A new energy automobile motor operation state diagnosis method

Abstract

The application discloses a new energy automobile motor operation state diagnosis method and belongs to the motor diagnosis technical field.The application generates a vibration energy signal from a motor vibration signal, obtains short-scale, medium-scale and long-scale energy deviation signals through three different length windows, segments each scale energy deviation signal and carries out EMD decomposition, separates long-term deviation component signals and local deviation fluctuation signals, then extracts combined long-term deviation, combined long-term fluctuation and combined long-term skewness to form a long-term characteristic matrix, extracts combined local impact energy value, combined local impact strength and combined local impact form value to form a local characteristic matrix, finally processes the two types of normalized characteristic matrices through a double-branch fusion convolutional neural network, and outputs a motor operation state diagnosis result.The application improves the motor operation state representation ability and fault diagnosis precision.

Description

Technical Field

[0001] This invention relates to the field of motor diagnostic technology, and specifically to a method for diagnosing the operating status of a new energy vehicle motor. Background Technology

[0002] As the core power source, the drive motor of new energy vehicles operates under complex conditions such as variable speed, variable load, and strong impact for extended periods. Its vibration signals exhibit typical nonlinear, non-stationary, and highly disturbed characteristics. Therefore, vibration signal-based condition monitoring and fault diagnosis are crucial for ensuring driving safety. Current diagnostic methods for motor operation often employ fixed-time-window extraction of vibration energy, combined with Empirical Mode Decomposition (EMD) to separate signal components, and then use single features or conventional neural networks to complete fault identification.

[0003] These methods generally rely on a single scale or a fixed-length analysis window, making it difficult to simultaneously account for the trend-based deviations of motors during long-term operation and the localized impact characteristics caused by short-term faults. The multi-scale dynamic characteristics of the signal cannot be fully characterized. Furthermore, directly performing EMD decomposition on the original signal results in the overlap of long-term trend components and local fluctuation components, failing to achieve effective decoupling and accurate characterization. Diagnostic models with single-feature input struggle to learn the essential differences in faults at different scales, lacking sufficient sensitivity to early, weak, and complex faults, leading to misdiagnosis and missed diagnosis, and failing to meet the high reliability and high precision diagnostic requirements of new energy vehicle drive motors. Therefore, existing technologies suffer from low accuracy in diagnosing motor operating conditions. Summary of the Invention

[0004] In view of the above-mentioned shortcomings in the prior art, the present invention provides a method for diagnosing the operating status of a new energy vehicle motor, which solves the problem of low accuracy in diagnosing the operating status of the motor in the prior art.

[0005] To achieve the above-mentioned objectives, the technical solution adopted by this invention is: a method for diagnosing the operating status of a new energy vehicle motor, comprising the following steps:

[0006] S1. Generate vibration energy signals from motor vibration signals, and obtain short-scale, medium-scale, and long-scale energy deviation signals based on three different window lengths;

[0007] S2. Segment the energy deviation signals at short-scale, medium-scale, and long-scale respectively, and then perform EMD decomposition on each segment to obtain the long-term deviation component signal and the local deviation wave component signal of each segment.

[0008] S3. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviation component signals corresponding to short-scale, medium-scale, and long-scale to obtain the long-term feature matrix.

[0009] S4. Extract the joint local impact energy value, joint local impact intensity, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale to obtain the local feature matrix;

[0010] S5. A dual-branch fusion convolutional neural network is used to process the normalized long-term feature matrix and local feature matrix to obtain the motor operating status diagnosis results.

[0011] Furthermore, S1 includes the following sub-steps:

[0012] S11. Calculate the instantaneous energy value for each sampling point in the motor vibration signal to obtain the vibration energy signal;

[0013] S12. Set the window length to short, medium, and long.

[0014] S13. Taking each sampling point in the vibration energy signal as the center, calculate the average energy of the three window lengths in the neighborhood of the center respectively;

[0015] S14. Subtract the three energy averages from the energy value of each sampling point to obtain the three energy deviation values;

[0016] S15. The energy deviation values ​​of each sampling point belonging to the same window length are used to form the corresponding energy deviation signal, thus obtaining the short-scale, medium-scale, and long-scale energy deviation signals.

[0017] Furthermore, S2 includes the following sub-steps:

[0018] S21. The energy deviation signals of short scale, medium scale and long scale are segmented and processed respectively to obtain multiple short-scale energy deviation sub-signals, multiple medium-scale energy deviation sub-signals and multiple long-scale energy deviation sub-signals.

[0019] S22. Perform EMD decomposition on the short-scale energy deviation sub-signal, the medium-scale energy deviation sub-signal, and the long-scale energy deviation sub-signal respectively to obtain the multi-order intrinsic mode functions and residual terms;

[0020] S23. The residual term is taken as the long-term deviation component of the sub-signal. The values ​​of the 1st, 2nd and 3rd orders in the multi-order intrinsic mode functions are added together according to the sampling points to obtain the local deviation wave sub-signal.

[0021] Furthermore, S3 includes the following sub-steps:

[0022] S31. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviations of the molecular signals corresponding to short-scale, medium-scale, and long-scale.

[0023] S32. Combine the joint long-term deviation, joint long-term volatility, and joint long-term skewness of the same segment to form a long-term feature vector.

[0024] S33. The long-term feature vectors corresponding to each segment are used to form a long-term feature matrix.

