A motor electromagnetic fault detection method, device, terminal equipment and storage medium

Abstract

The present application discloses a motor electromagnetic fault detection method, device, terminal device and storage medium. The present application uses infrared thermal imaging equipment to collect the red-hot image of the motor, and obtains the temperature value corresponding to each pixel of the red-hot image through temperature mapping; when the temperature value corresponding to any pixel in the preset area exceeds the first threshold, the noise sensor collects the first noise signal generated when the motor is running; through fast Fourier transform and characteristic frequency matching, according to the first noise signal, a number of third noise signals are screened from all preset second noise signals; through spectrum subtraction, each third noise signal is removed from the first noise signal in turn to obtain a fourth noise signal; through a pre-trained decision tree model, the fourth noise signal and the red-hot image are combined to obtain the electromagnetic fault detection result. Through the present application, electromagnetic faults can be accurately identified in scenarios where multiple faults occur simultaneously without changing the motor structure.

Description

Technical Field

[0001] The present application relates to the field of motor data processing, and in particular to a motor electromagnetic fault detection method, device, terminal equipment and storage medium. Background Art

[0002] Electromagnetic faults are usually caused by abnormal electromagnetic characteristics of the motor during operation. Electromagnetic faults can lead to reduced equipment performance, reduced efficiency, and even damage to the equipment. Common electromagnetic faults include winding short circuits, current imbalance, electromagnetic interference, magnetic circuit abnormalities, etc. These faults not only affect the normal operation of the equipment, but may also cause more serious secondary faults, such as mechanical damage and overheating.

[0003] The existing electromagnetic fault detection method obtains the motor operation data by connecting additional detection equipment, such as analyzing the input current and output current of the monitoring equipment and detecting the change of the magnetic flux density inside the motor by connecting the magnetic flux sensor. The data collected by the above sensors is further combined with machine learning methods to train the corresponding fault identification model or build a corresponding simulation model for fault judgment. However, the above method requires changes to the motor structure, resulting in unnecessary downtime operations, and the constructed model is usually highly complex and requires a large number of early feature processing procedures to eliminate noise data generated by other faults.

[0004] Therefore, how to effectively eliminate the impact of other faults on electromagnetic fault judgment without changing the motor structure and ensure calculation efficiency is a technical problem that needs to be solved in current electromagnetic fault detection. Summary of the invention

[0005] The present application provides a motor electromagnetic fault detection method, device, terminal equipment and storage medium to solve technical problems.

[0006] In order to solve the above technical problems, in a first aspect, an embodiment of the present application provides a motor electromagnetic fault detection method, comprising:

[0007] The red-hot image of the motor is collected in real time by infrared thermal imaging equipment, and the temperature value corresponding to each pixel point of the red-hot image is obtained by temperature mapping;

[0008] When a temperature value corresponding to any pixel point in the preset area exceeds a first threshold, a first noise signal generated when the motor is running is collected by a noise sensor;

[0009] By fast Fourier transform and characteristic frequency matching, according to the first noise signal, a plurality of third noise signals are screened from all preset second noise signals; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical failure scenarios;

[0010] By spectrum subtraction, the third noise signals are sequentially removed from the first noise signal to obtain a fourth noise signal;

[0011] The electromagnetic fault detection result is obtained by combining the pre-trained decision tree model with the fourth noise signal and the red-hot image; the decision tree model is obtained by training the historical operation data of the motor in normal operation scenarios and electromagnetic fault scenarios.

[0012] Compared with the prior art, the embodiments of the present application have the following beneficial effects: since electromagnetic faults are usually caused by the interaction between electric current and magnetic field, and mechanical faults are usually caused by mechanical damage such as model breakage of the hardware equipment of the motor, the noises corresponding to the two have different characteristics, and the occurrence of electromagnetic faults and mechanical faults will cause abnormal heating of the motor. Therefore, through the red thermal imaging device and the noise collection device, there is no need to change the mechanical structure of the motor itself, and only the corresponding sensors need to be set up externally to achieve the detection of electromagnetic faults; and when multiple faults occur in the motor at the same time, the noise caused by the mechanical fault usually makes the noise of the electromagnetic fault difficult to identify. Therefore, by simulating the second noise signal under various types of mechanical faults in advance, based on the second noise signal, the noise generated in the actual motor operating environment is denoised, and the fourth noise signal that may be caused by the electromagnetic fault is retained, so that only a decision tree is needed to achieve fault detection, avoiding the extraction of small effective features from complex noise by constructing a complex simulation model or machine learning model, effectively improving the computing efficiency, and solving the negative impact of other faults on electromagnetic fault detection.

