Transmission system low-quality and few-sample diagnosis method based on diffusion process second-order moment measurement

Abstract

The invention relates to the technical field of intelligent operation and maintenance of machine equipment, in particular to a transmission system low-quality and few-sample diagnosis method based on diffusion process second-order moment measurement, which comprises the following steps: acquiring an electromechanical composite transmission system fault diagnosis data set containing error tags, and randomly extracting samples from the electromechanical composite transmission system fault diagnosis data set to obtain a kernel sample construction set and a network parameter update set; calculating a diffusion process second-order moment measurement matrix; calculating a kernel sample of each health category in the kernel sample construction set; calculating the similarity between a diffusion process second-order moment measurement matrix of the feature vectors of the samples in the network parameter updating set and kernel sample representations of various health categories based on the classifier, and calculating the probability of the health categories to which the samples in the network parameter updating set belong according to the similarity; and constructing a loss function capable of attenuating the influence of the abnormal value, and carrying out iterative training on the feature extractor and the classifier. According to the method, learning can be effectively carried out in a small sample training data set containing error tags, and good diagnosis performance is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent operation and maintenance of machine equipment, and more specifically to a low-quality and small-sample diagnosis method for a transmission system based on second-order moment measurement of a diffusion process. Background Art

[0002] With the increasingly close integration of industry and information technology, modern mechanical equipment is developing towards automation, intelligence and precision. Among them, mechanical transmission systems, as important subsystems of mechanical equipment, are widely used in transportation facilities, electric power, machinery manufacturing and other fields. Their health directly affects the safety and reliability of mechanical equipment. Once a failure occurs, it may lead to economic losses or even casualties. In recent years, accidents caused by mechanical transmission system failures have occurred from time to time. In order to prevent failures and take appropriate maintenance measures, it is necessary to monitor and diagnose the transmission system comprehensively.

[0003] Intelligent diagnosis methods based on deep learning networks have been widely studied due to their excellent end-to-end recognition capabilities, which have promoted the transformation of mechanical transmission systems from "periodic maintenance" to "predictive maintenance". However, intelligent diagnosis methods based on deep learning usually rely on a large number of training samples. The lack of training samples will lead to serious overfitting, which in turn affects the diagnostic performance. In the operation and maintenance of mechanical transmission systems, it is difficult to obtain enough labeled fault samples. In addition, the lack of professional knowledge of staff, information communication errors and human negligence inevitably lead to incorrect labels, which brings great challenges to the engineering application of fault diagnosis models. In the problem of intelligent fault diagnosis of transmission systems under small sample constraints, existing methods only consider the marginal distribution of features to evaluate sample similarity when training the network, ignoring the correlation between different features, and incorrect labels will cause the network's diagnostic performance to drop significantly.

[0004] Therefore, how to improve the diagnostic accuracy of intelligent fault diagnosis algorithms has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a low-quality and small-sample diagnosis method for transmission systems based on the second-order moment measurement of the diffusion process, which can effectively learn in a small sample training data set containing erroneous labels to obtain good diagnostic performance, thereby solving the problem of reduced diagnostic effect due to scarce fault samples and erroneous labels in actual engineering.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] The low-quality and few-sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process includes the following steps:

[0008] Obtain a mechatronic transmission system fault diagnosis dataset D containing wrong labels, randomly extract samples from it to obtain a core sample construction set S and a network parameter update set Q;

[0009] Based on the feature extractor, multi-dimensional features are extracted for each sample in the core sample construction set S and the network parameter update set Q, and the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process are calculated;

[0010] Calculate the core sample P of each healthy category in the core sample construction set S based on the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process;

[0011] Based on the classifier, the second-order moment measurement matrix of the diffusion process of the sample feature vector in the network parameter update set Q is calculated, and the similarity between the kernel samples P of various health categories is calculated. According to the similarity, the probability of the healthy category to which the sample in the network parameter update set Q belongs is calculated;

[0012] A loss function that can attenuate the impact of outliers is constructed, and under the guidance of this loss function, the feature extractor and classifier are iteratively trained.

