CN114282580A - Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system - Google Patents
Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system Download PDFInfo
- Publication number
- CN114282580A CN114282580A CN202111668865.1A CN202111668865A CN114282580A CN 114282580 A CN114282580 A CN 114282580A CN 202111668865 A CN202111668865 A CN 202111668865A CN 114282580 A CN114282580 A CN 114282580A
- Authority
- CN
- China
- Prior art keywords
- fault diagnosis
- classifier
- fault
- permanent magnet
- dimensional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Control Of Electric Motors In General (AREA)
Abstract
The invention discloses a visual image-based permanent magnet drive motor demagnetization fault diagnosis model construction method, a fault diagnosis method and a system, wherein in the method, a motor surface magnetic flux leakage signal is used as an original signal for fault diagnosis and is converted into a two-dimensional Fourier spectrogram, and then global features and local features are extracted and fused. And finally, constructing a classifier by using the fused features, and particularly preferably performing fault diagnosis on the softmax classifier. The method adopts two-dimensional Fourier transform to convert the one-dimensional time domain signal into a spectrogram, and effective fault characteristics hidden in the time domain signal are shown; visual features of the images are extracted through a self-encoder method, feature effectiveness is improved, fault diagnosis is carried out through a softmax classifier with a simple structure and a small amount of calculation, and the problem of demagnetization fault diagnosis of a permanent magnet driving motor for the electric automobile is effectively solved.
Description
Technical Field
The invention relates to a motor demagnetization fault diagnosis technology, in particular to a permanent magnet drive motor demagnetization fault diagnosis model construction method based on visual images, a fault diagnosis method and a system, and particularly can be applied to electric automobiles.
Background
The new energy automobile industry in China has gone through the development process of 20 years, and China has become the production country and the consumption country of the largest new energy automobile in the world, wherein the scale of the pure electric automobile accounts for more than 50% of the world, and the first global level is internationally advanced. Along with the continuous improvement of new energy automobile intellectuality, integration, its inner structure is also complicated day by day. In recent two years, safety accidents of new energy automobiles present an increasing trend, the industry has gradually changed from mileage anxiety to safety anxiety, and safety problems have become one of the core problems to be solved in the development of new energy automobiles. In 5 months in 2020, three mandatory national standards of 'safety requirements for electric vehicles', 'safety requirements for electric motor cars' and 'safety requirements for power storage batteries for electric vehicles' are continuously issued by the ministry of industry and informatization, and the safety technical requirements for the whole new energy vehicles are definitely provided.
The power center of the electric automobile is the motor, the permanent magnet driving motor adopts the permanent magnet to generate a motor magnetic field, and the permanent magnet driving motor has simple structure and high efficiency and control precision and is widely applied to the fields of electric automobiles, aerospace, wind power generation and the like. The power of domestic and international electric vehicles is mostly provided by permanent magnet synchronous motors. The magnetic steel sheet of the permanent magnet driving motor is mostly made of neodymium iron boron permanent magnet materials, and the Curie temperature of the magnetic steel sheet is low. Therefore, overload of the motor, damage of a heat dissipation system and the like can cause magnetic loss of the permanent magnet, and demagnetization of the permanent magnet is easily caused. The demagnetization fault can aggravate torque ripple and motor loss, seriously reduce the performance of the automobile and cause property loss and casualties in serious cases.
The state signals for demagnetization fault diagnosis at present are motor stator current, voltage, torque, magnetic flux and the like. The change of the state signals is caused by the indirect influence of the demagnetization fault of the permanent magnet, and is easily interfered by other faults. The motor status signal not only contains fault characteristics, but also is mixed with redundant signals and other interference signals. Therefore, the extraction of the fault characteristics is important in the motor fault diagnosis. The traditional method focuses on one-dimensional characteristics of time domain and frequency domain, and requires complex and professional signal processing knowledge. And the state signal is expanded to a two-dimensional image, so that fault diagnosis is facilitated. However, in the conventional method, only a global feature extraction method is adopted to extract image features, or only a local feature extraction method is adopted to extract features, but both methods have certain defects. For example, the global feature extraction cannot be applied to image occlusion, rotation and the like, and the local features cannot give consideration to the problem of low feature efficiency caused by the comprehensive features.
In summary, the existing diagnostic methods have the following technical problems:
1) in the traditional feature extraction based on one-dimensional time domain or frequency domain signals, the signal preprocessing is complex, and the professional requirement is strong.
2) The global feature extraction cannot be applied to image shielding, rotation and the like, and meanwhile, the local features cannot give consideration to the problem of low feature efficiency caused by comprehensive features.
