; multiple models can be trained independently, allowing for independent judgment of motor anomalies under different operating conditions, reducing the false judgment rate; it has good generalization across motor models and operating conditions, requiring only the addition of a sub-model when changing motor models, without needing to retrain all models. Attached Figure Description

[0054] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings.

[0055] Figure 1 An exemplary system architecture diagram is shown, in which one embodiment of this disclosure can be applied.

[0056] Figure 2 A flowchart of a motor fault detection method based on a voiceprint classification model according to one embodiment of this application is shown.

[0057] Figure 3 A flowchart illustrating a training method for a single-class support vector machine model according to one embodiment of this application is shown.

[0058] Figure 4 The flowchart illustrates a process according to one embodiment of this application, invoking a custom Mel filter bank to process each frame of signal, extracting the Mel frequency cepstral coefficient features of each frame, and combining the Mel frequency cepstral coefficient features of all frames into a first feature matrix.

[0059] Figure 5 A schematic diagram of the structure of an embodiment of a motor fault detection device based on a voiceprint classification model according to one embodiment of the present disclosure is shown;

[0060] Figure 6 A schematic diagram of the structure of a computer system of an electronic device according to one embodiment of the present disclosure is shown. Detailed Implementation

[0061] The specific embodiments of this application are described below with reference to the accompanying drawings and examples. Through the content described in this specification, those skilled in the art can clearly and completely understand the technical solution, the technical problem solved, and the resulting technical effects of this application. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, for ease of description, only the parts related to this application are shown in the accompanying drawings.

[0062] It should be noted that the structures, proportions, sizes, etc., shown in the accompanying drawings are only for the purpose of assisting those skilled in the art in understanding and reading the contents described in the specification, and are not intended to limit the conditions under which this application can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportional relationships, or adjustments to the size should fall within the scope of the technical content disclosed in this application, provided that they do not affect the effects and purposes that this application can produce.

[0063] The use of terms such as "first," "second," and "the" does not imply quantity limitation and may indicate singular or plural. The terms "comprising," "including," "having," and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules is not limited to the listed steps or units, but may also include steps or units not listed, or other steps or units inherent to these processes, methods, products, or devices. The terms "connected," "linked," and "coupled" used in this application are not limited to physical or mechanical connections, but may also include direct or indirect electrical connections.

[0064] Figure 1 An exemplary system architecture is shown, illustrating an embodiment of the motor fault detection method based on a voiceprint classification model of this disclosure. (See also...) Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0065] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104, thereby loading pages to be tested, etc. Terminal devices 101, 102, and 103 can be hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc. When terminal devices 101, 102, and 103 are software, they can be installed on the terminal devices listed above. They can be implemented as multiple software programs or software modules (e.g., used to run motor fault detection software) or as a single software program or software module. No specific limitations are made here.

[0066] In some cases, the motor fault detection method provided in this disclosure can be executed by terminal devices 101, 102, and 103, and correspondingly, the motor fault detection device can be set in terminal devices 101, 102, and 103. In this case, the system architecture 100 may not include server 105.

[0067] In some cases, the motor fault detection method provided in this disclosure can be executed jointly by terminal devices 101, 102, 103 and server 105. This disclosure does not limit this. Correspondingly, the motor fault detection device can also be separately installed in terminal devices 101, 102, 103 and server 105.

[0068] In some cases, the motor fault detection method provided in this disclosure can be executed by server 105. Accordingly, the motor fault detection device can also be set in server 105. In this case, the system architecture 100 may not include terminal devices 101, 102, and 103.

[0069] It should be noted that server 105 can be either hardware or software. When server 105 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When server 105 is software, it can be implemented as multiple software programs or software modules, or as a single software program or software module. No specific limitations are made here.

[0070] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0071] Figure 2 A flowchart of a motor fault detection method based on a voiceprint classification model according to one embodiment of this application is shown. Figure 2 As shown, the motor fault detection method includes several steps:

[0072] Step S1: Acquire the audio signal of the motor under test;

[0073] Step S2: Preprocess the audio signal of the motor under test;

[0074] Step S3: Set up a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into the first feature matrix; where, the custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for motor audio characteristics.