[0013] Furthermore, the sample representation inside the dataset D is {(X 0 ,Y 0 ),(X 1 ,Y 1 )…(X n ,Y n )}, where any sample Represents the multi-source sensor data of the sample, L represents the time dimension, C represents the channel dimension, and Y a Indicates the corresponding label; X a is a time series, represented by {x 1 ,x 2 ,…,x L}, where any value in the time series is represented by x b,C Represents the data of the Cth channel of the signal at time b.

[0014] Furthermore, N health categories are randomly selected from the data set D, and K samples are randomly selected from each health category to form a core sample construction set S; several additional samples of the same health category as the core sample construction set S are selected from the data set D to form a network parameter update set Q, and the samples in the network parameter update set Q do not overlap with the samples in the core sample construction set S.

[0015] Furthermore, the feature extractor consists of multiple one-dimensional convolutional layers. For any sample X a , its one-dimensional convolution is calculated as follows:

[0016]

[0017] Among them, d is the size of the convolution kernel; C is the total number of channels; i and j are variables in the summation process; w is the weight of the convolution kernel; and b is the bias.

[0018] Input sample X a The multidimensional features captured by the feature extractor F are expressed as in, represents the nonlinear mapping of the feature extractor, represents the learnable parameters in the feature extractor.

[0019] Furthermore, the extracted multidimensional features t a Input to the multidimensional correlation distribution feature extraction layer, which is used to calculate the second-order moment measurement matrix of the diffusion process based on the correlation distribution features; the calculation method of the second-order moment measurement matrix of the diffusion process is as follows:

[0020] t a With L×C dimensions, L represents the time dimension and C represents the channel dimension. The multidimensional feature t is calculated. a The Euclidean space distance matrix of :

[0021]

[0022] in, Represents multidimensional features t a The Euclidean space distance between the i-th eigenvector and the j-th eigenvector in ;

[0023] Calculate multidimensional features t a The second-order moment measurement matrix of the diffusion process is:

[0024]

[0025] Furthermore, for the healthy category c in the core sample construction set S, the calculation formula of its core sample is:

[0026]

[0027] Among them, P c Construct the core samples of healthy category c in the core sample set S; represents the sample X of the health category c a The second-order moment measurement matrix of the diffusion process; Construct sample X in set S for the core sample a The weight of .

[0028] Furthermore, the sample X in the core sample construction set S a Weight The calculation formula is:

[0029]

[0030] in, represents the sample X in the healthy category c in the core sample construction set S a The sum of the similarity of the second-order moment measurement matrix of the diffusion process is calculated as follows:

[0031]

[0032] in, represents the sample X in the healthy category c in the core sample construction set S a and X b The second-order moment of the diffusion process measures the matrix similarity.

[0033] Furthermore, two different samples X in the healthy category c in the core sample construction set S a and X b The calculation process of the second-order moment measurement matrix similarity of the diffusion process is:

[0034] Through the feature extractor, samples X are extracted respectively a and sample X b The multidimensional features of a and t b , and calculate the multidimensional features t a and t b The second-order moment measurement matrix M of the diffusion process a and M b ;

[0035] Calculate two samples X a and X b The second-order moment measurement matrix similarity of the diffusion process is:

[0036] ρ(t a ,t b )=tr(M a T M b )

[0037] Here, tr(·) represents the trace of the matrix.

[0038] Furthermore, the calculation formula for the probability of the healthy category to which the sample in the network parameter update set Q belongs is:

[0039]

[0040] Among them, P(Y i =c) represents sample X i The probability of belonging to health category c; Represents sample X i The core sample P of the healthy category c in the core sample construction set S cThe similarity between them; k is one of the health categories k in N; N is the set of all health categories in the core sample construction set S; For sample X i and the kernel sample P of healthy category k k The similarity between

[0041]

[0042] Among them, M i T Update the network parameters with sample X in set Q i The second-order moment metric matrix of the diffusion process.