Therefore, it is necessary to research a more accurate and effective method and system for diagnosing demagnetization faults of a permanent magnet driving motor.
Disclosure of Invention
The invention aims to solve at least part of technical problems in the prior art, and further provides a method for constructing a demagnetization fault diagnosis model of a permanent magnet driving motor based on a visual image, and a fault diagnosis method and system. According to the method, a leakage flux signal of a fault motor is selected and extracted as an original signal, and a one-dimensional time domain signal is converted into a two-dimensional image, namely a two-dimensional Fourier spectrogram, so that the fault signal characteristics are enriched; in addition, the global features and the local features are fused, and the two problems in the prior art are effectively overcome.
On one hand, the invention provides a visual image-based method for constructing a demagnetization fault diagnosis model of a permanent magnet drive motor, which comprises the following steps: acquiring signals, performing two-dimensional image conversion on the acquired signals, extracting global features and local features of the two-dimensional images, fusing the global features and the local features, and constructing a fault diagnosis classifier by utilizing the fused features, wherein the fault diagnosis classifier specifically comprises the following steps:
step 1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
step 2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
and step 3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
and 4, step 4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
the classifier is applied to demagnetization fault diagnosis of the permanent magnet synchronous motor.
The invention selects the leakage magnetic signal as the original signal of the fault diagnosis, considers that the leakage magnetic signal can directly reflect the demagnetization fault, compared with other voltage and current signals which indirectly reflect the demagnetization fault, the effectiveness of the fault characteristic can be improved, the interference by other faults is less, and secondly, the leakage magnetic signal outside the motor shell is generally measured, so the motor does not need to be disassembled, and the artificial interference can not be introduced.
Furthermore, the invention converts the magnetic leakage signal into the two-dimensional Fourier spectrogram, so that the characteristic of the fault data after being converted into the image is more effectively embodied by considering the two-dimensional Fourier transformation in the image formed by the one-dimensional signal, and the invention is favorable for fault diagnosis. The invention uses the spectrogram obtained by converting the one-dimensional signal into the image and then performing Fourier transform on the image for fault diagnosis, can remove the influence of the amplitude of the one-dimensional signal, and only performs characteristic extraction on the frequency domain of the image formed by the one-dimensional signal.
Further optionally, in step 3, a global feature and a local feature of the two-dimensional fourier spectrogram are extracted by using an auto-encoder and a scale-invariant feature transform, respectively.
According to the invention, researches show that for a two-dimensional Fourier spectrogram, ideal demagnetization fault diagnosis cannot be obtained by a single local feature or a global feature, and more effective fault features can be extracted from the two-dimensional Fourier spectrogram by fusing the global feature and the local feature, so that a better diagnosis effect is obtained, and the method is particularly applied to the field of automobiles.
Further optionally, when extracting the local features of the two-dimensional fourier spectrogram by using scale-invariant feature transformation, feature points are searched on different scale spaces;
wherein, the scale space L (x, y, σ) of the image is a convolution of a blurred gaussian and a two-dimensional fourier spectrogram, which is expressed as:
L(x，y，σ)＝G(x，y，σ)*I(x，y)
the feature points are represented as:
D(x，y，σ)＝(G(x，y，kσ)-G(x，y，σ))*I(x，y)
in the formula, D (x, y, σ) is a feature point corresponding to an element I (x, y) on the two-dimensional fourier spectrogram, G (x, y, σ) is a gaussian fuzzy representation of the element I (x, y) on the two-dimensional fourier spectrogram, k is a calculation constant, and σ is a standard deviation of normal distribution;
each feature point is assigned a reference direction, which is expressed as:
θ(x，y)＝tan-1((L(x，y+1)-L(x，y+1))/(L(x+1，y)-L(x-1，y)))
in the formula, θ (x, y) represents a reference direction, and the function L () represents a scale space value.
Further optionally, when the magnetic leakage signal is converted into the two-dimensional fourier spectrogram in step 2, first, N × N points that are 2 times larger than the motor operation time during data acquisition are taken from the magnetic leakage signal to form a two-dimensional matrix f (x, y), and then the two-dimensional matrix f (x, y) is converted into the two-dimensional spectrogram H (v, p) through two-dimensional fourier transform, where v and p are coordinates of a horizontal axis and a vertical axis of a frequency domain coordinate in the two-dimensional spectrogram of the image respectively;
wherein, the two-dimensional Fourier transform formula is as follows:
wherein, translating the frequency origin to the center, the translation formula is:
the two-dimensional fourier spectrogram I (x, y) and the two-dimensional spectrogram H (v, p) have the same meaning and can both represent fourier spectrograms.
