A method, device and medium for identifying abnormal sound and vibration characteristics of large-capacity propulsion motor
In the recognition of the sound and vibration abnormality characteristic of large-capacity propulsion motor, the fundamental frequency and torque-related frequency of electromagnetic noise are determined by using the rotation speed, the number of rotor slots and the number of poles. Combined with the weighting process of principal component analysis, the electromagnetic noise characteristic vector is extracted, and the current solution is solved. There are some techniques that have low recognition accuracy, which has achieved higher recognition accuracy.
Patent Information
- Application Number
- CN202411252209.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-09-09
AI Technical Summary
In the prior art, in the recognition of acoustic and vibration abnormality characteristics of large-capacity propulsion motors, the recognition accuracy of principal component analysis is reduced due to the same weight of noise, especially when the electromagnetic noise characteristics are not significant enough, it is difficult to accurately express the electromagnetic noise characteristics.
By obtaining the motor speed, number of rotor slots and pole pairs, the fundamental frequency and torque-related frequency of electromagnetic noise are determined, combined with principal component analysis, the sound signal is weighted, the characteristic vector of electromagnetic noise is extracted, and the single classification model is used for identification .
The accuracy of recognition of the acoustic and vibrating abnormal noise characteristics of large-capacity propulsion motors is improved, the influence of other noises is reduced, the electromagnetic noise characteristics are more significant, and the recognition ability of the single-class model is enhanced.
Smart Images

Figure CN118762723B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of sound recognition technology, and in particular to a method, device and medium for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor. Background Art
[0002] Large-capacity propulsion motors refer to motors with higher power and torque, which are designed to drive large equipment or systems, such as ship propulsion, large industrial machinery, etc. Large-capacity propulsion motors will generate various noise signals during operation due to their high power and high load characteristics. These signals include not only the sound generated during normal operation of the motor, but also the sound signals of abnormal vibration and sound.
[0003] In the prior art, the acoustic and vibration characteristics of large-capacity propulsion motors are generally identified by extracting features from the sound signals collected when the large-capacity propulsion motors are running. When using principal component analysis (PCA) to extract features from the collected sound signals, since the collected sound signals contain not only electromagnetic noise but also other noises, all noises have the same weight when traditional principal component analysis is used to extract features. If the features of the electromagnetic noise in the collected sound signals are not significant enough, the feature vector extracted by the principal component analysis cannot accurately express the features of the electromagnetic noise, resulting in a decrease in the accuracy of the acoustic and vibration characteristics identification. Summary of the invention
[0004] In order to solve the above technical problems, the purpose of this application is to provide a method, device and medium for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor. The technical solutions adopted are as follows:
[0005] In a first aspect, an embodiment of the present application provides a method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor, the method comprising the following steps:
[0006] Obtain the sound signals of each segment during the stable operation of a large-capacity propulsion motor, as well as the motor's speed, number of rotor slots, and number of pole pairs;
[0007] Determine the fundamental frequency of the electromagnetic noise based on the rotation speed of the motor and the number of rotor slots; determine the correlation between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain based on the amplitude difference between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain;
[0008] The torque-related frequency is determined based on the number of rotor slots and pole pairs of the motor and the fundamental frequency of the electromagnetic noise; the correlation between the torque-related frequency and any frequency of the sound signal in the frequency domain is determined based on the amplitude difference between the torque-related frequency and any frequency of the sound signal in the frequency domain;
[0009] The amplitude of the frequency of the sound signal in the frequency domain is weighted by using the correlation and association as weights, and the characteristic vector of the electromagnetic noise in the sound signal is obtained by combining the principal component analysis;
[0010] The single classification model is trained using the feature vector of electromagnetic noise, and the trained single classification model is used to output the recognition result of the acoustic and vibration abnormal sound feature in the sound signal of the motor to be identified.
[0011] Preferably, the method for determining the fundamental frequency of the electromagnetic noise is:
[0012] Calculate the product of the number of rotor slots and the speed of the motor as the theoretical fundamental frequency of the electromagnetic noise;
[0013] In the frequency domain of the sound signal, the frequency corresponding to the peak closest to the theoretical fundamental frequency of the electromagnetic noise is taken as the fundamental frequency of the electromagnetic noise.
[0014] Preferably, the method for determining the relevance includes:
[0015] For any frequency of the sound signal in the frequency domain, obtain the amplitude of any frequency;
[0016] Obtain the amplitude of the frequency corresponding to the fundamental frequency of the electromagnetic noise in the frequency domain of the sound signal;
[0017] Respectively obtaining the maximum and minimum values of the amplitude of any frequency and the amplitude of the fundamental frequency of the electromagnetic noise;
[0018] A negative correlation mapping is performed on the modulo result of the maximum value and the minimum value to determine the correlation degree.