[0013] In some embodiments of the first aspect of the present application, the method of screening a plurality of third noise signals from all the second noise signals according to the first noise signal by fast Fourier transform and characteristic frequency matching includes:

[0014] extracting a plurality of fifth noise signals from the first noise signal through independent component analysis;

[0015] By fast Fourier transform, each of the fifth noise signals is converted into a first spectrum, and each of the second noise signals is converted into a corresponding second spectrum;

[0016] By matching characteristic frequencies, sequentially calculating the similarity between each of the first frequency spectra and each of the second frequency spectra;

[0017] The second noise signal corresponding to the second frequency spectrum whose similarity is greater than a second threshold is used as the third noise signal.

[0018] Compared with the prior art, the above embodiment has the following beneficial effects: through independent component analysis, different noises can be separated from the first noise signal, and the actually collected fifth noise signal can be matched with the second noise signal in each mechanical fault scenario, so as to accurately judge the mechanical fault occurring during the current motor operation, so that the noise signal caused by the electromagnetic fault can be accurately separated from the first noise signal subsequently.

[0019] In some embodiments of the first aspect of the present application, the calculating the similarity between each of the first spectra and each of the second spectra in sequence by characteristic frequency matching includes:

[0020] Extracting a number of corresponding first characteristic frequencies from the first spectrum, and extracting a number of corresponding second characteristic frequencies from the second spectrum;

[0021] The Euclidean distances between all the second characteristic frequencies and all the first characteristic frequencies are calculated as the similarities between the corresponding first spectrum and the second spectrum.

[0022] Compared with the prior art, the above embodiment has the following beneficial effects: by extracting characteristic frequencies from the spectrum, the two characteristic frequencies are aligned, so that the similarity between the noise signals corresponding to the two spectra can be accurately evaluated by calculating the Euclidean distance between the two characteristic frequencies, thereby improving the accuracy of mechanical fault judgment.

[0023] In some embodiments of the first aspect of the present application, the calculating the Euclidean distance between all the second characteristic frequencies and all the first characteristic frequencies as the similarity between the corresponding first spectrum and the second spectrum includes:

[0024] The calculation formula of the Euclidean distance is specifically:

[0025]

[0026] in, is the Euclidean distance; is the total number of characteristic frequencies; The first spectrum The first characteristic frequency; The second spectrum The second characteristic frequency; For the The amplitude of the first characteristic frequency; For the The amplitude of the second characteristic frequency.

[0027] Compared with the prior art, the above embodiment has the following beneficial effects: by calculating the Euclidean distance of the characteristic frequency, the accuracy of the similarity evaluation is ensured, and at the same time, it has the advantage of low computational complexity, thereby improving the real-time performance of electromagnetic fault detection.

[0028] In some embodiments of the first aspect of the present application, removing the third noise signals from the first noise signal in sequence by spectrum subtraction to obtain the fourth noise signal includes:

[0029] By fast Fourier transform, the first noise signal is converted into a third frequency spectrum, and each of the third noise signals is converted into a corresponding fourth frequency spectrum;

[0030] Determining a common frequency point between the third spectrum and each of the fourth spectrums by characteristic frequency matching;

[0031] Subtracting the corresponding fourth spectrum from the third spectrum according to the common frequency points in sequence to obtain a fifth spectrum;

[0032] The fifth spectrum is converted into the fourth noise signal by inverse fast Fourier transform.

[0033] Compared with the prior art, the above embodiment has the following beneficial effects: through spectrum subtraction, the interference components in the noise can be efficiently eliminated to ensure that the extracted fourth noise signal is purer, thereby improving the accuracy of subsequent electromagnetic fault detection; at the same time, through fast Fourier transform, the data processing efficiency is improved and the real-time detection is guaranteed.