Further optionally, the fault diagnosis classifier adopts a softmax classifier, the fused features are input into the softmax classifier, classification probability values are obtained, the sum of the classification probability values of all the classes of faults is 1, and the classification probability values are represented as:
wherein, yiRepresenting the sofmax classifier input, n representing its dimension, and e being the natural base.
Common intelligent classification models include support vector machines, decision trees, neural networks, and the like. The classification model can obtain better diagnosis effect, and aiming at the fusion characteristics of the invention, the existing network can be adopted to construct the classifier, the construction process is not optimized, and the prior art can be referred.
However, the present invention further contemplates that the internal parameters and hyper-parameters of the above network models affect their performance. And the hyper-parameters of the classification model are different for different classification objects. Therefore, for different fault diagnoses, a professional and complex hyper-parameter optimization adjustment process is often required, such as the maximum number of partitions of a decision tree or the frame constraints of a support vector machine. Although neural networks exist as high-level networks with good mobility, most of them have a deep network layer, a long training process, and poor transparency and interpretability. In view of the above-mentioned shortcomings of the network, in order to further enhance the performance of the fault diagnosis classifier of the present invention, the present invention further contemplates providing the aforementioned softmax classifier.
The Softmax classifier is generally used for classification as the last layer of a convolutional neural network or some other complex neural network with a larger number of layers. The structure is simple, the interpretability is strong, and the poor transparency and the poor analyzability in the complex neural network are mainly caused by a pooling layer, a connecting layer and the like in the middle layer. The softmax classifier can achieve good effect under the combined characteristics, simplifies the diagnosis process and increases the interpretability of the diagnosis process. After the one-dimensional signal is converted into the two-dimensional Fourier spectrogram, the signal processing process is simplified by only paying attention to the frequency domain characteristic of the magnetic leakage signal after the fault is converted into the spectrogram, the fault characteristic is highlighted, and effective characteristics are obtained by combining the global characteristic and the local characteristic, so that the classifier with softmax has few parameters, and the diagnosis process can be simplified by using a classifier with strong explanatory property.
In a second aspect, the present invention provides a method for diagnosing a demagnetization fault of a permanent magnet driving motor based on the above method for constructing a fault diagnosis model, which includes: the method comprises the following steps of constructing a fault diagnosis classifier and diagnosing faults based on the fault diagnosis classifier, wherein the fault diagnosis classifier specifically comprises the following steps:
s1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
s2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
s3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
s4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
s5: and acquiring a magnetic leakage signal of the permanent magnet synchronous motor to be diagnosed, extracting and fusing the characteristics according to the steps S2-S3, and inputting the fused characteristics into a fault diagnosis classifier to obtain a fault diagnosis result.
In a third aspect, the present invention provides a system based on the method for constructing a demagnetization fault diagnosis model of a permanent magnet drive motor or the method for diagnosing a demagnetization fault of a permanent magnet drive motor, including:
the signal acquisition module is used for acquiring/obtaining magnetic leakage signals of the fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
the image conversion module is used for converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
the characteristic extraction module is used for extracting global characteristics and local characteristics of the two-dimensional Fourier spectrogram;
the characteristic fusion module is used for carrying out characteristic fusion on the global characteristic and the local characteristic;
the fault diagnosis classifier building module is used for taking the fused features as the input of the fault diagnosis classifier, taking the fault types as the output of the fault diagnosis classifier, and performing network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
and the diagnosis module is used for diagnosing the fault by utilizing the fault diagnosis classifier and the fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
In a fourth aspect, the present invention provides an electronic terminal, which comprises a processor and a memory connected to each other, wherein the processor is programmed or configured to execute the method for constructing the demagnetization fault diagnosis model of the permanent magnet synchronous motor or the steps of the method for diagnosing the demagnetization fault of the permanent magnet synchronous motor.
In a fifth aspect, the present invention provides a system based on the method for constructing a demagnetization fault diagnosis model of a permanent magnet drive motor or the method for diagnosing a demagnetization fault of a permanent magnet drive motor, where the system is an electric vehicle system, and the system includes: the system comprises a whole vehicle monitoring system, can communication, a vehicle driving control system, a magnetic flux sensor and a driving motor; the driving motor is a permanent magnet driving motor, the whole vehicle monitoring system issues a control instruction to the vehicle driving control system through can communication, and the vehicle driving control system controls the driving motor so as to enable the electric vehicle to operate;
the magnetic flux sensor measures a magnetic flux leakage signal on the surface of the driving motor and transmits the signal to the automobile driving control system, the automobile driving control system is loaded with or calls the fault diagnosis classifier generated by the method, and then the fault diagnosis is carried out by utilizing the fault diagnosis classifier and fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
In a sixth aspect, the present invention provides a readable storage medium, wherein the readable storage medium stores therein a computer program programmed or configured to execute the method for constructing a demagnetization fault diagnosis model of a permanent magnet synchronous motor or the method for diagnosing a demagnetization fault of a permanent magnet synchronous motor.