[0025] Furthermore, the process of obtaining the joint long-term deviation includes: summing the signal values ​​of each sampling point in the molecular signal for the long-term deviation of each segment to obtain the long-term deviation; and summing the long-term deviations corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term deviation.

[0026] The process of obtaining the joint long-term volatility includes: extracting the standard deviation of the long-term deviation of each segment into a molecular signal; and summing the standard deviations corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term volatility.

[0027] The process of obtaining the joint long-term skewness includes: extracting the skewness from the long-term deviation of each segment into a molecular signal; and summing the skewnesses corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term skewness.

[0028] Furthermore, S4 includes the following sub-steps:

[0029] S41. Extract the joint local impact energy value, joint local impact intensity value, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale.

[0030] S42. Combine the joint local impact energy value, joint local impact intensity, and joint local impact morphology value of the same segment to form a local feature vector;

[0031] S43. The local feature vectors corresponding to each segment are used to form a local feature matrix.

[0032] Furthermore, the process of obtaining the joint local impact energy value includes: setting a signal threshold, filtering out signal values ​​greater than the signal threshold from the local deviation wave sub-signals of each segment to obtain peak values; adding up the peak values ​​corresponding to the local deviation wave sub-signals of each segment to obtain the impact energy value; and adding up the impact energy values ​​corresponding to the short-scale, medium-scale, and long-scale segments within the same segment to obtain the joint local impact energy value.

[0033] The process of obtaining the joint local impact intensity includes: taking the absolute value of each signal value in the local deviation wave sub-signal of each segment to obtain the signal absolute value, filtering the largest signal absolute value to obtain the peak amplitude; and adding the peak amplitudes corresponding to the short-scale, medium-scale and long-scale in the same segment to obtain the joint local impact intensity.

[0034] The process of obtaining the joint local impact morphology value includes: calculating the kurtosis of the local deviation wavelet signal for each segment; and adding the kurtosis corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint local impact morphology value.

[0035] Furthermore, the dual-branch fusion convolutional neural network in S5 includes: a first high-dimensional feature extraction unit, a second high-dimensional feature extraction unit, a channel attention feature fusion unit, a spatial attention feature fusion unit, a first Concat layer, and a classification unit;

[0036] The input of the first high-dimensional feature extraction unit is used to input the normalized long-term feature matrix, and its output is connected to the first input of the channel attention feature fusion unit and the first input of the spatial attention feature fusion unit, respectively.

[0037] The input of the second high-dimensional feature extraction unit is used to input the normalized local feature matrix, and its output is connected to the second input of the channel attention feature fusion unit and the second input of the spatial attention feature fusion unit, respectively.

[0038] The input of the first Concat layer is connected to the output of the channel attention feature fusion unit and the spatial attention feature fusion unit, respectively, and its output is connected to the input of the classification unit.

[0039] The output of the classification unit serves as the output of the dual-branch fusion convolutional neural network.

[0040] Furthermore, S5 includes the following sub-steps:

[0041] S51. The normalized long-term feature matrix is ​​processed by the first high-dimensional feature extraction unit to obtain long-term high-dimensional features.

[0042] S52. The normalized local feature matrix is ​​processed by the second high-dimensional feature extraction unit to obtain local high-dimensional features;

[0043] S53. Channel attention feature fusion unit is used to perform channel attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional channel fusion features;

[0044] S54. Spatial attention feature fusion unit is used to perform spatial attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional spatial fusion features;

[0045] S55. The first Concat layer is used to splice the high-dimensional channel fusion feature and the high-dimensional space fusion feature to obtain the spliced ​​feature;

[0046] S56. Classify the spliced ​​features using classification units to obtain the motor operating status diagnosis results.

[0047] Furthermore, the channel attention feature fusion unit includes: a first channel attention generation module, a second channel attention generation module, a multiplier M1, a multiplier M2, and an adder A1;

[0048] The input of the first channel attention generation module is connected to the first input of multiplier M1 and serves as the first input of the channel attention feature fusion unit; the input of the second channel attention generation module is connected to the first input of multiplier M2 and serves as the second input of the channel attention feature fusion unit; the output of the first channel attention generation module is connected to the second input of multiplier M1; the output of the second channel attention generation module is connected to the second input of multiplier M2; the input of adder A1 is connected to the outputs of multiplier M1 and multiplier M2 respectively, and its output serves as the output of the channel attention feature fusion unit.

[0049] The spatial attention feature fusion unit includes: a first spatial attention generation module, a second spatial attention generation module, a multiplier M3, a multiplier M4, and an adder A2;

[0050] The input terminal of the first spatial attention generation module is connected to the first input terminal of the multiplier M3 and serves as the first input terminal of the spatial attention feature fusion unit; the input terminal of the second spatial attention generation module is connected to the first input terminal of the multiplier M4 and serves as the second input terminal of the spatial attention feature fusion unit; the output terminal of the first spatial attention generation module is connected to the second input terminal of the multiplier M3; the output terminal of the second spatial attention generation module is connected to the second input terminal of the multiplier M4; the input terminal of the adder A2 is connected to the output terminals of the multipliers M3 and M4 respectively, and its output terminal serves as the output terminal of the spatial attention feature fusion unit.