[0034] In some embodiments of the first aspect of the present application, the obtaining of the electromagnetic fault detection result by combining the pre-trained decision tree model with the fourth noise signal and the red hot image includes:

[0035] Extracting time domain features from the fourth noise signal, and extracting frequency domain features from the fourth noise signal by fast Fourier transform;

[0036] In combination with the structural diagram of the motor, a first region corresponding to the stator and the rotor of the motor is segmented from the red-hot image, and a temperature value corresponding to each pixel point in the first region is extracted as a thermal map feature;

[0037] The time domain features, frequency domain features and heat map features are spliced ​​and input into the decision tree model to obtain electromagnetic fault detection results.

[0038] Compared with the prior art, the above embodiment has the following beneficial effects: by combining the extraction of time domain and frequency domain features and comprehensively considering multiple information, the performance of the decision tree model in fault detection is improved, and the accuracy of electromagnetic fault detection is improved. At the same time, by introducing thermal map features, temperature information can be combined with noise signals to provide a more comprehensive basis for fault diagnosis.

[0039] In some embodiments of the first aspect of the present application, the step of segmenting the first region corresponding to the stator and the rotor of the motor from the red-hot image in combination with the structural diagram of the motor includes:

[0040] Dividing the red-hot image into a plurality of sub-regions by an image segmentation algorithm;

[0041] Overlapping the divided red-hot image with the structure diagram of the motor, and selecting the sub-region whose overlapping area with the target region exceeds a third threshold from all the sub-regions; the target region is the region corresponding to the stator and rotor of the motor in the structure diagram;

[0042] All the sub-regions are concatenated to form the first region.

[0043] Compared with the prior art, the above embodiment has the following beneficial effects: through the combination of image segmentation algorithm and structure diagram, the key components of the motor can be accurately identified, ensuring that the extracted thermal map features are more relevant and effective, thereby improving the accuracy of subsequent detection results.

[0044] In a second aspect, the embodiment of the present application further provides a motor electromagnetic fault detection device, including: a red-hot image acquisition module, a noise acquisition module, a noise signal screening module, an electromagnetic fault noise signal extraction module, and an electromagnetic fault detection result acquisition module;

[0045] The red-hot image acquisition module is used to acquire the red-hot image of the motor in real time through infrared thermal imaging equipment, and obtain the temperature value corresponding to each pixel point of the red-hot image through temperature mapping;

[0046] The noise collection module is used to collect a first noise signal generated when the motor is running through a noise sensor when the temperature value corresponding to any pixel point in the preset area exceeds a first threshold value;

[0047] The noise signal screening module is used to screen a plurality of third noise signals from all preset second noise signals according to the first noise signal by fast Fourier transform and characteristic frequency matching; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical failure scenarios;

[0048] The electromagnetic fault noise signal extraction module is used to remove each of the third noise signals from the first noise signal in turn through spectrum subtraction to obtain a fourth noise signal;

[0049] The electromagnetic fault detection result acquisition module is used to obtain the electromagnetic fault detection result through a pre-trained decision tree model in combination with the fourth noise signal and the red hot image; the decision tree model is obtained through training of historical operating data of the motor in normal operating scenarios and electromagnetic fault scenarios.

[0050] In a third aspect, the present application also provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the above-mentioned motor electromagnetic fault detection method when executing the computer program.

[0051] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned motor electromagnetic fault detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A schematic flow chart of a motor electromagnetic fault detection method provided in some embodiments of the present application;

[0053] Figure 2 This is a schematic diagram of the structure of a motor electromagnetic fault detection device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0054] The existing electromagnetic fault detection method obtains the motor operation data by connecting additional detection equipment, such as analyzing the input current and output current of the monitoring equipment and detecting the change of the magnetic flux density inside the motor by connecting the magnetic flux sensor. The data collected by the above sensors is further combined with machine learning methods to train the corresponding fault identification model or build a corresponding simulation model for fault judgment. However, the above method requires changes to the motor structure, resulting in unnecessary downtime operations, and the constructed model is usually highly complex and requires a large number of early feature processing procedures to eliminate noise data generated by other faults.

[0055] In order to solve the above technical problems, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0056] Embodiment 1

[0057] Please refer to Figure 1 , a motor electromagnetic fault detection method provided by an embodiment of the present application, includes S10 to S50, specifically:

[0058] S10: using an infrared thermal imaging device to collect a red-hot image of the motor in real time, and obtaining a temperature value corresponding to each pixel point of the red-hot image through temperature mapping.