Advantageous effects
1. The invention directly converts the motor state signal into the two-dimensional Fourier spectrogram, presents more effective fault high-dimensional characteristics and avoids a complex and professional signal processing process.
2. The invention creatively provides a global and local feature fusion extraction method, makes up the defect of single global feature and local feature and improves the diagnosis effect.
3. The invention provides a softmax classifier for demagnetization fault diagnosis, which has only one layer, and has the advantages of high calculation speed, simple structure and strong interpretability.
Drawings
Fig. 1 is a schematic diagram of a demagnetization fault system of a permanent magnet drive motor for an electric vehicle in an embodiment of the invention.
Fig. 2 is a schematic diagram of a demagnetization fault diagnosis method of a permanent magnet driving motor for an electric vehicle based on a visual image according to the method of the embodiment of the invention.
Fig. 3 is a two-dimensional fourier spectrum diagram of three demagnetization-failed motors in the embodiment of the invention.
FIG. 4 is a fused feature map of global features and local features in an embodiment of the present invention.
Fig. 5 is a schematic diagram of softmax in an embodiment of the invention.
Detailed Description
The purpose and effect of the present invention will be more apparent from the following further description of the present invention with reference to the accompanying drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
In this embodiment, the present invention will be described in detail by taking a demagnetization fault diagnosis of a drive motor of an electric vehicle as an example. Fig. 1 is a schematic diagram of a demagnetization fault system of a permanent magnet drive motor for an electric vehicle according to an embodiment of the present invention. As shown in fig. 1, the system includes: the whole vehicle monitoring system 1 issues a control instruction to the vehicle drive control system 3 through can communication 2, and the vehicle drive control system 3 controls the drive motor 5 so as to enable the electric vehicle to operate. For fault diagnosis, the overall monitoring system 1 integrates a fault diagnosis result display, and all algorithms and calculation processes of fault diagnosis are in the automobile drive control system 3. The magnetic flux leakage signal on the surface of the driving motor 5 is measured by the magnetic flux sensor 4, the signal is transmitted to the automobile driving control system 3, the system 3 realizes the motor fault diagnosis by the demagnetization fault diagnosis method of the permanent magnet driving motor for the electric automobile based on the visual image, and the diagnosis result is displayed to a user in the whole automobile monitoring system 1.
Example 1:
as shown in fig. 2, the method and system for diagnosing demagnetization faults of a permanent magnet drive motor based on a visual image applied to an electric vehicle in the embodiment includes three parts, namely signal acquisition, one-dimensional signal to two-dimensional image conversion, image feature extraction of a self-encoder and a softmax classifier. The present embodiment selects to construct the softmax classifier, and it should be understood that the softmax classifier is the best example of the present invention, but the present invention is not limited thereto, and other classification network models may be selected without departing from the concept of the present invention.
The method for diagnosing the demagnetization fault of the permanent magnet driving motor based on the visual image comprises the following steps:
1) and collecting one-dimensional magnetic flux leakage signals of the motor as original signals for fault diagnosis, and converting the original signals into a two-dimensional Fourier spectrogram.
According to the system diagram in fig. 1, magnetic flux leakage signals are acquired for three states of a driving motor of an electric automobile, 2 kinds of demagnetization faults (namely demagnetization fault 1 (demagnetization 30%) and demagnetization fault 2 (demagnetization 100%)) and 1 kind of normal motors occur, and the magnetic flux leakage signals on the surface of the motor are measured by adopting a non-contact alternating-current magnetic sensor at 1000 r/min.
Then, according to two-dimensional Fourier transform, the leakage flux signals of the motor 3 and under various and variable working conditions are converted into a two-dimensional Fourier spectrum image data set for fault diagnosis, so that the expansion from one-dimensional time domain leakage flux signals to two-dimensional images is realized, and fig. 3 is a two-dimensional Fourier spectrum image of variable rotating speed of the driving motor of the electric automobile in three states under no-load.
When the frequency of the acquired leakage magnetic time domain signal is larger than 2 times of the motor operation during data acquisition, N multiplied by N points are taken to form a two-dimensional matrix f (x, y), and then the two-dimensional matrix f (x, y) is converted into a two-dimensional spectrogram H (v, p) through two-dimensional Fourier transform. The two-dimensional Fourier transform formula is as follows:
translating the frequency origin to the center, wherein the translation formula is as follows:
2) and respectively extracting global and local features of the two-dimensional Fourier spectrogram by using an autoencoder and a Scale Invariant Feature Transform (SIFT) method, and performing feature fusion.