[0075] Step S4: Standardize the extracted first feature matrix; the standardization process uses Z-score standardization, Min-Max normalization, or MaxAbs normalization.

[0076] Step S5: Use multiple single-class support vector machine models based on a three-level classification strategy to determine whether there is a model that matches the first feature matrix. If yes, proceed to step S6; otherwise, proceed to step S7. The three-level classifications are arranged in the order of judgment as power, rotor frequency, and cylinder block.

[0077] Step S6: Determine that the motor is operating normally;

[0078] Step S7: Determine if a motor malfunction has occurred and output a malfunction report.

[0079] According to the motor fault detection method provided in one embodiment of this application, the single-class support vector machine model (i.e., OneClassSVM) refers to an unsupervised / semi-supervised learning model trained using only samples of a single class (i.e., normal state) to determine whether a new sample belongs to that class (i.e., whether it is abnormal). Its goal is to learn the compact boundary of the training samples in the feature space, such that normal samples are located inside the boundary and abnormal samples are located outside the boundary.

[0080] According to the motor fault detection method provided in one embodiment of this application, those skilled in the art can set the training parameters of OneClassSVM based on experience. Preferably, nu is 0.01 to 0.1, the kernel function is RBF, and gamma is set to 'scale' or determined according to the reciprocal of the feature dimension.

[0081] According to the motor fault detection method provided in one embodiment of this application, the Mel frequency cepstral coefficient feature (i.e., MFCC feature) is an audio feature representation method based on human auditory perception characteristics or a custom frequency domain division method. In this invention, the extraction of MFCC features employs a custom linearly divided Mel filter bank, typically setting 48 triangular filters covering the entire frequency band from 0 to half the sampling rate. Six to 48-dimensional MFCC coefficients are extracted from each frame to form a feature matrix. This feature matrix can effectively characterize the acoustic signature differences of various types of motors during operation.

[0082] Figure 3 A flowchart illustrating a training method for a single-class support vector machine model according to one embodiment of this application is shown. Figure 3 As shown, the training method for a single-class support vector machine model includes the following steps:

[0083] Step S41: Collect the audio signal of the training motor operating normally;

[0084] Step S42: Preprocess the audio signal; wherein, the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation.

[0085] Step S43: Set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix.

[0086] Step S44: Standardize the extracted second feature matrix;

[0087] Step S45: Using the second feature matrix as input, train independent single-class support vector machine models according to the type of training motor; wherein, the type of motor corresponds to the three-level classification.

[0088] This invention trains OneClassSVM models according to a three-level classification of motors. Specifically: Audio recordings of the training motors during normal operation are collected, and three OneClassSVM models are trained for high, medium, and low power respectively. Then, multiple OneClassSVM models are trained for motors with different rotor frequencies at different power levels. Finally, multiple OneClassSVM models are trained for motors with different cylinder blocks at the aforementioned rotor frequencies. During monitoring, the OneClassSVM models are used sequentially according to the three levels of power, rotor frequency, and cylinder block to determine whether the collected audio signals of the motor under test match. If a OneClassSVM model matches the audio signal of the motor under test, it means that the motor under test is working normally; otherwise, it means that there is a problem with the operation of the motor under test.

[0089] Figure 4 A flowchart illustrating one embodiment of this application is provided, showing the process of invoking a custom Mel filter bank to process each frame of signal, extracting the Mel frequency cepstral coefficient features of each frame, and combining the Mel frequency cepstral coefficient features of all frames into a first feature matrix. Figure 4 As shown, setting up a custom Mel filter bank and using it to process each frame of signal to extract the Mel frequency cepstral coefficient features of each frame includes the following steps:

[0090] Step S31: Initialize the custom Mel filter bank, set the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate;

[0091] Step S32: Linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points;

[0092] Step S33: Calculate the frequency resolution based on the number of FFT points, and convert the frequency points into FFT frequency domain index values;

[0093] Step S34: Generate the response curves of multiple filters in the form of a triangular window function to form a filter bank matrix.

[0094] According to the above-described embodiment of the present application, the motor fault detection method is divided into three types: high power, medium power, and low power. Among them, high power is greater than 70% of the preset motor power threshold, medium power is greater than 30% of the preset motor power threshold and less than or equal to 70% of the preset motor power threshold, and low power is less than or equal to 30% of the preset motor power threshold.

[0095] According to the motor fault detection method provided in one embodiment of this application, the rotor frequency type refers to the type of motor classified according to the rotor speed and the number of magnetic pole pairs, and the cylinder type refers to the type of motor classified according to the cylinder structure. Cylinder structures include, but are not limited to, air-cooled cast iron integral cylinders, water-cooled jacketed cylinders, and aluminum alloy stretched thin-walled cylinders.

[0096] According to the motor fault detection method provided in one embodiment of this application, the first feature matrix is ​​a 6- to 48-dimensional Mel-frequency cepstral coefficient feature matrix. The lower dimension is used for rapid monitoring, and the higher dimension is used for detailed diagnosis. Preferably, the first feature matrix is ​​a 24-dimensional Mel-frequency cepstral coefficient feature matrix.

[0097] According to the motor fault detection method provided in one embodiment of this application, the noise reduction process can be selected by those skilled in the art based on the motor model and / or usage scenario. For example, a Butterworth bandpass filter with a cutoff frequency of 50Hz to 22kHz can be used for noise reduction, or spectral subtraction can be used to remove steady-state background noise.

[0098] According to the motor fault detection method provided in one embodiment of this application, the Mel frequency cepstral coefficient features of each frame can be obtained by taking the logarithm of the energy output of each filter and then performing a discrete cosine transform (DCT) to finally obtain the Mel frequency cepstral coefficient features of each frame (this general method will lose a lot of information for industrial equipment audio acquisition and judgment).

[0099] According to the motor fault detection method provided in one embodiment of this application, the acquisition device does not need to be in direct contact with the motor when acquiring audio signals of the motor under test and the training motor. The type and location of the acquisition device can be selected by those skilled in the art based on their professional knowledge. For example, an omnidirectional condenser microphone with a sampling rate of 48kHz and a bit depth of 16 bits is used, and the microphone is placed 0.5 meters in front of the motor housing, axially aligned with the motor bearing.

[0100] According to the motor fault detection method provided in one embodiment of this application, the number of filters can be adjusted according to actual monitoring needs, for example: 48 or 32.

[0101] According to the motor fault detection method provided in one embodiment of this application, the preprocessing further includes windowing each frame of signal using a Hanning window, and then performing a fast Fourier transform on each frame of signal.

[0102] Secondly, Figure 5 A schematic diagram of one embodiment of a motor fault detection device based on a voiceprint classification model according to an embodiment of the present disclosure is shown. Figure 5 As shown, the motor fault detection device based on the voiceprint classification model includes:

[0103] First audio acquisition module 201: used to acquire audio signals of the motor under test;

[0104] First audio preprocessing module 202: used to preprocess the audio signal of the motor under test;

[0105] First audio filtering module 203: used to set a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a first feature matrix; wherein, the custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for motor audio characteristics.

[0106] First audio normalization module 204: used to normalize the extracted first feature matrix; wherein, the normalization process adopts Z-score normalization, Min-Max normalization or MaxAbs normalization;

[0107] Obstacle detection module 205: Used to determine whether there is a model that matches the first feature matrix using multiple single-class support vector machine models based on a three-level classification strategy. If yes, the motor is determined to be operating normally; if no, the motor is determined to be obstructed and an obstacle report is output. The three-level classification is arranged in the order of judgment as power, rotor frequency and cylinder.

[0108] It should be noted that the sub-modules in the subsequent embodiments are not all labeled in the figures.

[0109] According to one embodiment of the present application, the motor fault detection device further includes a training module 206 for a single-class support vector machine model, the training module 206 comprising:

[0110] The second audio acquisition submodule is used to acquire audio signals from the training motor during normal operation.

[0111] The second audio preprocessing submodule is used to preprocess the audio signal; the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation.

[0112] The second audio filtering submodule is used to set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix.

[0113] The second audio normalization submodule is used to normalize the extracted second feature matrix.

[0114] Model training submodule: Used to train independent single-class support vector machine models based on the type of training motor, using the second feature matrix as input; where the type of motor corresponds to the three-level classification.