[0019] Preferably, the method for determining the torque-related frequency is:
[0020] Obtain the least common multiple of the number of rotor slots and the number of pole pairs of the motor;
[0021] The product of the least common multiple and the fundamental frequency of the electromagnetic noise is taken as the torque related frequency.
[0022] Preferably, the method for determining the degree of association includes:
[0023] Obtain the amplitude of the torque-related frequency corresponding to the frequency domain of the sound signal;
[0024] Calculate the difference between the amplitude of any frequency and the amplitude of the torque-related frequency;
[0025] The correlation degree is determined based on the difference; wherein the correlation degree is negatively correlated with the difference.
[0026] Preferably, the weighted processing method is:
[0027] Calculating a sum of the correlation and the association;
[0028] The product of the sum result and the amplitude of any frequency is used as the weighted amplitude of any frequency.
[0029] Preferably, the specific method of obtaining the characteristic vector of electromagnetic noise in the sound signal by combining principal component analysis includes:
[0030] The weighted amplitudes of all frequencies of the sound signal of each motor in the frequency domain are normalized and used as each row in the feature matrix;
[0031] Obtain multiple eigenvectors after the covariance matrix decomposition of the feature matrix; sort the eigenvectors after the covariance matrix decomposition from large to small according to the contribution rate of the eigenvector;
[0032] The contribution rates of the eigenvectors are added in sequence according to the sorting order until the sum of the additions is greater than a preset threshold. The eigenvector at this time and all the eigenvectors before it are selected as the principal components after principal component analysis of the feature matrix, and are recorded as the eigenvectors of the electromagnetic noise.
[0033] Preferably, the single classification model is trained by using the feature vector of electromagnetic noise, and the trained single classification model is used to output the recognition result of the acoustic and vibration abnormal sound feature in the sound signal of the motor to be identified, specifically:
[0034] The feature vector of electromagnetic noise is used as input to train a single-classification model. There are two types of recognition results for abnormal sound and vibration in the output click sound signal: "normal" and "abnormal sound and vibration".
[0035] The feature vector of the electromagnetic noise in the sound signal of the motor to be identified is used as the input of the trained single-classification model, and the recognition result of the abnormal sound and vibration in the sound signal of the motor is output.
[0036] In a second aspect, an embodiment of the present invention provides a device for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor. The device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor as described in the first aspect.
[0037] In a third aspect, an embodiment of the present invention provides a medium for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor, wherein the medium stores a computer program, and when the computer program is executed by a processor, a method for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor as described in the first aspect is implemented.
[0038] It can be seen from the above embodiments that the method, device and medium for identifying the acoustic and vibration characteristics of a large-capacity propulsion motor provided in the embodiments of the present application have at least the following beneficial effects:
[0039] The present application obtains the fundamental frequency and torque-related frequency of electromagnetic noise respectively through the rotational speed, number of rotor slots and number of pole pairs of a large-capacity propulsion motor during operation, and uses them as reference frequencies to evaluate the frequency performance of the sound signal in the frequency domain; in addition, by combining the fundamental frequency and torque-related frequency of the electromagnetic noise with the amplitude difference between any frequency of the sound signal in the frequency domain, the analysis result is weighted for the frequency amplitude of the collected sound signal in the frequency domain, and the frequency related to the electromagnetic noise in the sound signal is enhanced, making the electromagnetic noise feature more significant; then, the processed sound signal is subjected to principal component analysis to obtain the characteristic vector of the electromagnetic noise, and the characteristic vector of the electromagnetic noise is used to train a single classification model to realize the recognition of the acoustic and vibration characteristics of the large-capacity propulsion motor. The beneficial effect is that the use of principal component analysis can further identify and extract the characteristics of the electromagnetic noise, while reducing the influence of other noises, and can more effectively train the single classification model for the recognition of electromagnetic noise, thereby improving the recognition accuracy of the acoustic and vibration characteristics of the large-capacity propulsion motor. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0041] Figure 1 A flowchart of a method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor provided in one embodiment of the present application;
[0042] Figure 2 A flowchart of a process for obtaining a characteristic vector of electromagnetic noise in a sound signal provided by one embodiment of the present application;
[0043] Figure 3 A flow chart of a method for determining the fundamental frequency of electromagnetic noise provided by one embodiment of the present application;
[0044] Figure 4 A flow chart of a method for determining relevance provided in one embodiment of the present application;
[0045] Figure 5 A flow chart of a method for determining torque-related frequency provided in one embodiment of the present application;
[0046] Figure 6A flow chart of a method for determining a degree of association provided in one embodiment of the present application;
[0047] Figure 7 A flowchart of a specific method for obtaining a characteristic vector of electromagnetic noise in a sound signal by combining principal component analysis provided in one embodiment of the present application;
[0048] Figure 8 A schematic diagram of the structure of a device for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION
[0049] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail a method, device and medium for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor proposed in the present application, as well as its specific implementation method, structure, characteristics and effects. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0050] Unless otherwise specified and limited, terms such as "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such articles or devices. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the article or device including the element. In addition, the term "and\or" used herein includes any and all combinations of one or more related listed items. All technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of this application.