[0059] Preferably, in some embodiments of the present application, the mapping relationship between the temperature and pixel value of different infrared thermal imaging devices is different, and the present application does not limit the temperature mapping method; after obtaining the temperature value, the temperature value is filtered or smoothed to improve the accuracy of the temperature value, specifically: through Gaussian filtering, high-frequency noise in the temperature value is removed to prevent abnormal changes in the temperature value corresponding to a certain pixel point in the temperature value, thereby affecting subsequent judgments.

[0060] S20: When the temperature value corresponding to any pixel point in the preset area exceeds a first threshold, a first noise signal generated when the motor is running is collected by a noise sensor.

[0061] Preferably, when the infrared thermal imaging device is installed and fixed, a red-hot image can be taken at a corresponding installation position to determine the area where the motor is located in the red-hot image, thereby obtaining a preset area in the red-hot image where the temperature value of the pixel point needs to be determined.

[0062] Preferably, in some embodiments of the present application, the first noise signal is collected by a multi-channel microphone array; the microphones in the microphone array are respectively arranged at different positions of the motor.

[0063] S30: Through fast Fourier transform and characteristic frequency matching, according to the first noise signal, a plurality of third noise signals are screened from all preset second noise signals; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical failure scenarios.

[0064] Furthermore, in some embodiments of the present application, the method of screening a plurality of third noise signals from all the second noise signals according to the first noise signal by fast Fourier transform and characteristic frequency matching includes:

[0065] extracting a plurality of fifth noise signals from the first noise signal through independent component analysis;

[0066] By fast Fourier transform, each of the fifth noise signals is converted into a first spectrum, and each of the second noise signals is converted into a corresponding second spectrum;

[0067] The second noise signal corresponding to the second frequency spectrum whose similarity is greater than a second threshold is used as the third noise signal.

[0068] Through independent component analysis, different noises can be separated from the first noise signal, and the actually collected fifth noise signal can be matched with the second noise signal in each mechanical fault scenario, so as to accurately judge the mechanical fault occurring during the current motor operation, so that the noise signal caused by the electromagnetic fault can be accurately separated from the first noise signal in the future.

[0069] Preferably, in some embodiments of the present application, several fifth noise signals can be extracted from the first noise signal by the following preferred implementation:

[0070] S31: performing mean processing on the first noise signal collected from each channel to centralize the data;

[0071] S32: performing principal component analysis on the data processed by S31, and whitening the data corresponding to each channel in turn, thereby reducing the correlation between different signals and improving independence;

[0072] S33: Input the data acquired in S32 into a FastICA (Fast Independent Component Analysis) algorithm to acquire each fifth noise signal.

[0073] Furthermore, in some embodiments of the present application, the similarity between each of the first spectrum and each of the second spectrum is calculated in sequence through characteristic frequency matching, including:

[0074] Extracting a number of corresponding first characteristic frequencies from the first spectrum, and extracting a number of corresponding second characteristic frequencies from the second spectrum;

[0075] The Euclidean distances between all the second characteristic frequencies and all the first characteristic frequencies are calculated as the similarities between the corresponding first spectrum and the second spectrum.

[0076] By extracting characteristic frequencies from the spectrum and ensuring that the two characteristic frequencies are aligned, the similarity between the noise signals corresponding to the two spectra can be accurately evaluated by calculating the Euclidean distance between the two characteristic frequencies, thereby improving the accuracy of mechanical fault judgment.

[0077] Preferably, in some embodiments of the present application, when extracting the corresponding first characteristic frequencies from the first spectrum and extracting the corresponding second characteristic frequencies from the second spectrum, the method further includes:

[0078] S34: sequentially determining peak values ​​in each spectrum as temporary characteristic frequencies;

[0079] S35: Since the number of peaks in different spectra may be different, the number of temporary feature frequencies in the two spectra will be different. Therefore, before calculating the similarity, feature alignment is required to screen the temporary feature frequencies. Specifically, assuming that the number of temporary feature frequencies determined in the current first spectrum is less than the number of temporary feature frequencies determined in the second spectrum, the nearest neighbor matching method is used to search in the second spectrum for the temporary feature frequencies that are nearest to each temporary feature frequency in the first spectrum in turn as the second feature frequencies, and all temporary feature frequencies in the first spectrum are used as the first feature frequencies.