In this embodiment, a self-encoder and a Scale Invariant Feature Transform (SIFT) method are adopted to extract global and local features of a two-dimensional fourier spectrogram respectively, and feature fusion is performed. Fig. 4 is a diagram of two feature extractions, the extraction process including:
2.1) extracting global features of the two-dimensional Fourier spectrogram by adopting a self-encoder;
2.2) extracting local features of the two-dimensional Fourier spectrogram by adopting a Scale Invariant Feature Transform (SIFT) method.
2.3) fusing the global feature and the local feature. The fusion technical means has diversity, and the invention does not specifically restrict the fusion technical means.
Regarding the extraction of global features of two-dimensional Fourier spectrogram by using an auto-encoder:
the method adopts a double-layer self-encoder to extract the global characteristics of the two-dimensional Fourier spectrogram, the self-encoder is also a neural network, and the required settings comprise maximum convolution times, L2 network weight regularization parameters, sparse regularization controller parameters, sparse regularization item parameters and whether to stretch data or not.
Extracting local features of a two-dimensional Fourier spectrogram by adopting a Scale Invariant Feature Transform (SIFT) method;
the SIFT method searches for feature points in different scale spaces, the acquisition of the scale spaces needs to be realized by using gaussian blur, and when the two-dimensional template is m × n, the gaussian blur calculation formula of elements I (x, y) in the image is as follows:
where σ is the standard deviation of the normal distribution. The convolution of the blurred Gaussian and the I (x, y) of the original image is the scale space L (x, y, sigma) of the image, and the specific formula is
L(x,y，σ)＝G(x,y,σ)*I(x,y) (4)
Then extreme value detection is carried out to determine characteristic points
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y) (5)
Where k is a calculation constant. In order to make the descriptor rotation invariant, it is necessary to assign a reference direction to each keypoint using local features of the image. The direction is calculated as follows:
θ(x,y)＝tan-1((L(x,y+1)-L(x,y+1))/(L(x+1,y)-L(x-1,y))) (6)
3) and (5) adopting a softmax classifier to perform fault diagnosis. The schematic diagram of softmax is shown in fig. 5, where feature values extracted from the encoder and SIFT are all values subjected to complex weighting and nonlinear processing, the values may be any values, classification probability values are obtained through calculation processing of a softmax classifier, the sum of various probability values is 1, and the calculation formula is:
wherein, yiRepresenting the sofmax classifier input and n its dimensionality. For multi-classification, it is also necessary to determine how close the actual output is to the desired output by a cross-entropy loss function:
wherein t isiRepresenting the true value.
Based on the theoretical statement, a diagnosis model (fault diagnosis classifier) which can be used for realizing demagnetization fault diagnosis of the permanent magnet synchronous motor is constructed, and then based on the fault diagnosis classifier, demagnetization fault diagnosis can be realized on the motor.
4) Acquiring a magnetic leakage signal of the permanent magnet synchronous motor to be diagnosed, extracting and fusing the characteristics according to the modes of 1) -2), and inputting the fused characteristics into a fault diagnosis classifier to obtain a fault diagnosis result.
The present embodiment is an electric vehicle as an example, and it should be understood that the method of the present invention may also be applied to motor leakage fault diagnosis in other fields without departing from the concept of the present invention.
Example 2:
the present embodiment provides a system based on the above fault diagnosis model building method or fault diagnosis method, which includes:
the signal acquisition module is used for acquiring/obtaining magnetic leakage signals of the fault motor under various faults, the magnetic leakage signals are one-dimensional time domain signals, and the signal acquisition module can be realized by a software module, namely the signal acquisition module is used for acquiring the magnetic leakage signals acquired by hardware and can also be realized in a hardware mode, such as a magnetic flux sensor.
The image conversion module is used for converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
the characteristic extraction module is used for extracting global characteristics and local characteristics of the two-dimensional Fourier spectrogram;
the characteristic fusion module is used for carrying out characteristic fusion on the global characteristic and the local characteristic;
the fault diagnosis classifier building module is used for taking the fused features as the input of the fault diagnosis classifier, taking the fault types as the output of the fault diagnosis classifier, and performing network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
and the diagnosis module is used for diagnosing the fault by utilizing the fault diagnosis classifier and the fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
It should be understood that, the specific implementation process of the above unit module refers to the method content, and the present invention is not described herein in detail, and the division of the above functional module unit is only a division of a logic function, and there may be another division manner in the actual implementation, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or may not be executed. Meanwhile, the integrated unit can be realized in a hardware form, and can also be realized in a software functional unit form.