[0051] The following is a detailed description of a method, device and medium for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor provided by the present application in conjunction with the accompanying drawings.
[0052] See also Figure 1 , which shows a flowchart of a method for identifying abnormal sound characteristics of a large-capacity propulsion motor provided by an embodiment of the present application, the method comprising the following steps:
[0053] The first step is to obtain the sound signals of each segment during the stable operation of the large-capacity propulsion motor, as well as the motor's speed, number of rotor slots and number of pole pairs.
[0054] The motor is fixed on the support device to keep the motor stationary. The sound sensor is installed at a distance of 5 cm from the motor axis with an angle of 0°. Each time, the motor is started from rest, and then the sound signal of stable operation 10 seconds after starting is collected. A total of N sound signals of stable operation of the large-capacity propulsion motor are collected. The collection time of each sound signal is 5 seconds, and the sampling frequency is 8kHz. At the same time, the speed n, the number of rotor slots z, and the number of pole pairs p of the motor are obtained.
[0055] In this embodiment, the value of N is 120, and the missing values of the collected large-capacity propulsion motor sound signal are filled using the cubic interpolation method. The cubic interpolation method is a well-known technology, and the specific process will not be repeated here.
[0056] In addition, for the convenience of description, in this application, large-capacity propulsion motor and motor have the same meaning, both referring to large-capacity propulsion motor.
[0057] It should be noted that the various parameters set in the process of obtaining the sound signal of the large-capacity propulsion motor during stable operation can be set by the implementer.
[0058] At this point, the above method can be used to obtain various sound signals of a large-capacity propulsion motor during stable operation, as well as the motor's speed, number of rotor slots, and number of pole pairs.
[0059] The second step is to extract features of the electromagnetic noise in each sound signal segment, and obtain the feature vector of the electromagnetic noise in the sound signal in combination with principal component analysis.
[0060] The collected sound signals of large-capacity propulsion motors mainly include electromagnetic noise, mechanical noise and air noise, among which electromagnetic noise is the main component of the sound signal, which can provide direct information about the electromagnetic field of large-capacity propulsion motors, help detect stator and rotor faults, and is directly related to the internal structure of large-capacity propulsion motors. Therefore, when performing principal component analysis on the collected sound signals of large-capacity propulsion motors, the weight of electromagnetic noise should be greater, so that the extracted characteristic vector of electromagnetic noise can better reflect the information of electromagnetic noise. Among them, principal component analysis is a well-known technology and will not be repeated.
[0061] In an embodiment of the present application, a weighting function is used to weight the electromagnetic noise before performing principal component analysis, emphasizing the frequency components of the electromagnetic noise in the collected sound signal, which can make the characteristics of the electromagnetic noise more significant and facilitate subsequent feature extraction and analysis.
[0062] In the embodiment of the present application, the process flow chart of obtaining the characteristic vector of the electromagnetic noise in the sound signal is as shown in the attached Figure 2 As shown, the specific implementation steps are:
[0063] The first small step is to determine the fundamental frequency of the electromagnetic noise based on the motor speed and the number of rotor slots; based on the amplitude difference between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain, determine the correlation between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain.
[0064] The generation of electromagnetic noise is closely related to electromagnetic vibration, which is excited by the electromagnetic force exerted by the air gap magnetic field inside the motor on the iron core. The air gap magnetic field itself is determined by the magnetomotive force generated by the stator and rotor windings of the motor and the magnetic permeance of the air gap. The electromagnetic force waveform generated by this magnetic field is rotating, including radial and tangential components.