[0080] Further, in some embodiments of the present application, the calculating the Euclidean distance between all the second characteristic frequencies and all the first characteristic frequencies as the similarity between the corresponding first spectrum and the second spectrum includes:

[0081] The calculation formula of the Euclidean distance is specifically:

[0082]

[0083] in, is the Euclidean distance; is the total number of characteristic frequencies; The first spectrum The first characteristic frequency; The second spectrum The second characteristic frequency; For the The amplitude of the first characteristic frequency; For the The amplitude of the second characteristic frequency.

[0084] S40: removing the third noise signals from the first noise signal in sequence through spectrum subtraction to obtain a fourth noise signal.

[0085] Furthermore, in some embodiments of the present application, removing the third noise signals from the first noise signal in sequence by spectrum subtraction to obtain the fourth noise signal includes:

[0086] By fast Fourier transform, the first noise signal is converted into a third frequency spectrum, and each of the third noise signals is converted into a corresponding fourth frequency spectrum;

[0087] Determining a common frequency point between the third spectrum and each of the fourth spectrums by characteristic frequency matching;

[0088] Subtracting the corresponding fourth spectrum from the third spectrum according to the common frequency points in sequence to obtain a fifth spectrum;

[0089] The fifth spectrum is converted into the fourth noise signal by inverse fast Fourier transform.

[0090] Through spectral subtraction, the interference components in the noise can be effectively eliminated to ensure that the extracted fourth noise signal is purer, thereby improving the accuracy of subsequent electromagnetic fault detection; at the same time, through fast Fourier transform, the data processing efficiency is improved and the real-time detection is guaranteed.

[0091] Preferably, in some embodiments of the present application, the common frequency point of the third spectrum and each of the fourth spectra is determined by characteristic frequency matching, and the common frequency point may be determined by the above-mentioned S35.

[0092] Preferably, in some embodiments of the present application, the fifth spectrum may be obtained by the following preferred implementation:

[0093] S41: subtracting the amplitude of the noise signal in the fourth spectrum from the amplitude of the noise signal in the third spectrum corresponding to the current common frequency point;

[0094] S42: Determine whether the subtraction result of S41 is less than 0. If so, set the current subtraction result to 0. Otherwise, continue to execute S41 until all common frequency points are traversed.

[0095] S50: Obtaining electromagnetic fault detection results through a pre-trained decision tree model in combination with the fourth noise signal and the red-hot image; the decision tree model is obtained through training with historical operating data of the motor in normal operating scenarios and electromagnetic fault scenarios.

[0096] Furthermore, in some embodiments of the present application, the electromagnetic fault detection result is obtained by combining the pre-trained decision tree model with the fourth noise signal and the red-hot image, including:

[0097] Extracting time domain features from the fourth noise signal, and extracting frequency domain features from the fourth noise signal by fast Fourier transform;

[0098] In combination with the structural diagram of the motor, a first region corresponding to the stator and the rotor of the motor is segmented from the red-hot image, and a temperature value corresponding to each pixel point in the first region is extracted as a thermal map feature;

[0099] The time domain features, frequency domain features and heat map features are spliced ​​and input into the decision tree model to obtain electromagnetic fault detection results.

[0100] By combining the extraction of time domain and frequency domain features and comprehensively considering multiple information, the performance of the decision tree model in fault detection is improved, and the accuracy of electromagnetic fault detection is improved. At the same time, by introducing thermal map features, temperature information can be combined with noise signals to provide a more comprehensive basis for fault diagnosis.

[0101] Preferably, in some embodiments of the present application, the decision tree model is trained and acquired by the following steps:

[0102] S51: taking the noise signal and the red-hot image collected in the normal operation scenario as positive samples, and taking the noise signal and the red-hot image collected in the electromagnetic fault scenario as negative samples;

[0103] S52: Perform feature processing on the sample data, and extract the time domain features, frequency domain features and heat map features of each sample in turn, wherein the time domain features include: mean, standard deviation, peak value and root mean square; the frequency domain features include: feature frequency and amplitude; the heat map features include: temperature average, temperature maximum and hot spot position;

[0104] S53: Standardize the feature values ​​obtained in S52, unify the dimensions of the standardized data, and then perform feature concatenation to obtain a feature vector for input into a decision tree model;

[0105] S54: Iteratively train the decision tree model using the feature vector of the sample data obtained in S53 until the accuracy of the decision tree model reaches a preset condition.

[0106] Further, in some embodiments of the present application, the step of segmenting the first region corresponding to the stator and the rotor of the motor from the red-hot image in combination with the structural diagram of the motor includes:

[0107] Dividing the red-hot image into a plurality of sub-regions by an image segmentation algorithm;

[0108] Overlapping the divided red-hot image with the structure diagram of the motor, and selecting the sub-region whose overlapping area with the target region exceeds a third threshold from all the sub-regions; the target region is the region corresponding to the stator and rotor of the motor in the structure diagram;

[0109] All the sub-regions are concatenated to form the first region.

[0110] By combining the image segmentation algorithm with the structure diagram, the key components of the motor can be accurately identified, ensuring that the extracted thermal map features are more relevant and effective, thereby improving the accuracy of subsequent detection results.

[0111] Preferably, in some embodiments of the present application, the structural diagram of the motor is a cross-sectional diagram of the motor corresponding to a red-hot image shooting angle.

[0112] In summary, a motor electromagnetic fault detection method provided in an embodiment of the present application has the following beneficial effects: since electromagnetic faults are usually caused by the interaction between current and magnetic field, and mechanical faults are usually caused by mechanical damage such as model breakage of the motor's hardware equipment, the noises corresponding to the two have different characteristics, and the occurrence of electromagnetic faults and mechanical faults will cause abnormal heating of the motor. Therefore, through red thermal imaging equipment and noise collection equipment, there is no need to change the mechanical structure of the motor itself, and only corresponding sensors need to be set up externally to achieve electromagnetic fault detection; and when multiple faults occur in the motor at the same time, the noise caused by mechanical faults usually makes the noise of electromagnetic faults difficult to identify. Therefore, by simulating the second noise signal under various types of mechanical faults in advance, based on the second noise signal, the noise generated in the actual motor operating environment is denoised, and the fourth noise signal that may be caused by electromagnetic faults is retained, so that fault detection can be achieved by only using a decision tree, avoiding the extraction of small effective features from complex noise by constructing complex simulation models or machine learning models, effectively improving computing efficiency, and solving the negative impact of other faults on electromagnetic fault detection.

[0113] Embodiment 2

[0114] refer to Figure 2 , a motor electromagnetic fault detection device provided in an embodiment of the present application, includes: a red-hot image acquisition module 11, a noise acquisition module 12, a noise signal screening module 13, an electromagnetic fault noise signal extraction module 14 and an electromagnetic fault detection result acquisition module 15.

[0115] Further, in some embodiments of the present application, the red-hot image acquisition module 11 is used to acquire the red-hot image of the motor in real time through an infrared thermal imaging device, and obtain the temperature value corresponding to each pixel of the red-hot image through temperature mapping; the noise acquisition module 12 is used to acquire the first noise signal generated when the motor is running through a noise sensor when the temperature value corresponding to any pixel in the preset area exceeds a first threshold; the noise signal screening module 13 is used to screen a plurality of third noise signals from all preset second noise signals according to the first noise signal through fast Fourier transform and characteristic frequency matching; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical fault scenarios; the electromagnetic fault noise signal extraction module 14 is used to remove each of the third noise signals from the first noise signal in turn through spectrum subtraction to obtain a fourth noise signal; the electromagnetic fault detection result acquisition module 15 is used to obtain the electromagnetic fault detection result through a pre-trained decision tree model in combination with the fourth noise signal and the red-hot image; the decision tree model is obtained through training of historical operation data of the motor in normal operation scenarios and electromagnetic fault scenarios.

[0116] It can be understood that the above-mentioned device item embodiment corresponds to the method item embodiment of the present invention. A motor electromagnetic fault detection device provided by the embodiment of the present invention can implement any method item embodiment of the present invention, that is, the motor electromagnetic fault detection method provided in embodiment one.

[0117] In summary, a motor electromagnetic fault detection device provided by an embodiment of the present application has the following beneficial effects: since electromagnetic faults are usually caused by the interaction between current and magnetic field, and mechanical faults are usually caused by mechanical damage such as model breakage of the motor's hardware equipment, the noises corresponding to the two have different characteristics, and the occurrence of electromagnetic faults and mechanical faults will cause abnormal heating of the motor. Therefore, through red thermal imaging equipment and noise collection equipment, there is no need to change the mechanical structure of the motor itself, and only corresponding sensors need to be set up externally to achieve electromagnetic fault detection; and when multiple faults occur in the motor at the same time, the noise caused by mechanical faults usually makes the noise of electromagnetic faults difficult to identify. Therefore, by simulating the second noise signal under various types of mechanical faults in advance, based on the second noise signal, the noise generated in the actual motor operating environment is denoised, and the fourth noise signal that may be caused by electromagnetic faults is retained, so that fault detection can be achieved by only using a decision tree, avoiding the extraction of small effective features from complex noise by constructing complex simulation models or machine learning models, effectively improving computing efficiency, and solving the negative impact of other faults on electromagnetic fault detection.

[0118] Embodiment 3

[0119] Based on the above-mentioned embodiment of the motor electromagnetic fault detection method, another embodiment of the present application provides a motor electromagnetic fault detection terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the motor electromagnetic fault detection method of any embodiment of the present application is implemented.

[0120] Exemplarily, in this embodiment, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the motor electromagnetic fault detection device.

[0121] The motor electromagnetic fault detection device can be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The motor electromagnetic fault detection terminal device can include, but is not limited to, a processor and a memory.

[0122] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The processor is the control center of the motor electromagnetic fault detection device, and uses various interfaces and lines to connect the various parts of the entire motor electromagnetic fault detection device. The memory may be used to store the computer program and / or module, and the processor implements various functions of the motor electromagnetic fault detection device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0123] Embodiment 4

[0124] Based on the above-mentioned embodiment of the motor electromagnetic fault detection method, another embodiment of the present application provides a storage medium, wherein the storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the motor electromagnetic fault detection method of any embodiment of the present application.

[0125] In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0126] The specific embodiments described above further describe the purpose, technical solutions and beneficial effects of the present application in detail. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the scope of protection of the present application. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A method for detecting electromagnetic faults of a motor, characterized in that: include: The red-hot image of the motor is collected in real time by infrared thermal imaging equipment, and the temperature value corresponding to each pixel point of the red-hot image is obtained by temperature mapping; When a temperature value corresponding to any pixel point in the preset area exceeds a first threshold, a first noise signal generated when the motor is running is collected by a noise sensor; By fast Fourier transform and characteristic frequency matching, according to the first noise signal, a plurality of third noise signals are screened from all preset second noise signals; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical failure scenarios; By spectrum subtraction, the third noise signals are sequentially removed from the first noise signal to obtain a fourth noise signal; The electromagnetic fault detection result is obtained by combining the fourth noise signal and the red-hot image with a pre-trained decision tree model; the decision tree model is obtained by training the historical operation data of the motor in normal operation scenarios and electromagnetic fault scenarios; Among them, the method of screening a plurality of third noise signals from all the second noise signals according to the first noise signal through fast Fourier transform and characteristic frequency matching includes: extracting a plurality of fifth noise signals from the first noise signal through independent component analysis; converting each of the fifth noise signals into a first spectrum and converting each of the second noise signals into a corresponding second spectrum through fast Fourier transform; calculating the similarity between each of the first spectrum and each of the second spectrum in turn through characteristic frequency matching; and taking the second noise signal corresponding to the second spectrum whose similarity is greater than a second threshold as the third noise signal.

2. A motor electromagnetic fault detection method as claimed in claim 1, characterized in that: The calculating the similarity between each of the first frequency spectra and each of the second frequency spectra in sequence through characteristic frequency matching includes: Extracting a number of corresponding first characteristic frequencies from the first spectrum, and extracting a number of corresponding second characteristic frequencies from the second spectrum; The Euclidean distances between all the second characteristic frequencies and all the first characteristic frequencies are calculated as the similarities between the corresponding first spectrum and the second spectrum.