Example 3:
the embodiment provides an electronic terminal, which comprises a processor and a memory which are connected with each other, wherein the processor is programmed or configured to execute the method for constructing the demagnetization fault diagnosis model of the permanent magnet synchronous motor or the steps of the demagnetization fault diagnosis method of the permanent magnet synchronous motor.
When the method for constructing the demagnetization fault diagnosis model of the permanent magnet synchronous motor is executed, the following steps are specifically executed:
step 1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
step 2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
and step 3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
and 4, step 4: and performing network training by using the fusion characteristics of various faults of known fault types to obtain the fault diagnosis classifier.
When the demagnetization fault diagnosis method of the permanent magnet synchronous motor is executed, the following steps are specifically executed:
s1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
s2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
s3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
s4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
s5: and acquiring a magnetic leakage signal of the permanent magnet synchronous motor to be diagnosed, extracting and fusing the characteristics according to the steps S2-S3, and inputting the fused characteristics into a fault diagnosis classifier to obtain a fault diagnosis result.
The specific implementation process of each step refers to the explanation of the foregoing method.
It should be understood that in the embodiments of the present invention, the Processor may be a Central Processing Unit (CPU), and the Processor may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The memory may include both read-only memory and random access memory, and provides instructions and data to the processor. The portion of memory may also include non-volatile random access memory. For example, the memory may also store device type information.
Example 4:
the present embodiment provides a fault diagnosis system, which is an electric vehicle system, and includes: the system comprises a whole vehicle monitoring system, can communication, a vehicle driving control system, a magnetic flux sensor and a driving motor; the driving motor is a permanent magnet driving motor, the whole vehicle monitoring system issues a control instruction to the vehicle driving control system through can communication, and the vehicle driving control system controls the driving motor so as to enable the electric vehicle to operate;
the magnetic flux sensor measures a magnetic flux leakage signal on the surface of the driving motor and transmits the signal to the automobile driving control system, the automobile driving control system loads or calls the fault diagnosis classifier generated by the method, and then the fault diagnosis is carried out by utilizing the fault diagnosis classifier and fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
Example 5:
the present embodiment provides a readable storage medium, wherein the readable storage medium stores a computer program programmed or configured to execute the method for constructing the demagnetization fault diagnosis model of the permanent magnet synchronous motor or the method for diagnosing the demagnetization fault of the permanent magnet synchronous motor.
When the method for constructing the demagnetization fault diagnosis model of the permanent magnet synchronous motor is executed, the following steps are specifically executed:
step 1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
step 2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
and step 3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
and 4, step 4: and performing network training by using the fusion characteristics of various faults of known fault types to obtain the fault diagnosis classifier.
When the demagnetization fault diagnosis method of the permanent magnet synchronous motor is executed, the following steps are specifically executed:
s1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
s2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
s3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
s4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
s5: and acquiring a magnetic leakage signal of the permanent magnet synchronous motor to be diagnosed, extracting and fusing the characteristics according to the steps S2-S3, and inputting the fused characteristics into a fault diagnosis classifier to obtain a fault diagnosis result.
The specific implementation process of each step refers to the explanation of the foregoing method.
The readable storage medium is a computer readable storage medium, which may be an internal storage unit of the controller according to any of the foregoing embodiments, for example, a hard disk or a memory of the controller. The readable storage medium may also be an external storage device of the controller, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the controller. Further, the readable storage medium may also include both an internal storage unit of the controller and an external storage device. The readable storage medium is used for storing the computer program and other programs and data required by the controller. The readable storage medium may also be used to temporarily store data that has been output or is to be output.
Based on such understanding, the technical solution of the present invention essentially or partially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned readable storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. The present application is directed to methods, apparatus (systems), and computer program products according to embodiments of the application, wherein the instructions that execute via the flowcharts and/or processor of the computer program product create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.
Claims (10)
1. A permanent magnet driving motor demagnetization fault diagnosis model construction method based on visual images is characterized by comprising the following steps: the method comprises the following steps: acquiring signals, performing two-dimensional image conversion on the acquired signals, extracting global features and local features of the two-dimensional images, fusing the global features and the local features, and constructing a fault diagnosis classifier by utilizing the fused features, wherein the fault diagnosis classifier specifically comprises the following steps:
step 1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
step 2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
and step 3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
and 4, step 4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
the classifier is applied to demagnetization fault diagnosis of the permanent magnet synchronous motor.
2. The method of claim 1, wherein: and 3, extracting global features and local features of the two-dimensional Fourier spectrogram by adopting a self-encoder and scale-invariant feature conversion respectively.