[0043] Furthermore, the loss function is expressed as:

[0044]

[0045] in, For data X i Features extracted by the feature extractor; N is the set of all health categories; k is one of the health categories in the N health set; For the sample X i Corresponding label health category Y i Nuclear samples; P k is the core sample of healthy category k.

[0046] θ i is the weight of the loss function, which is used to attenuate the impact of outliers. The weight of the wrong label is attenuated by using the outlier attribute of the wrong label. The weight is calculated as follows:

[0047]

[0048] Among them, sample X i is any sample in the network parameter update set; Y i For sample X i Health category label; Y j For sample X j Health category label; Y i =Y j Represents sample X i and sample X j The health category label is the same, i.e. X j is the network parameter update concentrated health category label Y j With Y i The same any sample; For sample X i With its health category Y i Nuclear samples similarity; For sample Xj With its health category Y j Nuclear samples The similarity.

[0049] It can be seen from the above technical solutions that, compared with the prior art, the present invention has the following beneficial effects:

[0050] 1) The present invention uses a feature extractor to capture the multi-dimensional features of samples, and by calculating the second-order moment measurement matrix of the diffusion process, it fully utilizes the multi-source sensor time series data and enhances the feature extraction capability. By calculating the core samples of the healthy category and predicting and classifying by similarity in the network parameter update set, the design and training process of the traditional classifier is simplified, and the accuracy and efficiency of fault diagnosis are improved.

[0051] 2) The present invention introduces an error label suppression learning strategy, using core samples generated based on similarity weighted aggregation to reduce the negative impact of the wrong label samples in the core sample construction set, and constructs a loss function by attenuating the weight of the wrong label samples in the core sample construction set for the category core samples and the weight of the classification loss of the wrong label samples in the network parameter update set, effectively avoiding the reduction of diagnostic accuracy due to label errors. In general, the present invention can achieve accurate fault diagnosis in a data set containing wrong labels for a mechanical transmission system under small sample constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0053] Figure 1 A flow chart of a low-quality and small-sample diagnosis method for a transmission system based on second-order moment measurement of a diffusion process provided by the present invention;

[0054] Figure 2 The principle diagram of the core sample network learning framework provided by the present invention;

[0055] Figure 3 A schematic diagram of the wrong label suppression learning strategy in the weighted aggregation of core samples provided by the present invention;

[0056] Figure 4 A schematic diagram of the error label suppression learning strategy in the loss function optimization process provided by the present invention;

[0057] Figure 5 This is a schematic diagram showing the accuracy comparison between the method of the present invention and the existing small sample learning method in motor diagnosis;

[0058] Figure 6 This is a schematic diagram showing the comparison of the accuracy of the method of the present invention and the existing small sample learning method in gearbox diagnosis;

[0059] Figure 7 This is a schematic diagram showing the comparison of the accuracy of the method of the present invention and the existing small sample learning method in axle box diagnosis. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] like Figure 1-Figure 2 As shown, the embodiment of the present invention discloses a low-quality and small sample diagnosis method for a transmission system based on the second-order moment measurement of a diffusion process, comprising the following steps:

[0062] S1. Obtain a fault diagnosis dataset D of a mechatronic transmission system containing error labels, and randomly extract samples from it to obtain a core sample construction set S and a network parameter update set Q;

[0063] S2, based on the feature extractor, multi-dimensional feature extraction is performed on each sample in the core sample construction set S and the network parameter update set Q, and the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process are calculated;

[0064] S3, calculating the core sample P of each health category in the core sample construction set S based on the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process;

[0065] S4, based on the classifier, calculate the similarity between the diffusion process second-order moment measurement matrix of the sample feature vector in the network parameter update set Q and various health category core samples P, and calculate the probability of the healthy category to which the sample in the network parameter update set Q belongs according to the similarity;

[0066] S5. Construct a loss function that can attenuate the impact of outliers. Under the guidance of this loss function, iteratively train the feature extractor and classifier.