[0051] The beneficial effects of this invention are as follows: By generating energy deviation signals using analysis windows of three different scales (short, medium, and long), this invention can simultaneously consider the long-term trend deviation of the motor and the local impact characteristics caused by short-term faults, thus achieving a comprehensive characterization of the multi-scale dynamic features of the vibration signal. By segmenting the multi-scale signal and then performing EMD decomposition, the overlapping of long-term deviation components and local fluctuation components is avoided, achieving effective decoupling and accurate characterization of the two types of components. On this basis, long-term deviation features and local impact features are extracted separately, which can comprehensively reflect the essential differences of the fault from multiple dimensions, significantly improving the sensitivity to early weak faults and complex faults, and reducing the probability of misjudgment and missed judgment. Furthermore, by using a dual-branch fusion convolutional neural network to deeply fuse and learn the long-term feature matrix and the local feature matrix, the correlation information between features of different scales and types can be fully explored, further improving the accuracy and reliability of fault identification. This improves the overall diagnostic accuracy of the operating status of the drive motor of new energy vehicles, meeting the high reliability and high precision diagnostic requirements under complex working conditions. Attached Figure Description

[0052] Figure 1 A flowchart of a method for diagnosing the operating status of a new energy vehicle motor;

[0053] Figure 2 This is a schematic diagram of the structure of a dual-branch fused convolutional neural network;

[0054] Figure 3 This is a schematic diagram of the structure of the first high-dimensional feature extraction unit and the second high-dimensional feature extraction unit;

[0055] Figure 4 This is a schematic diagram of the structure of the channel attention feature fusion unit;

[0056] Figure 5 This is a schematic diagram of the structure of the first channel attention generation module and the second channel attention generation module;

[0057] Figure 6 This is a schematic diagram of the spatial attention feature fusion unit.

[0058] Figure 7 This is a schematic diagram of the structure of the first spatial attention generation module and the second spatial attention generation module;

[0059] Figure 8 This is a structural diagram of a classification unit. Detailed Implementation

[0060] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0061] like Figure 1 As shown, a method for diagnosing the operating status of a new energy vehicle motor includes the following steps:

[0062] S1. Generate vibration energy signals from motor vibration signals, and obtain short-scale, medium-scale, and long-scale energy deviation signals based on three different window lengths;

[0063] S2. Segment the energy deviation signals at short, medium and long scales respectively, and then perform EMD decomposition on each segment to obtain the long-term deviation component signal and the local deviation wave component signal of each segment.

[0064] S3. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviation component signals corresponding to short-scale, medium-scale, and long-scale to obtain the long-term feature matrix.

[0065] S4. Extract the joint local impact energy value, joint local impact intensity, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale to obtain the local feature matrix;

[0066] S5. A dual-branch fusion convolutional neural network is used to process the normalized long-term feature matrix and local feature matrix to obtain the motor operating status diagnosis results.

[0067] In this embodiment, S1 includes the following sub-steps:

[0068] S11. Calculate the instantaneous energy value for each sampling point in the motor vibration signal to obtain the vibration energy signal;

[0069] S12. Set the window length to short, medium, and long.

[0070] S13. Taking each sampling point of the vibration energy signal as the center, calculate the average energy under three window lengths in its neighborhood to obtain the average energy of the short window, the average energy of the medium window, and the average energy of the long window for the corresponding sampling point.

[0071] S14. Subtract the average short-window energy value of the corresponding sampling point from the energy value of each sampling point to obtain the short-window energy deviation value; subtract the average medium-window energy value of the corresponding sampling point from the energy value of each sampling point to obtain the medium-window energy deviation value; subtract the average long-window energy value of the corresponding sampling point from the energy value of each sampling point to obtain the long-window energy deviation value.

[0072] S15. The short-scale energy deviation signal is composed of the short-window energy deviation values ​​of all sampling points, the medium-scale energy deviation signal is composed of the medium-window energy deviation values ​​of all sampling points, and the long-scale energy deviation signal is composed of the long-window energy deviation values ​​of all sampling points.

[0073] In this embodiment, the formula for calculating the instantaneous energy value is: ,in, The first in the vibration energy signal The instantaneous energy value of each sampling point. The first in the motor vibration signal The signal value at each sampling point This is the number of the sampling point.

[0074] In this embodiment, the short window length is set to 10 sampling points, the medium window length is set to 20 sampling points, and the long window length is set to 30 sampling points. That is, for each sampling point, the average energy within the neighborhood range of length 10, 20, and 30 is obtained respectively.

[0075] This invention calculates the instantaneous energy value of the motor vibration signal point by point, which can transform the original vibration signal into an energy representation that better reflects the degree of abnormal operation and highlights the energy mutation characteristics caused by faults. By using three window lengths (short, medium, and long), the average energy value of the neighborhood is calculated with each sampling point as the center. The energy deviation value is obtained by subtracting the energy value of each sampling point from the average value of the corresponding window. This can effectively eliminate energy fluctuations and random noise interference under normal operating conditions and retain the abnormal deviation components caused by faults. The resulting multi-scale energy deviation signal can comprehensively characterize the changes in the operating state of the motor at different time scales and effectively improve the ability to detect weak faults and early faults.

[0076] In this embodiment, the motor vibration signal is a vibration acceleration signal acquired in real time by a vibration acceleration sensor installed on the housing, end cover, or bearing seat of the new energy vehicle drive motor.

[0077] In this embodiment, S2 includes the following sub-steps:

[0078] S21. The energy deviation signals of short scale, medium scale and long scale are segmented respectively to obtain multiple short-scale energy deviation sub-signals, multiple medium-scale energy deviation sub-signals and multiple long-scale energy deviation sub-signals. The segmentation method of the energy deviation signals of the three scales is the same.

[0079] S22. Perform EMD decomposition on the short-scale energy deviation sub-signal, the medium-scale energy deviation sub-signal, and the long-scale energy deviation sub-signal respectively to obtain the multi-order intrinsic mode functions and residual terms;

[0080] S23. The residual term is taken as the long-term deviation component of the sub-signal. The values ​​of the 1st, 2nd and 3rd orders in the multi-order intrinsic mode functions are added together according to the sampling points to obtain the local deviation wave sub-signal.

[0081] A segment of energy deviation sub-signal corresponds to multiple eigenmode functions and residual terms, and corresponds to a long-term deviation sub-signal and a local deviation wave sub-signal.

[0082] In this embodiment, the energy deviation signals of short scale, medium scale and long scale are segmented respectively. The same segmentation rule is used for the energy deviation signals of the three scales: non-overlapping segmentation with 64 sampling points as a segment, and the starting sampling point of the segmentation of each scale signal is consistent, so as to obtain multiple short-scale energy deviation sub-signals, multiple medium-scale energy deviation sub-signals and multiple long-scale energy deviation sub-signals.

[0083] In this embodiment, the first-order intrinsic mode function, the second-order intrinsic mode function, and the third-order intrinsic mode function belonging to the same energy deviation sub-signal are summed at the same sampling point to obtain the local deviation wave sub-signal.

[0084] This invention reduces the impact of signal non-stationarity and operating condition fluctuations on subsequent decomposition by segmenting the multi-scale energy deviation signal, effectively suppressing mode aliasing. By performing EMD decomposition on each segmented sub-signal, the residual term is used as the long-term deviation component sub-signal, which can accurately characterize the slow deviation and long-term degradation trend during motor operation. The local deviation fluctuation sub-signal is obtained by superimposing the first three intrinsic mode functions, which can highlight the local impact and short-term sudden change characteristics caused by the fault, thereby achieving effective decoupling and accurate characterization of the long-term trend component and the local fluctuation component.

[0085] In this embodiment, S3 includes the following sub-steps:

[0086] S31. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviations of the molecular signals corresponding to short-scale, medium-scale, and long-scale.

[0087] S32. Combine the joint long-term deviation, joint long-term volatility, and joint long-term skewness of the same segment to form a long-term feature vector.

[0088] S33. Take the long-term feature vector corresponding to each segment as a column vector, arrange the columns in chronological order, and form a long-term feature matrix.

[0089] In this embodiment, the size of the long-term feature matrix is ​​3×T, where T is the number of segments.

[0090] In this embodiment, the process of obtaining the joint long-term deviation includes: adding the signal values ​​of each sampling point in the long-term deviation of each segment into the sub-signal to obtain the long-term deviation; and adding the long-term deviations corresponding to the short-scale, medium-scale, and long-scale sub-segments within the same segment to obtain the joint long-term deviation, wherein the same segment refers to the same segment number, that is, adding the long-term deviation of the short-scale sub-signal of the m-th segment, the long-term deviation of the medium-scale sub-signal of the m-th segment, and the long-term deviation of the long-scale sub-signal of the m-th segment into the sub-signal.

[0091] The process of obtaining the joint long-term volatility includes: extracting the standard deviation of the long-term deviation of each segment into a molecular signal; and summing the standard deviations corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term volatility.

[0092] The process of obtaining the joint long-term skewness includes: extracting the skewness from the long-term deviation of each segment into a molecular signal; and summing the skewnesses corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term skewness.

[0093] The formula for calculating the joint long-term deviation is:

[0094] ,

[0095] in, To combine long-term deviation, For the first Long-term deviation of the scale in the molecular signal The signal value at each sampling point The number of sampling points, The numbers 1, 2, and 3 correspond to short-scale, medium-scale, and long-scale, respectively.

[0096] The formula for calculating the combined long-term volatility is:

[0097] ,

[0098] in, To combine long-term volatility, For the first The scale has long deviated from the standard deviation of the molecular signal.

[0099] The formula for calculating the joint long-term skewness is:

[0100] ,

[0101] in, To combine long-term skewness, For the first The skewness of the molecular signal due to long-term deviation at the scale.

[0102] This invention extracts and fuses the sum, standard deviation, and skewness of long-term deviations into molecular signals at short, medium, and long scales, respectively, to construct joint long-term deviation, joint long-term volatility, and joint long-term skewness. This allows for a comprehensive characterization of the slow drift, performance degradation, and trend anomalies of a motor during long-term operation from three dimensions: the degree of amplitude accumulation offset, the degree of data discrete volatility, and the asymmetry of signal distribution. The fusion and superposition of multi-scale features enhances the characterization ability of weak long-term offset faults and improves the stability and discriminative power of the features.

[0103] In this embodiment, S4 includes the following sub-steps:

[0104] S41. Extract the joint local impact energy value, joint local impact intensity value, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale.

[0105] S42. Combine the joint local impact energy value, joint local impact intensity, and joint local impact morphology value of the same segment to form a local feature vector;

[0106] S43. Take the local feature vector corresponding to each segment as a column vector, arrange the columns in chronological order, and form a local feature matrix.

[0107] In this embodiment, the size of the local feature matrix is ​​3×T, where T is the number of segments.

[0108] In this embodiment, the process of obtaining the joint local impact energy value includes: setting a signal threshold, filtering out signal values ​​greater than the signal threshold from the local deviation wave sub-signals of each segment to obtain peak values; adding up the peak values ​​corresponding to the local deviation wave sub-signals of each segment to obtain the impact energy value; and adding up the impact energy values ​​corresponding to the short-scale, medium-scale, and long-scale segments within the same segment to obtain the joint local impact energy value.

[0109] The process of obtaining the joint local impact intensity includes: taking the absolute value of each signal value in the local deviation wave sub-signal of each segment to obtain the signal absolute value, filtering the largest signal absolute value to obtain the peak amplitude; and adding the peak amplitudes corresponding to the short-scale, medium-scale and long-scale in the same segment to obtain the joint local impact intensity.

[0110] The process of obtaining the joint local impact morphology value includes: calculating the kurtosis of the local deviation wavelet signal for each segment; and adding the kurtosis corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint local impact morphology value.

[0111] In this embodiment, the signal threshold is 0.5 to 2 times the standard deviation of the local deviation of the wavelet signal in the current segment.

[0112] The formula for calculating the combined local impact energy value is:

[0113] ,

[0114] in, To combine local impact energy values, For the first The first type of local deviation wave sub-signal with a scale greater than the signal threshold A peak value, This represents the number of peak values ​​exceeding the signal threshold in the local deviation wave sub-signal at the corresponding scale.

[0115] The formula for calculating the combined local impact strength is:

[0116] ,

[0117] in, To combine local impact intensity, In the first The maximum absolute value of the signal is taken from the local deviation wave sub-signal of the various scales.

[0118] The formula for calculating the combined local impact shape value is:

[0119] ,

[0120] in, To combine local impact morphological values, For the first The kurtosis of the local deviation from the wavelet signal at this scale.

[0121] This invention constructs a joint local impact energy value, a joint local impact intensity value, and a joint local impact morphology value by fusing the threshold peak sum, maximum absolute amplitude, and kurtosis of local deviation wave sub-signals at short, medium, and long scales. This allows for the capture of instantaneous impact and local abrupt change characteristics caused by motor faults from three dimensions: impact energy magnitude, impact amplitude strength, and impact distribution morphology. The fusion of multi-scale local features can significantly improve the sensitivity to early weak faults, complex faults, and short-term impact faults, effectively enhancing the identification of fault features.

[0122] In this embodiment, the long-term feature matrix and the local feature matrix are normalized row by row (each row element is divided by the maximum value of that row). The maximum value is found in each row, and normalization is achieved by dividing each row element by the maximum value of the corresponding row.

[0123] like Figure 2 As shown, the dual-branch fusion convolutional neural network in S5 includes: a first high-dimensional feature extraction unit, a second high-dimensional feature extraction unit, a channel attention feature fusion unit, a spatial attention feature fusion unit, a first Concat layer, and a classification unit;

[0124] The input of the first high-dimensional feature extraction unit is used to input the normalized long-term feature matrix, and its output is connected to the first input of the channel attention feature fusion unit and the first input of the spatial attention feature fusion unit, respectively.

[0125] The input of the second high-dimensional feature extraction unit is used to input the normalized local feature matrix, and its output is connected to the second input of the channel attention feature fusion unit and the second input of the spatial attention feature fusion unit, respectively.

[0126] The input of the first Concat layer is connected to the output of the channel attention feature fusion unit and the output of the spatial attention feature fusion unit, respectively, and its output is connected to the input of the classification unit.

[0127] The output of the classification unit serves as the output of the dual-branch fusion convolutional neural network.

[0128] In this embodiment, S5 includes the following sub-steps:

[0129] S51. The normalized long-term feature matrix is ​​processed by the first high-dimensional feature extraction unit to obtain long-term high-dimensional features.

[0130] S52. The normalized local feature matrix is ​​processed by the second high-dimensional feature extraction unit to obtain local high-dimensional features;

[0131] S53. Channel attention feature fusion unit is used to perform channel attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional channel fusion features;

[0132] S54. Spatial attention feature fusion unit is used to perform spatial attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional spatial fusion features;

[0133] S55. The first Concat layer is used to splice the high-dimensional channel fusion feature and the high-dimensional space fusion feature to obtain the spliced ​​feature;

[0134] S56. Classify the spliced ​​features using classification units to obtain the motor operating status diagnosis results.

[0135] like Figure 3 As shown, the first high-dimensional feature extraction unit and the second high-dimensional feature extraction unit have the same structure, both including: a first convolutional layer, a second convolutional layer, a third convolutional layer, a fourth convolutional layer, a fifth convolutional layer, a sixth convolutional layer, a first pooling layer, a second pooling layer, a third pooling layer, and a second Concat layer.

[0136] The input of the first convolutional layer is connected to the input of the second convolutional layer and the input of the third convolutional layer, respectively, and serves as the input of the first high-dimensional feature extraction unit and the second high-dimensional feature extraction unit.

[0137] The first convolutional layer, the fourth convolutional layer, and the first pooling layer are connected in sequence; the second convolutional layer, the fifth convolutional layer, and the second pooling layer are connected in sequence; the third convolutional layer, the sixth convolutional layer, and the third pooling layer are connected in sequence.

[0138] The input of the second Concat layer is connected to the outputs of the first pooling layer, the second pooling layer, and the third pooling layer, respectively, and its output serves as the output of the first high-dimensional feature extraction unit and the second high-dimensional feature extraction unit.

[0139] In this embodiment, the size of the long-term feature matrix and the local feature matrix is ​​3×T. The convolutional kernels of the first, second, and third convolutional layers are all 3×1, with a stride of 1, padding of 1, and 1 output channel, and the output size is 1×T. The convolutional kernel of the fourth convolutional layer is 1×3, with a stride of 1, padding of 1, 8 output channels, and an output size of 8×T. The convolutional kernel of the fifth convolutional layer is 1×5, with a stride of 1, padding of 2, 8 output channels, and an output size of 8×T. The convolutional kernel of the sixth convolutional layer is 1×7, with a stride of 1, padding of 3, 8 output channels, and an output size of 8×T. The pooling window size of the first, second, and third pooling layers is 1×2, with a stride of 2 and an output size of 8×T / 2. The second concat layer concatenates the channels to obtain 24×T / 2.

[0140] This invention employs a three-branch parallel convolutional structure, utilizing temporal convolutional kernels of different sizes (1×3, 1×5, 1×7) to extract multi-scale temporal features from the input features. This enables the accurate capture of local patterns, short-term fluctuations, and long-term dependencies in the input matrix at different temporal resolutions, effectively improving the richness and representational power of the features.

[0141] like Figure 4 As shown, the channel attention feature fusion unit includes: a first channel attention generation module, a second channel attention generation module, a multiplier M1, a multiplier M2, and an adder A1;

[0142] The input of the first channel attention generation module is connected to the first input of multiplier M1 and serves as the first input of the channel attention feature fusion unit; the input of the second channel attention generation module is connected to the first input of multiplier M2 and serves as the second input of the channel attention feature fusion unit; the output of the first channel attention generation module is connected to the second input of multiplier M1; the output of the second channel attention generation module is connected to the second input of multiplier M2; the input of adder A1 is connected to the outputs of multiplier M1 and multiplier M2 respectively, and its output serves as the output of the channel attention feature fusion unit.

[0143] like Figure 5 As shown, the first channel attention generation module and the second channel attention generation module have the same structure, both including the following connected in sequence: a first global average pooling layer, a first fully connected layer and a first Sigmoid layer.

[0144] First global average pooling layer: Performs global average pooling on the input 24×T / 2 features along the temporal dimension, with an output size of 24×1 (maintaining 24 channels, compressing the temporal length to 1). Multiplier M1: Multiplies the first input feature (24×T / 2) element-wise with the channel weights M1 (24×1) along the channel dimension, with an output size of 24×T / 2. Multiplier M2: Multiplies the second input feature (24×T / 2) element-wise with the channel weights M2 (24×1) along the channel dimension, with an output size of 24×T / 2.

[0145] like Figure 6 As shown, the spatial attention feature fusion unit includes: a first spatial attention generation module, a second spatial attention generation module, a multiplier M3, a multiplier M4, and an adder A2;

[0146] The input terminal of the first spatial attention generation module is connected to the first input terminal of the multiplier M3 and serves as the first input terminal of the spatial attention feature fusion unit; the input terminal of the second spatial attention generation module is connected to the first input terminal of the multiplier M4 and serves as the second input terminal of the spatial attention feature fusion unit; the output terminal of the first spatial attention generation module is connected to the second input terminal of the multiplier M3; the output terminal of the second spatial attention generation module is connected to the second input terminal of the multiplier M4; the input terminal of the adder A2 is connected to the output terminals of the multipliers M3 and M4 respectively, and its output terminal serves as the output terminal of the spatial attention feature fusion unit.

[0147] like Figure 7As shown, the first spatial attention generation module and the second spatial attention generation module include: a second global average pooling layer, a global max pooling layer, a third Concat layer, a seventh convolutional layer, and a second Sigmoid layer;

[0148] Second Global Average Pooling Layer: Performs global average pooling along the channel dimension, with an output size of 1×T / 2. Global Max Pooling Layer: Performs global max pooling along the channel dimension, with an output size of 1×T / 2. Third Concat Layer: Concatenates two 1×T / 2 features along the channel dimension, with an output size of 2×T / 2. Seventh Convolutional Layer: Convolutional kernel size is 1×1, output channel is 1, and output size is 1×T / 2. Second Sigmoid Layer: Outputs spatial attention weights, also with a size of 1×T / 2.

[0149] Multiplier M3: Multiplies the first input feature (24×T / 2) element-wise with the spatial weights M3 (1×T / 2), resulting in an output size of 24×T / 2. Multiplier M4: Multiplies the second input feature (24×T / 2) element-wise with the spatial weights M4 (1×T / 2), resulting in an output size of 24×T / 2.

[0150] This invention employs a first high-dimensional feature extraction unit to process the long-term feature matrix and a second high-dimensional feature extraction unit to process the local feature matrix, extracting high-dimensional features. Then, a channel attention feature fusion unit is used to perform channel attention weighting on the long-term high-dimensional features and the local high-dimensional features. In the channel attention feature fusion unit, the first and second channel attention generation modules respectively perform adaptive weighting (M1, M2) on the long-term high-dimensional features and the local high-dimensional features in terms of channel dimension. Then, element-level fusion (A1) is used to obtain high-dimensional channel fusion features, which can automatically focus on key feature channels that are more discriminative for state recognition and suppress redundant channel information.

[0151] In the spatial attention feature fusion unit, the first and second spatial attention generation modules respectively perform adaptive weighting of long-term high-dimensional features and local high-dimensional features in terms of spatial dimension (M3, M4), and then obtain high-dimensional spatial fusion features through element-level fusion (A2), which can accurately locate the spatial location that is more important for fault representation and enhance the feature response of local fault areas.

[0152] Finally, the high-dimensional channel fusion features and high-dimensional spatial fusion features are spliced ​​together through the first Concat layer, realizing dual attention-weighted fusion of the channel dimension and the spatial dimension, which significantly improves the representation ability and recognition of the features, thereby effectively improving the accuracy of motor operation status diagnosis.

[0153] like Figure 8As shown, the classification unit consists of the following sequentially connected layers: the eighth convolutional layer, the fourth pooling layer, the flattening layer, the second fully connected layer, and the classification layer. The eighth convolutional layer has a 1×3 kernel, a stride of 1, padding of 1, 48 output channels, and an output size of 48×T / 2. The fourth pooling layer has a pooling window of 1×2, a stride of 2, padding of 0, and an output size of 48×T / 4. The activation function of the classification layer is Softmax.

[0154] In this embodiment, the motor operating status diagnosis results include: normal status, rotor fault status, stator fault status, and bearing fault status.

[0155] In this embodiment, the dual-branch fused convolutional neural network is trained using the existing gradient descent method.

[0156] This invention generates energy deviation signals using analysis windows of short, medium, and long scales, simultaneously considering both the long-term trend of motor deviation and the local impact characteristics caused by short-term faults, thus achieving a comprehensive characterization of the multi-scale dynamic features of vibration signals. By segmenting the multi-scale signals and then performing EMD decomposition, the overlapping of long-term deviation components and local fluctuation components is avoided, achieving effective decoupling and accurate characterization of the two types of components. Based on this, long-term deviation features and local impact features are extracted separately, which can comprehensively reflect the essential differences of faults from multiple dimensions, significantly improving the sensitivity to early weak faults and complex faults, and reducing the probability of misjudgment and missed judgment. Furthermore, a dual-branch fusion convolutional neural network is used to deeply fuse and learn the long-term feature matrix and the local feature matrix, which can fully explore the correlation information between features of different scales and types, further improving the accuracy and reliability of fault identification. This improves the overall diagnostic accuracy of the operating status of new energy vehicle drive motors, meeting the high reliability and high accuracy diagnostic requirements under complex operating conditions.

[0157] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for diagnosing the operating status of a new energy vehicle motor, characterized in that, Includes the following steps: S1. Generate vibration energy signals from motor vibration signals, and obtain short-scale, medium-scale, and long-scale energy deviation signals based on three different window lengths; S2. Segment the energy deviation signals at short, medium and long scales respectively, and then perform EMD decomposition on each segment to obtain the long-term deviation component signal and the local deviation wave component signal of each segment. S3. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviation component signals corresponding to short-scale, medium-scale, and long-scale to obtain the long-term feature matrix. S4. Extract the joint local impact energy value, joint local impact intensity, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale to obtain the local feature matrix; S5. A dual-branch fusion convolutional neural network is used to process the normalized long-term feature matrix and local feature matrix to obtain the motor operating status diagnosis results. S1 includes the following steps: S11. Calculate the instantaneous energy value for each sampling point in the motor vibration signal to obtain the vibration energy signal; S12. Set the window length to short, medium, and long. S13. Taking each sampling point in the vibration energy signal as the center, calculate the average energy of the three window lengths in the neighborhood of the center respectively; S14. Subtract the three energy averages from the energy value of each sampling point to obtain the three energy deviation values; S15. The energy deviation values ​​of each sampling point belonging to the same window length are used to form the corresponding energy deviation signal, and the short-scale, medium-scale and long-scale energy deviation signals are obtained. S2 includes the following steps: S21. The energy deviation signals of short scale, medium scale and long scale are segmented and processed respectively to obtain multiple short-scale energy deviation sub-signals, multiple medium-scale energy deviation sub-signals and multiple long-scale energy deviation sub-signals. S22. Perform EMD decomposition on the short-scale energy deviation sub-signal, the medium-scale energy deviation sub-signal, and the long-scale energy deviation sub-signal respectively to obtain the multi-order intrinsic mode functions and residual terms; S23. The residual term is taken as the long-term deviation component of the sub-signal. The values ​​of the 1st, 2nd and 3rd orders in the multi-order intrinsic mode functions are added together according to the sampling points to obtain the local deviation wave sub-signal. The process of obtaining the joint long-term deviation includes: summing the signal values ​​of each sampling point in the molecular signal for the long-term deviation of each segment to obtain the long-term deviation; and summing the long-term deviations corresponding to the short-scale, medium-scale, and long-scale signals in the same segment to obtain the joint long-term deviation. The process of obtaining the joint long-term volatility includes: extracting the standard deviation of the long-term deviation of each segment into a molecular signal; and summing the standard deviations corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term volatility. The process of obtaining the joint long-term skewness includes: extracting the skewness from the long-term deviation of each segment into a molecular signal; and summing the skewnesses corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint long-term skewness. The process of obtaining the joint local impact energy value includes: setting a signal threshold, filtering out signal values ​​greater than the signal threshold from the local deviation wave sub-signals of each segment to obtain peak values; adding up the peak values ​​corresponding to the local deviation wave sub-signals of each segment to obtain the impact energy value; and adding up the impact energy values ​​corresponding to the short-scale, medium-scale, and long-scale segments within the same segment to obtain the joint local impact energy value. The process of obtaining the joint local impact intensity includes: taking the absolute value of each signal value in the local deviation wave sub-signal of each segment to obtain the signal absolute value, filtering the largest signal absolute value to obtain the peak amplitude; and adding the peak amplitudes corresponding to the short-scale, medium-scale and long-scale in the same segment to obtain the joint local impact intensity. The process of obtaining the joint local impact morphology value includes: calculating the kurtosis of the local deviation wavelet signal for each segment; and adding the kurtosis corresponding to the short-scale, medium-scale, and long-scale signals within the same segment to obtain the joint local impact morphology value.

2. The method for diagnosing the operating status of a new energy vehicle motor according to claim 1, characterized in that, S3 includes the following steps: S31. Extract the joint long-term deviation, joint long-term volatility, and joint long-term skewness from the long-term deviations of the molecular signals corresponding to short-scale, medium-scale, and long-scale. S32. Combine the joint long-term deviation, joint long-term volatility, and joint long-term skewness of the same segment to form a long-term feature vector. S33. The long-term feature vectors corresponding to each segment are used to form a long-term feature matrix.

3. The method for diagnosing the operating status of a new energy vehicle motor according to claim 1, characterized in that, S4 includes the following steps: S41. Extract the joint local impact energy value, joint local impact intensity value, and joint local impact morphology value from the local deviation wave sub-signals corresponding to short-scale, medium-scale, and long-scale. S42. Combine the joint local impact energy value, joint local impact intensity, and joint local impact morphology value of the same segment to form a local feature vector; S43. The local feature vectors corresponding to each segment are used to form a local feature matrix.

4. The method for diagnosing the operating status of a new energy vehicle motor according to claim 1, characterized in that, The dual-branch fusion convolutional neural network in S5 includes: a first high-dimensional feature extraction unit, a second high-dimensional feature extraction unit, a channel attention feature fusion unit, a spatial attention feature fusion unit, a first concat layer, and a classification unit; The input of the first high-dimensional feature extraction unit is used to input the normalized long-term feature matrix, and its output is connected to the first input of the channel attention feature fusion unit and the first input of the spatial attention feature fusion unit, respectively. The input of the second high-dimensional feature extraction unit is used to input the normalized local feature matrix, and its output is connected to the second input of the channel attention feature fusion unit and the second input of the spatial attention feature fusion unit, respectively. The input of the first Concat layer is connected to the output of the channel attention feature fusion unit and the spatial attention feature fusion unit, respectively, and its output is connected to the input of the classification unit. The output of the classification unit serves as the output of the dual-branch fusion convolutional neural network.

5. The method for diagnosing the operating status of a new energy vehicle motor according to claim 4, characterized in that, S5 includes the following steps: S51. The normalized long-term feature matrix is ​​processed by the first high-dimensional feature extraction unit to obtain long-term high-dimensional features. S52. The normalized local feature matrix is ​​processed by the second high-dimensional feature extraction unit to obtain local high-dimensional features; S53. Channel attention feature fusion unit is used to perform channel attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional channel fusion features; S54. Spatial attention feature fusion unit is used to perform spatial attention weighting on long-term high-dimensional features and local high-dimensional features to obtain high-dimensional spatial fusion features; S55. The first Concat layer is used to splice the high-dimensional channel fusion feature and the high-dimensional space fusion feature to obtain the spliced ​​feature; S56. Classify the spliced ​​features using classification units to obtain the motor operating status diagnosis results.

6. The method for diagnosing the operating status of a new energy vehicle motor according to claim 4 or 5, characterized in that, The channel attention feature fusion unit includes: a first channel attention generation module, a second channel attention generation module, a multiplier M1, a multiplier M2, and an adder A1; The input of the first channel attention generation module is connected to the first input of multiplier M1 and serves as the first input of the channel attention feature fusion unit; the input of the second channel attention generation module is connected to the first input of multiplier M2 and serves as the second input of the channel attention feature fusion unit; the output of the first channel attention generation module is connected to the second input of multiplier M1; the output of the second channel attention generation module is connected to the second input of multiplier M2; the input of adder A1 is connected to the outputs of multiplier M1 and multiplier M2 respectively, and its output serves as the output of the channel attention feature fusion unit. The spatial attention feature fusion unit includes: a first spatial attention generation module, a second spatial attention generation module, a multiplier M3, a multiplier M4, and an adder A2; The input terminal of the first spatial attention generation module is connected to the first input terminal of the multiplier M3 and serves as the first input terminal of the spatial attention feature fusion unit; the input terminal of the second spatial attention generation module is connected to the first input terminal of the multiplier M4 and serves as the second input terminal of the spatial attention feature fusion unit; the output terminal of the first spatial attention generation module is connected to the second input terminal of the multiplier M3; the output terminal of the second spatial attention generation module is connected to the second input terminal of the multiplier M4; the input terminal of the adder A2 is connected to the output terminals of the multipliers M3 and M4 respectively, and its output terminal serves as the output terminal of the spatial attention feature fusion unit.

Images