[0065] The radial component is the direction of the electromagnetic force perpendicular to the motor shaft. It acts directly on the stator and rotor of the motor, causing radial vibration of the stator core and rotor. This vibration is the main source of electromagnetic noise in the motor because the vibration will propagate through the motor structure and generate sound waves in the air.
[0066] Motor rotor laminations are divided into odd-number slots and even-number slots. Since odd-number slots allow for a higher slot fill rate, higher power output can be achieved within the same volume. Therefore, large-capacity propulsion motors usually use odd-number slot rotor laminations. The noise caused by this structure is the main part of the electromagnetic noise of large-capacity propulsion motors, that is, during the operation of the motor, the rotor core with odd-number slots is periodically subjected to changes in unilateral magnetic pull.
[0067] Specifically, during the rotation of the rotor, the number of rotor slots in the upper magnetic pole arc and the lower magnetic pole arc do not correspond, which causes the stator core to vibrate up and down periodically. Similarly, the rotor is subjected to a periodically changing unilateral magnetic pull, which causes the rotor to vibrate. Therefore, electromagnetic noise also has periodicity, which is the key to distinguishing it from other noises.
[0068] One of the sound signals during the stable operation of the large-capacity propulsion motor For example, use discrete Fourier transform to transform the sound signal Converting from the time domain to the frequency domain, the spectrum of the sound signal is expressed as .
[0069] The fundamental frequency is a direct reflection of the periodicity of the sound signal. The periodic characteristics of the sound signal can be understood through the fundamental frequency signal in the frequency domain. The electromagnetic noise is caused by the vibration generated by the rotating magnetic field in the motor. The rotation speed of this magnetic field is synchronized with the speed of the motor, and the resulting vibration is related to the number of rotor slots. The noise caused by this unilateral tension imbalance corresponds to the power supply frequency. Therefore, the method for calculating the theoretical fundamental frequency of the theoretical electromagnetic noise is: ; In the formula, It represents the theoretical fundamental frequency of electromagnetic noise, z represents the number of rotor slots of the motor, and n represents the speed of the motor in revolutions per second (r / s).
[0070] Since the fundamental frequency is one of the most significant peaks in the spectrum, it In the frequency domain, select the theoretical fundamental frequency of the electromagnetic noise from the distance The frequency corresponding to the nearest peak is taken as the fundamental frequency of the electromagnetic noise , the base frequency In fact, the fundamental frequency of electromagnetic noise is the base frequency of electromagnetic noise. In particular, since electromagnetic noise dominates the collected sound signal, the calculated fundamental frequency of electromagnetic noise is considered to be the collected sound signal. The fundamental frequency.
[0071] In the embodiment of the present application, the flow chart of the method for determining the fundamental frequency of electromagnetic noise is as shown in the attached Figure 3 shown.
[0072] Based on the above analysis, the sound signal In the frequency domain, the frequency components related to the fundamental frequency are more likely to be periodic signals, while the periodic frequency is more likely to be electromagnetic noise. Therefore, the correlation between the amplitude of each frequency in the frequency domain of the sound signal and the fundamental frequency of the electromagnetic noise can be calculated to characterize the possibility that the frequency is the electromagnetic noise frequency.
[0073] It should be noted that the closer the amplitudes of two frequency components are in the spectrum, the greater the correlation between the two frequencies. In the case of a frequency The closer the amplitude is to the fundamental frequency or the closer it is to the integer multiple of the fundamental frequency, that is, the harmonic, the higher the correlation between the frequency and the fundamental frequency, and the greater the possibility that the frequency is a periodic signal, that is, the closer it is to the electromagnetic noise signal, the greater the weight should be given.
[0074] For sound signals At any frequency in the frequency domain, the correlation between the fundamental frequency of the electromagnetic noise and any frequency is calculated, specifically: for any frequency of the sound signal in the frequency domain, the amplitude of any frequency is obtained; the amplitude of the frequency corresponding to the fundamental frequency of the electromagnetic noise in the frequency domain of the sound signal is obtained; the maximum and minimum values of the amplitude of the any frequency and the amplitude of the fundamental frequency of the electromagnetic noise are respectively obtained; the modulo results of the maximum value and the minimum value are negatively correlated and mapped to determine the correlation.
[0075] Optionally, the negative correlation mapping may be implemented by negative linear mapping, negative exponential mapping, or by setting adjustment parameters.
[0076] In the embodiment of the present application, when calculating the correlation, the negative correlation mapping in the calculation process is performed in the form of an inverse, and the correlation between the fundamental frequency of the electromagnetic noise and the i-th frequency of the sound signal in the frequency domain is , the expression is: ; In the formula, It represents the correlation between the fundamental frequency of electromagnetic noise and the ith frequency of the sound signal in the frequency domain. max represents the maximum value function and min represents the minimum value function. Represents the amplitude of the i-th frequency of the sound signal in the frequency domain, It represents the amplitude of the frequency corresponding to the fundamental frequency of electromagnetic noise in the frequency domain of the sound signal, and mod represents the modulo function.
[0077] In other implementations of the embodiments of the present application, the opposite number of the modulo result may be used as the exponent of an exponential function with a natural constant as the base, and the calculation result of the exponential function may be used as the correlation.
[0078] It should be understood that the greater the correlation between the fundamental frequency of the electromagnetic noise and the i-th frequency of the sound signal in the frequency domain, the greater the correlation between the frequency and the sound signal. The closer the fundamental frequency of the electromagnetic noise is, the more likely it is that the frequency is periodic, and the periodic frequency is the frequency of the electromagnetic noise. Before the principal component analysis, the frequency of the electromagnetic noise is weighted according to the magnitude of the correlation. The greater the correlation, the greater the weight, which makes the principal component more significant and facilitates principal component decomposition; otherwise, there should be a smaller weight.
[0079] In the embodiment of the present application, the flowchart of the method for determining the correlation is as shown in the attached Figure 4 shown.
[0080] The second small step is to determine the torque-related frequency based on the number of rotor slots and pole pairs of the motor and the fundamental frequency of the electromagnetic noise; based on the amplitude difference between the torque-related frequency and any frequency of the sound signal in the frequency domain, determine the correlation between the torque-related frequency and any frequency of the sound signal in the frequency domain.
[0081] The tangential component is the direction of the electromagnetic force along the motor shaft. Its impact on large-capacity propulsion motors is usually not as significant as the radial component, but it also contributes to electromagnetic noise. The tangential component causes the interaction between the permanent magnet and the armature teeth to produce cogging torque. Since the cogging torque changes periodically, torque fluctuations will occur. When the large-capacity propulsion motor is running, the torque fluctuations will affect the smoothness of the motor operation, bring high-order harmonics, and thus generate electromagnetic noise. Therefore, a certain high-order harmonic in the frequency domain is related to the cogging torque.
[0082] Based on this, the embodiment of the present application calculates the high-order harmonics, records them as torque-related frequencies, and then analyzes the relationship between other frequencies and the torque-related frequencies.
[0083] It is known that the cogging torque is only caused by the motor structural parameters. Therefore, the number of rotor slots and the number of pole pairs of the motor can be used as the basis for judging the motor cogging torque. The lowest common multiple of the number of rotor slots and the number of pole pairs is the frequency of the fundamental cogging torque. The greater the frequency of the fundamental cogging torque, the more fluctuations of the fundamental cogging torque per rotor revolution, and the higher the harmonic order.
[0084] Therefore, the high-order harmonics related to the torque fluctuation can be calculated through the structural parameters of the large-capacity propulsion motor, which is recorded as the torque-related frequency. The expression of the torque-related frequency is: ; In the formula, represents the torque-related frequency, z represents the number of rotor slots of the motor, p represents the number of pole pairs of the motor, LCM represents the least common multiple function, Indicates the fundamental frequency of electromagnetic noise.
[0085] It should be understood that the least common multiple of the number of rotor slots and the number of pole pairs represents the impact of the tangential component. The larger its value, the higher the harmonic order. Multiplying it with the fundamental frequency of electromagnetic noise represents the harmonic related to torque, that is, an integer multiple of the fundamental frequency of electromagnetic noise. The larger the torque-related frequency, the higher the order of the high-order harmonics in the electromagnetic noise brought by the tangential component.
[0086] In the embodiment of the present application, the flow chart of the method for determining the torque-related frequency is shown in the attached Figure 5 shown.
[0087] Based on the above analysis, the torque-related frequency is a high-order harmonic in the electromagnetic noise caused by the tangential component, so the frequency that is closer to the torque-related frequency amplitude should be given a greater weight when weighting. However, since the influence of the tangential component is not as great as that of the radial component, when calculating the correlation between the frequency and the torque-related frequency, an exponential function is used to reduce the subsequent weighting weight.
[0088] Therefore, for the sound signal At any frequency in the frequency domain, the correlation between the torque-related frequency and any frequency is calculated, specifically: obtaining the amplitude of the frequency corresponding to the torque-related frequency in the frequency domain of the sound signal; calculating the difference between the amplitude of any frequency and the amplitude of the torque-related frequency; determining the correlation based on the difference; wherein the correlation is negatively correlated with the difference.