[0067] The above steps are further explained below.

[0068] S1. Obtain a mechatronic transmission system fault diagnosis dataset D containing erroneous labels, and adopt an erroneous label suppression learning strategy to carry out core sample network training. The network mainly includes a feature extractor and a classifier.

[0069] The sample representation inside the dataset D is {(X0 ,Y 0 ),(X 1 ,Y 1 )…(X n ,Y n )}, where any sample Represents the multi-source sensor data of the sample, L represents the time dimension, C represents the channel dimension, and Y a Indicates the corresponding label; X a is a time series, represented by {x 1 ,x 2 ,…,x L}, where any value in the time series is represented by x b,C Represents the bth data point of the Cth channel of the signal. Different from the general small sample learning task, there are some wrong labels in the dataset D.

[0070] Conduct fault diagnosis network training, randomly select N health categories from the data set D, randomly select K samples in each health category to form a core sample construction set S; select several additional samples of the same health category as the core sample construction set S from the data set D to form a network parameter update set Q, and the samples in the network parameter update set Q do not overlap with the samples in the core sample construction set S.

[0071] S2. The core sample construction set S and the network parameter update set Q are used as the training samples for this round, and are input into the constructed feature extractor F to perform multi-dimensional feature extraction, and calculate the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process.

[0072] S21, the feature extractor is composed of multiple one-dimensional convolutional layers. For any sample X a , its one-dimensional convolution is calculated as follows:

[0073]

[0074] Among them, d is the size of the convolution kernel; C is the total number of channels; i and j are variables in the summation process; w is the weight of the convolution kernel; and b is the bias.

[0075] Input sample X a The multidimensional features captured by the feature extractor F are expressed as in, represents the nonlinear mapping of the feature extractor, represents the learnable parameters in the feature extractor.

[0076] S22, extract the multidimensional features t aInput to the multidimensional correlation distribution feature extraction layer, which is used to calculate the correlation distribution feature matrix based on the second-order moment of the diffusion process; the calculation method of the second-order moment measurement matrix of the diffusion process is as follows:

[0077] Definition a With L×C dimensions, L represents the time dimension and C represents the channel dimension. The multidimensional feature t is calculated. a The Euclidean space distance matrix of :

[0078]

[0079] in, Represents multidimensional features t a The Euclidean space distance between the i-th eigenvector and the j-th eigenvector in ;

[0080] Calculate multidimensional features t a The second-order moment measurement matrix of the diffusion process is:

[0081]

[0082] S3. Calculate the core sample P for each health category in the core sample construction set S, specifically including:

[0083] Calculate the core sample P of each health category in the core sample construction set S. For the health category c, the calculation formula of its core sample is:

[0084]

[0085] Among them, P c Construct the core sample of the healthy category c in the core sample set S, represents the sample X of healthy category c in the core sample construction set S a The second-order moment measurement matrix of the diffusion process is Construct sample X in set S for the core sample a The weight of .

[0086] The determination of is related to improving the robustness of error samples. The outlier properties of the error labels are used to attenuate the weight of the error labels. The principle is as follows: Figure 3 As shown, the sample X in the core sample construction set S a Weight The calculation formula is:

[0087]

[0088] in, represents the sample X in the healthy category c in the core sample construction set S a The sum of the similarity of the second-order moment measurement matrix of the diffusion process is calculated as follows:

[0089]

[0090] in, represents the sample X in the healthy category c in the core sample construction set S a and X b The second-order moment of the diffusion process measures the matrix similarity.