[0115] According to one embodiment of this application, the motor fault detection device includes setting a custom Mel filter bank, processing each frame of signal by calling the custom Mel filter bank, and extracting the Mel frequency cepstral coefficient features of each frame by calling the following sub-modules:

[0116] Initialization submodule: Used to initialize a custom Mel filter bank, setting the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate;

[0117] Frequency division point submodule: used to linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points;

[0118] Frequency Conversion Submodule: Used to calculate frequency resolution based on the number of FFT points and convert frequency points into FFT frequency domain index values;

[0119] The filter matrix submodule is used to generate the response curves of multiple filters in the form of a triangular window function, forming a filter bank matrix.

[0120] According to one embodiment of the present application, the motor fault detection device is divided into three types of power: high power, medium power, and low power. Among them, high power is greater than 70% of the preset motor power threshold, medium power is greater than 30% of the preset motor power threshold and less than or equal to 70% of the preset motor power threshold, and low power is less than or equal to 30% of the preset motor power threshold.

[0121] According to the above-described embodiment of the present application, the type of rotor frequency refers to the type of motor classified according to the rotor speed and the number of magnetic pole pairs, and the type of cylinder refers to the type of motor classified according to the cylinder structure.

[0122] According to one embodiment of the present application, the motor fault detection device provides a first feature matrix that is a 6 to 48 Vimer frequency cepstral coefficient feature matrix.

[0123] According to one embodiment of the present application, the motor fault detection device further includes preprocessing by applying a Hanning window to each frame of signal and then performing a fast Fourier transform on each frame of signal.

[0124] Figure 6 A schematic diagram of the structure of a computer system for implementing embodiments of the present disclosure is shown. Figure 6 The computer system 500 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure. Figure 6 As shown, the computer system 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which is capable of performing various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the computer system 500. The processing device 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0125] Typically, the following devices can be connected to I / O interface 505: input devices 506 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, etc.; output devices 507 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 508 including, for example, magnetic tapes, hard disks, etc.; and communication devices 509. Communication device 509 allows computer system 500 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 A computer system 500 with various electronic devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0126] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including methods for executing... Figure 1The flowchart illustrates the program code for the method. In such an embodiment, the computer program can be downloaded and installed from a network via communication device 509, or installed from storage device 508, or installed from ROM 502. When the computer program is executed by processing device 501, it performs the functions defined in the methods of embodiments of this disclosure.

[0127] It should be noted that the two embodiments given below are merely illustrative, and the values ​​assigned to the various parameters are not intended to be limiting. Those skilled in the art can reasonably set the relevant parameters when implementing the technical solutions in this application according to the actual circumstances.

[0128] Example 1: Training an independent OneClassSVM model and using it

[0129] • Collect the audio of the motor running, read it using audioread, and convert it into double-precision data;

[0130] • Divide the frame into 1024 frames with a frame length of 512 and a frame shift of 512, add a Hanning window and perform FFT transformation;

[0131] • Extracting acoustic features using a custom filter bank adapted to motor power, rotor frequency, and cylinder resonance. This method generates a triangular filter bank that is highly matched to the motor power range, rotor frequency, and cylinder vibration characteristics based on the motor operating audio sampling rate. This is used to accurately extract motor acoustic features and solve the problem that traditional filter banks cannot adapt to the frequency domain characteristics of motor operating conditions.

[0132] • Build multi-category labels based on high / medium / low power, rotor frequency, and cylinder type;

[0133] • After standardizing the features, train a multi-level OneClassSVM model;

[0134] • Input the signal to be tested into the model, and output the working condition classification and obstacle determination results.

[0135] Example 2: customFilterBank Feature Extraction Method

[0136] Step 1: Parameter Initialization and Sampling Rate Configuration

[0137] The custom filter bank construction function takes the audio sampling rate Fs as the input parameter; if no sampling rate is input, the default industrial standard sampling rate of 48000Hz is used; set the number of filters p=48, the number of FFT transform points nfft=1024, and the frequency coverage range from 0 to 1 / 2 of the sampling rate.

[0138] Step 2: Linear division of frequency range

[0139] The frequency range is divided into equal segments using a linear method, generating p+2 equal frequency points within the range of 0 to Fs / 2. This ensures that the filter bank uniformly covers the motor's full-frequency operating acoustic signal and adapts to different rotor frequencies and cylinder resonant frequency bands.

[0140] Step 3: Frequency Domain Index Calculation

[0141] The frequency resolution df = Fs / nfft is calculated based on the number of FFT points. The divided frequency points are converted into FFT frequency domain index values ​​to obtain the left boundary index n1, center index n0, and right boundary index n2 for each filter. This is adapted to the MATLAB array indexing rule that starts from 1, where nfft is the number of FFT transform points.

[0142] Step 4: Generating the triangular filter

[0143] Initialize a filter bank matrix with 24 rows, for example, 48K data points and nfft / 2 columns; iterate through the 48 filters and construct the response curve for each filter using a triangular window function:

[0144] • The response value increases linearly from the left boundary to the center of the filter;

[0145] • The response value decreases linearly from the center of the filter to the right boundary; finally, a triangular filter bank with bandpass filtering characteristics is generated to realize frequency domain weighting and feature enhancement of the motor acoustic signal.

[0146] Step 5: Filter bank output

[0147] The constructed filter bank matrix is ​​used as the output result for subsequent frequency domain filtering and feature extraction of motor acoustic signals.

[0148] The advantages of the motor fault detection method and apparatus based on the voiceprint classification model according to the embodiments of this application are as follows: It is applicable to motor audio collected through non-contact methods, adaptable to various motors, and reduces deployment costs; it classifies motors into three levels—power segment, rotor frequency, and cylinder structure—with strong adaptability to operating conditions, effectively distinguishing voiceprint differences under different loads; a custom filter bank enhances the recognizability of motor audio features and improves the accuracy of anomaly identification; it allows training the model using only normal samples, solving the problem of insufficient fault samples in industrial settings; the end-to-end process supports embedded real-time deployment, enabling online determination of motor status; multiple models can be trained independently, allowing for independent judgment of motor anomalies under different operating conditions, reducing the false judgment rate; it has good generalization across motor models and operating conditions, requiring only the addition of a sub-model when changing motor models, without needing to retrain all models.

[0149] Although this application has been described and illustrated with reference to specific embodiments thereof, such descriptions and illustrations are not intended to limit the application. It will be readily understood by those skilled in the art that various changes can be made and equivalent elements can be substituted in the embodiments without departing from the scope of protection of this application as defined by the claims. Differences may exist between the technical representation in this application and actual equipment due to variables in the manufacturing process, etc. Other embodiments of this application may exist that are not specifically described. The specification and illustrations should be regarded as illustrative rather than restrictive, and modifications can be made to suit the purpose and spirit of this application, all of which are within the scope of protection of the claims. While the methods disclosed herein have been described with reference to specific operations performed in a particular order, it should be understood that these operations can be rearranged, subdivided, or arranged to form equivalent methods without departing from the teachings of this application. Therefore, unless specifically indicated herein, the order and grouping of operations do not limit the application.

Claims

1. A method for motor fault detection based on a voiceprint classification model, characterized in that, It includes several steps: Step S1: Acquire the audio signal of the motor under test; Step S2: Preprocess the audio signal of the motor under test; Step S3: Set up a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a first feature matrix; wherein, the custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for motor audio characteristics. Step S4: Standardize the extracted first feature matrix; wherein the standardization process uses Z-score standardization, Min-Max normalization or MaxAbs normalization. Step S5: Use multiple single-class support vector machine models based on a three-level classification strategy to determine whether there is a model that matches the first feature matrix. If yes, determine that the motor is running normally; if no, determine that the motor has encountered a malfunction and output a malfunction report. The three-level classifications are arranged in the order of judgment as power, rotor frequency and cylinder.

2. The motor fault detection method according to claim 1, characterized in that, The training method for the single-class support vector machine model includes the following steps: Step S41: Collect the audio signal of the training motor operating normally; Step S42: Preprocess the audio signal; wherein the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation. Step S43: Set a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix; Step S44: Standardize the extracted second feature matrix; Step S45: Using the second feature matrix as input, train independent single-class support vector machine models according to the type of the training motor; wherein the type of the motor corresponds to the three-level classification.

3. The motor fault detection method according to claim 2, characterized in that, The process of setting a custom Mel filter bank, calling the custom Mel filter bank to process each frame of signal, and extracting the Mel frequency cepstral coefficient features of each frame includes the following steps: Step S31: Initialize the custom Mel filter bank, set the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate; Step S32: Linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points; Step S33: Calculate the frequency resolution based on the number of FFT points, and convert the frequency points into FFT frequency domain index values; Step S34: Generate the response curves of multiple filters in the form of a triangular window function to form a filter bank matrix.

4. The motor fault detection method according to claim 1, characterized in that, The power type is divided into three categories: high power, medium power, and low power; wherein, the high power is greater than 70% of the preset power threshold of the motor, the medium power is greater than 30% of the preset power threshold of the motor and less than or equal to 70% of the preset power threshold of the motor, and the low power is less than or equal to 30% of the preset power threshold of the motor.

5. The motor fault detection method according to claim 4, characterized in that, The rotor frequency type refers to the type of motor classified according to the rotor speed and the number of magnetic pole pairs, and the cylinder type refers to the type of motor classified according to the cylinder structure.

6. The motor fault detection method according to claim 4, characterized in that, The first characteristic matrix is ​​a cepstral coefficient characteristic matrix of 6 to 48 Vimer frequencies.

7. The motor fault detection method according to claim 1, characterized in that, The preprocessing also includes windowing each frame of signal using a Hanning window, and then performing a fast Fourier transform on each frame of signal.

8. A motor fault detection device based on a voiceprint classification model, characterized in that, It includes: First audio acquisition module: used to acquire audio signals of the motor under test; First preprocessing module: performs the aforementioned preprocessing on the audio signal of the motor under test; First filtering module: Set up a custom Mel filter bank, call the custom Mel filter bank to process each frame of signal, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a first feature matrix; wherein, the custom Mel filter bank refers to a triangular filter bank with non-standard Mel frequency scale and linear frequency division method that is optimized for motor audio characteristics. First standardization module: performs standardization processing on the extracted first feature matrix; wherein, the standardization processing adopts Z-score standardization, Min-Max normalization or MaxAbs normalization; Obstacle detection module: Uses multiple single-class support vector machine models based on a three-level classification strategy to determine whether there is a model of the motor that matches the first feature matrix. If yes, the motor is determined to be operating normally; if no, the motor is determined to be obstructed and an obstacle report is output. The three-level classification is arranged in the order of judgment as power, rotor frequency and cylinder.

9. The motor fault detection device according to claim 8, characterized in that, The motor fault detection device further includes a training module for a single-class support vector machine model, the training module comprising: The second audio acquisition submodule is used to acquire audio signals from the training motor during normal operation. The second audio preprocessing submodule is used to preprocess the audio signal; the preprocessing includes at least noise reduction, amplitude normalization and frame segmentation. The second audio filtering submodule is used to set up a custom Mel filter bank, call the custom Mel filter bank to process the signal of each frame, extract the Mel frequency cepstral coefficient features of each frame, and combine the Mel frequency cepstral coefficient features of all frames into a second feature matrix. The second audio normalization submodule is used to normalize the extracted second feature matrix. Model training submodule: Used to train independent single-class support vector machine models based on the type of training motor, using the second feature matrix as input; where the type of motor corresponds to the three-level classification.

10. The motor fault detection device according to claim 9, characterized in that, The process of setting a custom Mel filter bank, processing each frame of signal using the custom Mel filter bank, and extracting the Mel frequency cepstral coefficient features of each frame includes calling the following sub-modules: Initialization submodule: Used to initialize a custom Mel filter bank, setting the number of filters p to 48, the number of FFT points to 1024, and the frequency coverage range to 0 to 1 / 2 of the sampling rate; Frequency division point submodule: used to linearly divide the frequency range from 0 to Fs / 2 into p+2 equally divided frequency points; Frequency Conversion Submodule: Used to calculate frequency resolution based on the number of FFT points and convert frequency points into FFT frequency domain index values; The filter matrix submodule is used to generate the response curves of multiple filters in the form of a triangular window function, forming a filter bank matrix.

Images