[0089] In the embodiment of the present application, the difference between the amplitude of any frequency and the amplitude of the torque-related frequency is: the square of the difference between the amplitude of any frequency and the amplitude of the torque-related frequency. In other implementations of the embodiment of the present application, the absolute value of the difference between the amplitude of any frequency and the amplitude of the torque-related frequency can also be used as the difference between them.
[0090] Optionally, a negative correlation means that the dependent variable will decrease as the independent variable increases, and the dependent variable will increase as the independent variable decreases. This is determined by actual application and is not particularly limited in this application.
[0091] In the embodiment of the present application, when calculating the correlation, the relationship between the correlation and the difference is expressed in the form of a negative exponential function, and the correlation between the fundamental frequency of the electromagnetic noise and the i-th frequency of the sound signal in the frequency domain is , the specific expression is: ; In the formula, represents the correlation between the torque-related frequency and the i-th frequency of the sound signal in the frequency domain, e represents a natural constant, Represents the amplitude of the i-th frequency of the sound signal in the frequency domain, Indicates the amplitude of the torque-related frequency corresponding to the frequency of the sound signal in the frequency domain.
[0092] In other implementations of the embodiments of the present application, the reciprocal of the difference may be used as the correlation, so that there is a negative correlation between the difference and the correlation.
[0093] It should be understood that when calculating the relationship between the current frequency and the torque-related frequency, the closer the current frequency is to the torque-related frequency, the more likely the current frequency is the frequency of electromagnetic noise, and a larger weight should be given when weighting. Because the influence of the tangential component is smaller than that of the radial component, the above calculation method is used; conversely, a smaller weight is used when weighting the current frequency.
[0094] In the embodiment of the present application, the flowchart of the method for determining the degree of association is as shown in the attached Figure 6 shown.
[0095] In the third small step, the frequency amplitude of the sound signal in the frequency domain is weighted using the correlation and association as weights, and the characteristic vector of the electromagnetic noise in the sound signal is obtained by combining the principal component analysis.
[0096] Based on the above analysis, the correlation and association are used as adaptive weights to analyze the collected sound signals of large-capacity propulsion motors. Any frequency in is weighted, and the specific expression is: ; In the formula, Represents the weighted amplitude of the i-th frequency of the sound signal in the frequency domain, Represents the amplitude of the i-th frequency of the sound signal in the frequency domain, Indicates the correlation between the fundamental frequency of electromagnetic noise and the i-th frequency of the sound signal in the frequency domain, It represents the correlation between the torque-related frequency and the i-th frequency of the sound signal in the frequency domain.
[0097] According to the above formula, the N-segment acoustic frequency domain signals collected from the large-capacity propulsion motor are weightedly processed.
[0098] After weighting, the sound signal The amplitude of the frequency related to electromagnetic noise is enhanced, making the characteristics of electromagnetic noise more significant. Principal component analysis can further identify and extract the characteristics of electromagnetic noise and reduce the impact of other noises.
[0099] The frequency domain amplitudes of the sound signals of the N large-capacity propulsion motors after weighted processing are normalized. The embodiment of the present application adopts linear normalization processing, which is a well-known technology and the specific steps are not repeated here. In other implementations of the embodiment of the present application, the maximum and minimum normalization methods and the maximum absolute value normalization methods can also be used for normalization.
[0100] The normalized frequency domain signal is used to construct a feature matrix, where each row is the normalized value of the weighted amplitude of all frequencies in the frequency domain of the sound signal of a large-capacity propulsion motor, and each column is the normalized value of the weighted amplitude of the corresponding frequency in the frequency domain of the sound signal of different large-capacity propulsion motors at the same frequency. For example, the first row and the first column are the normalized values of the weighted amplitude of the first frequency of the first sound signal, and the amplitude is used to express the feature. Then the covariance matrix of the feature matrix is calculated, and the covariance matrix is decomposed to obtain the eigenvector.
[0101] Finally, the contribution rate of each eigenvector is calculated, wherein the calculation of the contribution rate is a well-known technique in principal component analysis and will not be described in detail here.
[0102] It should be understood that the greater the contribution rate, the more information about the original variable is contained in the principal component associated with the eigenvector, that is, the more information about the electromagnetic noise is contained.
[0103] The obtained eigenvectors are sorted from large to small according to the contribution rate, and the contribution rates of these eigenvectors are added in sequence in the sorted order. When the sum of the contribution rates added in sequence is greater than the threshold B, the eigenvector and all the eigenvectors before it are selected as the principal components after the principal component analysis of the feature matrix, and are recorded as the eigenvectors of electromagnetic noise. In this embodiment, the value of the threshold B is 0.8, which can be set by the implementer.
[0104] At this point, the characteristic vector of the electromagnetic noise in the sound signal can be obtained through the above method.
[0105] In the embodiment of the present application, a specific method flow chart of obtaining the characteristic vector of electromagnetic noise in the sound signal by combining principal component analysis is shown in the attached figure. Figure 7 shown.
[0106] The third step is to train a single-classification model using the feature vector of electromagnetic noise, and use the trained single-classification model to output the recognition result of the acoustic and vibration abnormal sound features in the sound signal of the motor to be identified.
[0107] The characteristic vector of electromagnetic noise obtained in the above steps is used as input to train the SVM single classification model (One Class SVM), and the various parameters are set as follows: the kernel function selects the radial basis function, and the kernel function parameter , judgment threshold, i.e. error tolerance , loss function selection -Insensitive loss function, the optimization algorithm selects the gradient descent method to obtain a trained single classification model.
[0108] When using the trained single-classification model to identify abnormal sound and vibration, the sound signal whose deviation from the support vector is less than or equal to the judgment threshold is marked as "normal", and the sound signal whose deviation from the support vector is greater than the judgment threshold is marked as "abnormal sound and vibration".
[0109] In other implementations of the present application, isolation forests can also be used to classify the feature vectors of electromagnetic noise to identify the acoustic vibration characteristics. The implementation process of the SVM single classification model and the isolation forest are both well-known technologies and will not be described in detail.
[0110] The sound signal of the large-capacity propulsion motor to be identified is collected, and the feature vector of the electromagnetic noise in the sound signal is obtained according to the above steps. The feature vector is used as the input of the trained single-classification model, and the recognition result of the abnormal sound and vibration in the sound signal of the motor is output. The recognition result includes two categories: "normal" and "abnormal sound and vibration".
[0111] If the sound signal collected by the single classification model is identified as "abnormal sound and vibration", the motor stator and rotor should be repaired in time.
[0112] Figure 8 A schematic diagram of the structure of a device for identifying the acoustic and vibration characteristics of a large-capacity propulsion motor provided in an embodiment of the present application. Figure 8 As shown, the device of this embodiment includes: at least one processor ( Figure 8Only one is shown), a memory, and a computer program stored in the memory and executable on at least one processor, which implements the steps in any one of the above-mentioned embodiments of the method for identifying the acoustic and vibration characteristics of a large-capacity propulsion motor when executing the computer program.
[0113] The device may include but is not limited to a processor and a memory. Those skilled in the art will appreciate that Figure 8 It is only an example of a device and does not constitute a limitation of the device. The terminal device may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include a network interface, a display screen, and an input device.
[0114] The processor may be a CPU, or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0115] The memory includes a readable storage medium, an internal memory, etc., wherein the internal memory may be the memory of a computer device, and the internal memory provides an environment for the operation of the operating system and computer-readable instructions in the readable storage medium. The readable storage medium may be a hard disk of a computer device, and in other embodiments, it may also be an external storage device of the computer device, for example, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Further, the memory may also include both an internal storage unit of the computer device and an external storage device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of a computer program, etc. The memory may also be used to temporarily store data that has been output or is to be output.
[0116] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned device can refer to the corresponding process in the above-mentioned method embodiment, which will not be repeated here. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned method embodiment can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.
[0117] Based on the same inventive concept as the above method, an embodiment of the present application also provides a medium for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor, wherein the medium stores a computer program, and when the computer program is executed by a processor, the method for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor as described in the first aspect is implemented. Its specific functions and technical effects brought about by it can be found in the method embodiment part, which will not be repeated here.
[0118] The various embodiments in the present application are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
[0119] It should be noted that, unless otherwise specified and limited, terms such as "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a circuit structure, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such article or device. In the absence of further restrictions, an element defined by the sentence "including one..." does not exclude the existence of other identical elements in the article or device including the element. In addition, the term "and\or" used herein includes any and all combinations of one or more related listed items.
[0120] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not invented by the present application.
[0121] It should be understood that the present application is not limited to the exact construction that has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof.
Claims
1. A method for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor, characterized in that: The method comprises the following steps: Obtain the sound signals of each segment during the stable operation of a large-capacity propulsion motor, as well as the motor's speed, number of rotor slots, and number of pole pairs; Determine the fundamental frequency of the electromagnetic noise based on the rotation speed of the motor and the number of rotor slots; determine the correlation between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain based on the amplitude difference between the fundamental frequency of the electromagnetic noise and any frequency of the sound signal in the frequency domain; The torque-related frequency is determined based on the number of rotor slots and pole pairs of the motor and the fundamental frequency of the electromagnetic noise; the correlation between the torque-related frequency and any frequency of the sound signal in the frequency domain is determined based on the amplitude difference between the torque-related frequency and any frequency of the sound signal in the frequency domain; The amplitude of the frequency of the sound signal in the frequency domain is weighted by using the correlation and association as weights, and the characteristic vector of the electromagnetic noise in the sound signal is obtained by combining the principal component analysis; The single classification model is trained using the feature vector of electromagnetic noise, and the trained single classification model is used to output the recognition result of the acoustic and vibration abnormal sound feature in the sound signal of the motor to be identified.
2. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 1, characterized in that: The method for determining the fundamental frequency of the electromagnetic noise is: Calculate the product of the number of rotor slots and the speed of the motor as the theoretical fundamental frequency of the electromagnetic noise; In the frequency domain of the sound signal, the frequency corresponding to the peak closest to the theoretical fundamental frequency of the electromagnetic noise is taken as the fundamental frequency of the electromagnetic noise.
3. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 2, characterized in that: The method for determining the relevance includes: For any frequency of the sound signal in the frequency domain, obtain the amplitude of any frequency; Obtain the amplitude of the frequency corresponding to the fundamental frequency of the electromagnetic noise in the frequency domain of the sound signal; Respectively obtaining the maximum and minimum values of the amplitude of any frequency and the amplitude of the fundamental frequency of the electromagnetic noise; A negative correlation mapping is performed on the modulo result of the maximum value and the minimum value to determine the correlation degree.
4. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 1, characterized in that: The torque-related frequency is determined as follows: Obtain the least common multiple of the number of rotor slots and the number of pole pairs of the motor; The product of the least common multiple and the fundamental frequency of the electromagnetic noise is taken as the torque related frequency.
5. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 4, characterized in that: The method for determining the degree of association includes: Obtain the amplitude of the torque-related frequency corresponding to the frequency domain of the sound signal; Calculate the difference between the amplitude of any frequency and the amplitude of the torque-related frequency; The correlation degree is determined based on the difference; wherein the correlation degree is negatively correlated with the difference.
6. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 1, characterized in that: The weighted processing method is: Calculating a sum of the correlation and the association; The product of the sum result and the amplitude of any frequency is used as the weighted amplitude of any frequency.
7. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 6, characterized in that: The specific method of obtaining the characteristic vector of electromagnetic noise in the sound signal by combining principal component analysis includes: The weighted amplitudes of all frequencies of the sound signal of each motor in the frequency domain are normalized and used as each row in the feature matrix; Obtain multiple eigenvectors after the covariance matrix decomposition of the feature matrix; sort the eigenvectors after the covariance matrix decomposition from large to small according to the contribution rate of the eigenvector; The contribution rates of the eigenvectors are added in sequence according to the sorting order until the sum of the additions is greater than a preset threshold. The eigenvector at this time and all the eigenvectors before it are selected as the principal components after principal component analysis of the feature matrix, and are recorded as the eigenvectors of the electromagnetic noise.
8. A method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as claimed in claim 1, characterized in that: The single classification model is trained by using the characteristic vector of electromagnetic noise, and the trained single classification model is used to output the recognition result of the acoustic and vibration abnormal sound feature in the sound signal of the motor to be recognized, specifically: The feature vector of electromagnetic noise is used as input to train a single-classification model. There are two types of recognition results for abnormal sound and vibration in the output motor sound signal: "normal" and "abnormal sound and vibration". The feature vector of the electromagnetic noise in the sound signal of the motor to be identified is used as the input of the trained single-classification model, and the recognition result of the abnormal sound and vibration in the sound signal of the motor is output.
9. A device for identifying the abnormal sound and vibration characteristics of a large-capacity propulsion motor, characterized in that: The device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a method for identifying abnormal acoustic and vibration characteristics of a large-capacity propulsion motor as described in any one of claims 1 to 8 is implemented.
10. A medium for identifying the characteristics of abnormal sound and vibration of a large-capacity propulsion motor, wherein the medium stores a computer program, characterized in that: When the computer program is executed by a processor, a method for identifying abnormal sound and vibration characteristics of a large-capacity propulsion motor as described in any one of claims 1 to 8 is implemented.