[0091] For two different samples X in the healthy category c in the core sample construction set S a and X b , the calculation process of the second-order moment measurement matrix similarity of the diffusion process is:

[0092] Through the feature extractor, samples X are extracted according to the method of S2. a and sample X b The multidimensional features of a and t b , and calculate the multidimensional features t a and t b The second-order moment measurement matrix M of the diffusion process a and M b ;

[0093] Calculate two samples X a and X b The second-order moment measurement matrix similarity of the diffusion process is:

[0094] ρ(t a ,t b )=tr(M a T M b )

[0095] Among them, tr(·) represents the trace of the matrix. Without loss of generality, for an m-order square matrix G, its trace is calculated as follows:

[0096] g ii Represents the value of the i-th row and i-th column of the square matrix G.

[0097] S4. Classify the samples in the network parameter update set Q into the healthy category with the highest similarity index:

[0098] Calculate the sample feature vectors in the network parameter update set Q and various healthy category core samples P c The similarity between them is used to classify the network parameter update set samples into the most similar categories according to the similarity. The category probability is calculated as follows:

[0099]

[0100] Among them, P(Y i=c) represents sample X i The probability of belonging to health category c; Represents sample X i The core sample P of the healthy category c in the core sample construction set S c The similarity between them; k is the health category k, N is the set of all health categories in the core sample construction set S, For sample X i and the kernel sample P of healthy category k k The similarity between .

[0101]

[0102] Among them, M i T Update the network parameters with sample X in set Q i The second-order moment metric matrix of the diffusion process.

[0103] S5. In order to avoid incorrect network updates caused by mislabeled query samples, a loss function with outlier decay is adopted. The expression of the loss function is:

[0104]

[0105] in, For sample X i The corresponding health category label Y i of nuclear samples; For data X i Features extracted by the feature extractor; N is the set of all health categories; k is one of the health categories in the N health set; P k is the core sample of healthy category k.

[0106] θ i is the weight of the loss function, which is used to attenuate the impact of outliers. It is calculated based on the outlier attenuation strategy. The principle is as follows Figure 4 As shown, specifically: using the outlier properties of the wrong label to attenuate the weight of the wrong label, the calculation method is as follows:

[0107]

[0108] Among them, sample X i is any sample in the network parameter update set; Y i For sample X i Health category label; Y j For sample X j Health category label; Y i =Y j Represents sample X i and sample X jThe health category label is the same, i.e. X j is the network parameter update concentrated health category label Y j With Y i The same any sample; For sample X i With its health category Y i Nuclear samples similarity; For sample X j With its health category Y j Nuclear samples The similarity.

[0109] When the sample X i Health Category Y i When it is wrong, the similarity Compared with the correct samples, the corresponding fraction numerator is small, so the weight is small. According to the above formula, the weight of the sample with wrong label is small, and the weight of the sample with correct label is large. When back-propagating to optimize the parameters, the influence of the wrong label sample can be attenuated.

[0110] Through the above loss function, the classification loss is calculated, and the model parameters are updated using the gradient descent algorithm. The feature extractor and classifier are updated and iterated in multiple small sample learning tasks. In each iteration, new task data is used for training, and the parameters of the feature extractor and classifier are continuously adjusted to improve the diagnostic performance and robustness of the model.

[0111] Next, the performance of the method of the present invention is further verified.

[0112] Taking the fault diagnosis of the electromechanical hybrid transmission system of urban rail trains as an example, the effectiveness of the method of the present invention is verified by the fault simulation data of the test bench. The test bench is designed according to the bogie of the subway train, and the ratio of the test bench to the real bogie is 1:2. The signals collected in the experiment include 18 channels with a sampling frequency of 64kHz, including the three-axis acceleration of the motor, gearbox and axle box and the three-phase current of the motor. In the experiment, 9 working conditions were considered. The inverter controlled different motor speeds to simulate different train speeds, and the electro-hydraulic load device applied different lateral loads to simulate the running state of straight or turning. The experiment simulated 19 different health states of the train transmission system, including motor fault, gearbox fault, axle box fault, etc. The training data set stores 10 samples for each health state, each sample is 0.05s. Among them, 5 samples of each state are randomly selected as the core sample construction set, and the other 5 samples are used as the network parameter update set. They are tested with samples outside the training data set.

[0113] The experiment uses the method proposed in the present invention to diagnose the faults of the motor, gearbox and axle box in the transmission system, and compares the diagnostic accuracy with that of four existing methods.

[0114] Method A: A baseline method that utilizes a typical kernel sample network and does not take special measures to handle mislabeling.

[0115] Method B: The kernel sample network and transfer matrix learning method are combined to achieve few-shot learning and improve robustness to incorrect labels.

[0116] Method C: Generalized cross entropy loss is used to train the kernel sample network to mitigate the impact of incorrect labels.

[0117] Method D: Train the kernel sample network based on the joint teaching strategy to mitigate the impact of label errors.

[0118] The basic structure of the core sample network is shown in Table 1:

[0119] Table 1. Summary of basic network structure

[0120]

[0121]

[0122] The training related parameters are shown in Table 2.

[0123] Table 2. Summary of training related parameters

[0124] Parameter name set up Number of training iterations 500 Learning Rate 0.001 Optimizer Adam

[0125] The experimental results are summarized in Tables 3-5 and Figure 5-Figure 7 .

[0126] Table 3. Summary of the accuracy of the datasets with different proportions of wrong labels in motor diagnosis

[0127]

[0128] Table 4 Summary of the accuracy of data sets with different proportions of wrong labels in gearbox diagnosis

[0129]

[0130]

[0131] Table 5 Summary of the accuracy of the data sets with different proportions of wrong labels in the axle box diagnosis

[0132]

[0133] It can be seen from the experimental results that the suppression of mislabeled samples in the present invention does not rely on large-scale training data sets, but designs specific mechanisms for different mislabeling scenarios, thereby achieving excellent diagnostic performance, which is significantly superior to existing methods.

[0134] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0135] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A low-quality and small sample diagnosis method for transmission systems based on the second-order moment measurement of the diffusion process, characterized in that: The following steps are involved: Obtain a mechatronic transmission system fault diagnosis dataset D containing wrong labels, randomly extract samples from it to obtain a core sample construction set S and a network parameter update set Q; Based on the feature extractor, multi-dimensional features are extracted for each sample in the core sample construction set S and the network parameter update set Q, and the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process are calculated; Calculate the core sample P of each healthy category in the core sample construction set S based on the weight of the sample feature vector and the second-order moment measurement matrix of the diffusion process; Based on the classifier, the second-order moment measurement matrix of the diffusion process of the sample feature vector in the network parameter update set Q is calculated, and the similarity between the kernel samples P of various health categories is calculated. According to the similarity, the probability of the healthy category to which the sample in the network parameter update set Q belongs is calculated; A loss function that can attenuate the impact of outliers is constructed, and under the guidance of this loss function, the feature extractor and classifier are iteratively trained.

2. The low-quality and small sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process according to claim 1 is characterized in that: The sample representation inside the dataset D is {(X0,Y0),(X1,Y1)…(X n ,Y n )}, where any sample Represents the multi-source sensor data of the sample, L represents the time dimension, C represents the channel dimension, and Y a Indicates the corresponding label; X a is a time series, represented by {x1,x2,…,x L }, where any value in the time series is represented by x b,C Represents the data of the Cth channel of the signal at time b.

3. The low-quality and small sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process according to claim 1 is characterized in that: N health categories are randomly selected from the data set D, and K samples are randomly selected from each health category to form a core sample construction set S; several additional samples of the same health category as the core sample construction set S are selected from the data set D to form a network parameter update set Q, and the samples in the network parameter update set Q do not overlap with the samples in the core sample construction set S.

4. The low-quality and small sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process according to claim 1 is characterized in that: The feature extractor consists of multiple one-dimensional convolutional layers. For any sample X a , its one-dimensional convolution is calculated as follows: Among them, d is the size of the convolution kernel; C is the total number of channels; i and j are variables in the summation process; w is the weight of the convolution kernel; and b is the bias. Input sample X a The multidimensional features captured by the feature extractor F are expressed as in, represents the nonlinear mapping of the feature extractor, represents the learnable parameters in the feature extractor.

5. The low-quality and small sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process according to claim 4 is characterized in that: The extracted multidimensional features t a Input to the multidimensional correlation distribution feature extraction layer, which is used to calculate the second-order moment measurement matrix of the diffusion process based on the correlation distribution features; the calculation method of the second-order moment measurement matrix of the diffusion process is as follows: t a With L×C dimensions, L represents the time dimension and C represents the channel dimension. The multidimensional feature t is calculated. a The Euclidean space distance matrix of : in, Represents multidimensional features t a The Euclidean space distance between the i-th eigenvector and the j-th eigenvector in ; Calculate multidimensional features t a The second-order moment measurement matrix of the diffusion process is:

6. The low-quality and small sample diagnosis method for transmission system based on second-order moment measurement of diffusion process according to claim 1 is characterized in that: For the healthy category c in the core sample construction set S, the calculation formula of its core sample is: Among them, P c Construct the core samples of healthy category c in the core sample set S; represents the sample X of healthy category c in the core sample construction set S a The second-order moment measurement matrix of the diffusion process; Construct sample X in set S for the core sample a The weight of .

7. The low-quality and small sample diagnosis method for transmission system based on second-order moment measurement of diffusion process according to claim 6 is characterized in that: Sample X in the core sample construction set S a Weight The calculation formula is: in, represents the sample X in the healthy category c in the core sample construction set S b The sum of the similarity of the second-order moment measurement matrix of the diffusion process; represents the sample X in the healthy category c in the core sample construction set S a The sum of the similarity of the second-order moment measurement matrix of the diffusion process is calculated as follows: in, represents the sample X belonging to the healthy category c in the core sample construction set S a and X b The second-order moment of the diffusion process measures the matrix similarity.

8. The low-quality and small sample diagnosis method for transmission system based on the second-order moment measurement of diffusion process according to claim 7 is characterized in that: Two different samples X in the healthy category c in the core sample construction set S a and X b The calculation process of the second-order moment measurement matrix similarity of the diffusion process is: Through the feature extractor, samples X are extracted respectively a and sample X b The multidimensional features of a and t b , and calculate the multidimensional features t a and t b The second-order moment measurement matrix M of the diffusion process a and M b ; Calculate two samples X a and X b The second-order moment measurement matrix similarity of the diffusion process is: ρ(t a ,t b )=tr(M a T M b ) Here, tr(·) represents the trace of the matrix.

9. The low-quality and small sample diagnosis method for transmission system based on second-order moment measurement of diffusion process according to claim 6 is characterized in that: The calculation formula for the probability of the healthy category to which the sample in the network parameter update set Q belongs is: Among them, P(Y i =c) represents the sample X i The probability of belonging to health category c; For sample X i The core sample P of the healthy category c in the core sample construction set S c The similarity between them; k is one of the health categories k in N; N is the set of all health categories in the core sample construction set S; For sample X i and the kernel sample P of healthy category k k The similarity between Among them, M i T Update the network parameters with sample X in set Q i The second-order moment metric matrix of the diffusion process.

10. The low-quality and small sample diagnosis method for transmission system based on second-order moment measurement of diffusion process according to claim 1, characterized in that: The expression of the loss function is: in, For data X i Features extracted by the feature extractor; N is the set of all health categories; k is one of the health categories in the health set N; For sample X i Corresponding label health category Y i Nuclear samples; P k is the core sample of healthy category k; is the weight of the loss function, which is used to attenuate the impact of outliers. The weight of the wrong label is attenuated by using the outlier attribute of the wrong label. The weight is calculated as follows: Among them, sample X i is any sample in the network parameter update set; Y i For sample X i Health category label; Y j For sample X j Health category label; Y i =Y j Represents sample X i and sample X j The health categories are the same; For sample X i With its health category Y i Nuclear samples similarity; For sample X j With its health category Y j Nuclear samples The similarity.

Images