3. A motor electromagnetic fault detection method as claimed in claim 2, characterized in that: The calculating the Euclidean distance between all the second characteristic frequencies and all the first characteristic frequencies as the similarity between the corresponding first spectrum and the second spectrum includes: The calculation formula of the Euclidean distance is specifically: ； in, is the Euclidean distance; is the total number of characteristic frequencies; The first spectrum The first characteristic frequency; The second spectrum The second characteristic frequency; For the The amplitude of the first characteristic frequency; For the The amplitude of the second characteristic frequency.

4. A motor electromagnetic fault detection method as claimed in claim 1, characterized in that: The step of removing the third noise signals from the first noise signal in sequence by spectrum subtraction to obtain a fourth noise signal includes: By fast Fourier transform, the first noise signal is converted into a third frequency spectrum, and each of the third noise signals is converted into a corresponding fourth frequency spectrum; Determining a common frequency point between the third spectrum and each of the fourth spectrums by characteristic frequency matching; Subtracting the corresponding fourth spectrum from the third spectrum according to the common frequency points in sequence to obtain a fifth spectrum; The fifth spectrum is converted into the fourth noise signal by inverse fast Fourier transform.

5. A motor electromagnetic fault detection method as claimed in claim 1, characterized in that: The method of obtaining an electromagnetic fault detection result by combining the pre-trained decision tree model with the fourth noise signal and the red-hot image includes: Extracting time domain features from the fourth noise signal, and extracting frequency domain features from the fourth noise signal by fast Fourier transform; In combination with the structural diagram of the motor, a first region corresponding to the stator and the rotor of the motor is segmented from the red-hot image, and a temperature value corresponding to each pixel point in the first region is extracted as a thermal map feature; The time domain features, frequency domain features and heat map features are spliced ​​and input into the decision tree model to obtain electromagnetic fault detection results.

6. A motor electromagnetic fault detection method as claimed in claim 5, characterized in that: The step of segmenting the first region corresponding to the stator and the rotor of the motor from the red-hot image in combination with the structural diagram of the motor includes: Dividing the red-hot image into a plurality of sub-regions by an image segmentation algorithm; Overlapping the divided red-hot image with the structure diagram of the motor, and selecting the sub-region whose overlapping area with the target region exceeds a third threshold from all the sub-regions; the target region is the region corresponding to the stator and rotor of the motor in the structure diagram; All the sub-regions are concatenated to form the first region.

7. A motor electromagnetic fault detection device, characterized in that: include: Red-hot image acquisition module, noise acquisition module, noise signal screening module, electromagnetic fault noise signal extraction module and electromagnetic fault detection result acquisition module; The red-hot image acquisition module is used to acquire the red-hot image of the motor in real time through infrared thermal imaging equipment, and obtain the temperature value corresponding to each pixel point of the red-hot image through temperature mapping; The noise collection module is used to collect a first noise signal generated when the motor is running through a noise sensor when the temperature value corresponding to any pixel point in the preset area exceeds a first threshold value; The noise signal screening module is used to screen a plurality of third noise signals from all preset second noise signals according to the first noise signal by fast Fourier transform and characteristic frequency matching; the second noise signal is obtained by collecting the noise generated when the motor is running under different mechanical failure scenarios; The electromagnetic fault noise signal extraction module is used to remove each of the third noise signals from the first noise signal in turn through spectrum subtraction to obtain a fourth noise signal; The electromagnetic fault detection result acquisition module is used to obtain the electromagnetic fault detection result by combining the fourth noise signal and the red-hot image through a pre-trained decision tree model; the decision tree model is obtained by training the historical operation data of the motor in normal operation scenarios and electromagnetic fault scenarios; Among them, the method of screening a plurality of third noise signals from all the second noise signals according to the first noise signal through fast Fourier transform and characteristic frequency matching includes: extracting a plurality of fifth noise signals from the first noise signal through independent component analysis; converting each of the fifth noise signals into a first spectrum and converting each of the second noise signals into a corresponding second spectrum through fast Fourier transform; calculating the similarity between each of the first spectrum and each of the second spectrum in turn through characteristic frequency matching; and taking the second noise signal corresponding to the second spectrum whose similarity is greater than a second threshold as the third noise signal.

8. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a motor electromagnetic fault detection method as claimed in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a motor electromagnetic fault detection method as claimed in any one of claims 1 to 6.

Images