3. The method of claim 2, wherein: when local features of the two-dimensional Fourier spectrogram are extracted by adopting scale-invariant feature transformation, feature points are searched on different scale spaces;
wherein, the scale space L (x, y, σ) of the image is a convolution of a blurred gaussian and a two-dimensional fourier spectrogram, which is expressed as:
L(x,y，σ)＝G(x,y,σ)*I(x,y)
the feature points are represented as:
D(x,y,σ)＝(G(x,y,kσ)-G(x,y,σ))*I(x,y)
in the formula, D (x, y, σ) is a feature point corresponding to an element I (x, y) on the two-dimensional fourier spectrogram, G (x, y, σ) is a gaussian fuzzy representation of the element I (x, y) on the two-dimensional fourier spectrogram, k is a calculation constant, and σ is a standard deviation of normal distribution;
each feature point is assigned a reference direction, which is expressed as:
θ(x,y)＝tan-1((L(x,y+1)-L(x,y+1))/(L(x+1,y)-L(x-1,y)))
in the formula, θ (x, y) represents a reference direction, and the function L () represents a scale space value.
4. The method of claim 1, wherein: when the magnetic leakage signal is converted into a two-dimensional Fourier spectrogram in the step 2, firstly, N multiplied by N points which are 2 times larger than the running time of a motor during data acquisition are taken from the magnetic leakage signal to form a two-dimensional matrix f (x, y), and then the two-dimensional matrix is converted into a two-dimensional spectrogram H (v, p) through two-dimensional Fourier transform, wherein v and p are respectively a horizontal axis coordinate and a vertical axis coordinate of a frequency domain coordinate in the image two-dimensional spectrogram;
wherein, the two-dimensional Fourier transform formula is as follows:
wherein, translating the frequency origin to the center, the translation formula is:
5. the method of claim 1, wherein: the fault diagnosis classifier is a softmax classifier, the fused features are input into the softmax classifier to obtain classification probability values, the sum of the classification probability values of all types of faults is 1, and the classification probability values are represented as follows:
wherein, yiRepresenting the sofmax neural network classifier input, n representing its dimension, and e being the natural base.
6. A demagnetization fault diagnosis method of a permanent magnet drive motor based on the model construction method of any one of claims 1 to 5 is characterized in that: the method comprises the following steps: the method comprises the following steps of constructing a fault diagnosis classifier and diagnosing faults based on the fault diagnosis classifier, wherein the fault diagnosis classifier specifically comprises the following steps:
s1: collecting magnetic leakage signals of a fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
s2: converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
s3: extracting global features and local features of the two-dimensional Fourier spectrogram, and performing feature fusion;
s4: taking the fused features as the input of a fault diagnosis classifier, taking the fault type as the output of the fault diagnosis classifier, and carrying out network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
s5: and acquiring a magnetic leakage signal of the permanent magnet synchronous motor to be diagnosed, extracting and fusing the characteristics according to the steps S2-S3, and inputting the fused characteristics into a fault diagnosis classifier to obtain a fault diagnosis result.
7. A system based on the method for constructing a model for diagnosing demagnetization faults of a permanent magnet drive motor according to claim 1 or the method for diagnosing demagnetization faults of a permanent magnet drive motor according to claim 6, characterized in that: the method comprises the following steps:
the signal acquisition module is used for acquiring/obtaining magnetic leakage signals of the fault motor under various faults, wherein the magnetic leakage signals are one-dimensional time domain signals;
the image conversion module is used for converting the magnetic leakage signal into a two-dimensional Fourier spectrogram;
the characteristic extraction module is used for extracting global characteristics and local characteristics of the two-dimensional Fourier spectrogram;
the characteristic fusion module is used for carrying out characteristic fusion on the global characteristic and the local characteristic;
the fault diagnosis classifier building module is used for taking the fused features as the input of the fault diagnosis classifier, taking the fault types as the output of the fault diagnosis classifier, and performing network training by using the fused features of various faults of known fault types to obtain the fault diagnosis classifier;
and the diagnosis module is used for diagnosing the fault by utilizing the fault diagnosis classifier and the fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
8. An electronic terminal, characterized by comprising a processor and a memory connected with each other, wherein the processor is programmed or configured to execute the steps of the method for constructing a demagnetization fault diagnosis model of a permanent magnet synchronous motor according to claim 1 or the method for diagnosing a demagnetization fault of a permanent magnet synchronous motor according to claim 6.
9. A system based on the method for constructing a model for diagnosing demagnetization faults of a permanent magnet drive motor according to claim 1 or the method for diagnosing demagnetization faults of a permanent magnet drive motor according to claim 6, characterized in that: the system is an electric vehicle system, comprising: the system comprises a whole vehicle monitoring system, can communication, a vehicle driving control system, a magnetic flux sensor and a driving motor; the driving motor is a permanent magnet driving motor, the whole vehicle monitoring system issues a control instruction to the vehicle driving control system through can communication, and the vehicle driving control system controls the driving motor so as to enable the electric vehicle to operate;
the magnetic flux sensor measures a magnetic flux leakage signal on the surface of the driving motor and transmits the signal to the automobile driving control system, the automobile driving control system loads or calls the fault diagnosis classifier generated by the method of claim 1, and then the fault diagnosis is carried out by utilizing the fault diagnosis classifier and fusion characteristics corresponding to the permanent magnet synchronous motor to be diagnosed to obtain a fault diagnosis result.
10. A readable storage medium, characterized by: the readable storage medium stores therein a computer program programmed or configured to execute the method for constructing a demagnetization fault diagnosis model of a permanent magnet synchronous motor according to claim 1 or the method for diagnosing a demagnetization fault of a permanent magnet synchronous motor according to claim 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111668865.1A CN114282580A (en) | 2021-12-31 | 2021-12-31 | Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111668865.1A CN114282580A (en) | 2021-12-31 | 2021-12-31 | Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114282580A true CN114282580A (en) | 2022-04-05 |
Family
ID=80879405
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111668865.1A Pending CN114282580A (en) | 2021-12-31 | 2021-12-31 | Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114282580A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115931319A (en) * | 2022-10-27 | 2023-04-07 | 圣名科技(广州)有限责任公司 | Fault diagnosis method, fault diagnosis device, electronic equipment and storage medium |
-
2021
- 2021-12-31 CN CN202111668865.1A patent/CN114282580A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115931319A (en) * | 2022-10-27 | 2023-04-07 | 圣名科技(广州)有限责任公司 | Fault diagnosis method, fault diagnosis device, electronic equipment and storage medium |
| CN115931319B (en) * | 2022-10-27 | 2023-10-10 | 圣名科技(广州)有限责任公司 | Fault diagnosis method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sun et al. | Convolutional discriminative feature learning for induction motor fault diagnosis | |
| Li et al. | Scconv: Spatial and channel reconstruction convolution for feature redundancy | |
| CN111860171B (en) | Method and system for detecting irregular-shaped target in large-scale remote sensing image | |
| Zhou et al. | Self-attention feature fusion network for semantic segmentation | |
| Chuang et al. | The diagnosis method of stator winding faults in PMSMs based on SOM neural networks | |
| CN109460738B (en) | A kind of road surface types evaluation method of the depth convolutional neural networks based on free of losses function | |
| Long et al. | Motor fault diagnosis based on scale invariant image features | |
| CN106777442A (en) | A kind of permanent-magnet brushless DC electric machine cogging torque Optimization Design | |
| CN106780582A (en) | Based on the image significance detection method that textural characteristics and color characteristic are merged | |
| CN114282580A (en) | Visual image-based permanent magnet driving motor demagnetization fault diagnosis model construction method and fault diagnosis method and system | |
| Guo et al. | Rolling bearing fault diagnosis method based on attention CNN and BiLSTM network | |
| CN114463341A (en) | Medical image segmentation method based on long and short distance features | |
| Minaz et al. | An effective method for detection of demagnetization fault in axial flux coreless PMSG with texture-based analysis | |
| Li et al. | Diagnosis of rotor demagnetization and eccentricity faults for IPMSM based on deep CNN and image recognition | |
| Tang et al. | Fault diagnosis of hydraulic generator bearing by VMD-based feature extraction and classification | |
| Tang et al. | A small object detection algorithm based on improved faster RCNN | |
| CN111860361B (en) | Automatic identifier and identification method for green channel cargo scanning image entrainment | |
| Wang et al. | First-order differential filtering spectrum division method and information fusion multi-scale network for fault diagnosis of bearings under different loads | |
| CN117332340A (en) | PMSM fault diagnosis method and system based on multi-sensor visual feature fusion | |
| CN107423739A (en) | Image characteristic extracting method and device | |
| Hou et al. | Imbalanced fault identification via embedding-augmented Gaussian prototype network with meta-learning perspective | |
| Shang et al. | Combining support vector machines and information gain ranking for classification of mars mcmurdo panorama images | |
| Zhang et al. | Transcending the limit of local window: Advanced super-resolution transformer with adaptive token dictionary | |
| CN112837420B (en) | Shape complement method and system for terracotta soldiers and horses point cloud based on multi-scale and folding structure | |
| Lin et al. | Object detection in RGB-D images via anchor box with multi-reduced region proposal network and multi-pooling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |