A rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier difference network

Abstract

The application provides a rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier difference network, and realizes fault diagnosis of rolling bearings under different working conditions. The method constructs a network structure containing a feature extractor and a double classifier, uses a fusion classifier difference index to measure the difference between the output results of the double classifier, and uses a two-stage joint feature alignment strategy to improve the transferability and distinguishability of the extracted features. In the first stage, the maximum mean difference index is used to improve the transferability of the features. On the basis of ensuring the transferability of the features in the first stage, in the second stage, the potential class information of the target domain is fully utilized, and the local maximum mean difference index is used to improve the distinguishability of the fault features. The method simultaneously improves the transferability and distinguishability of the extracted features, ensures the certainty of the output diagnosis results of the classifier, and significantly improves the diagnosis precision of the rolling bearing variable working condition fault diagnosis task.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-domain fault diagnosis of key components of rotating equipment such as rolling bearings, and more specifically, to an unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment maximum classifier difference network. Background Art

[0002] Rolling bearings are a vital component of modern rotating equipment. If damage to a rolling bearing is not promptly detected and repaired, it can trigger a chain reaction of failures, leading to equipment damage or serious safety incidents, resulting in irreparable losses. Therefore, research on rolling bearing fault diagnosis is of great significance. However, due to factors such as speed and load, the operating conditions of rotating equipment are complex and variable. This leads to significant differences in the data distribution between test and training samples, which in turn severely degrades the performance of data-driven diagnostic models. Therefore, complex and variable operating conditions remain a significant challenge for fault diagnosis in real industrial scenarios.

[0003] Domain adaptation is an effective method for fault diagnosis under variable operating conditions. The basic idea of ​​domain adaptation is to transfer knowledge learned from a source domain task to a new, related target domain task to achieve accurate fault diagnosis for target domain samples. However, most domain adaptation methods rely on a large number of labeled target domain samples, which often faces challenges in practical applications.

[0004] Unsupervised domain adaptation effectively addresses the problem of lack of labels for target domain samples, specifically the problem of variable operating condition diagnosis where the source task contains a large number of labeled samples and the target task contains only a small number of unlabeled samples. Unsupervised domain adaptation methods can be broadly categorized into those based on discrepancy metrics and those based on adversarial learning. Discrepancy metric-based methods achieve domain adaptation by comparing the differences between data from different distributions. An et al. designed an improved joint maximum mean discrepancy (JMD) to achieve synchronous alignment of the marginal and conditional probability distributions under unsupervised conditions. Zhou et al. dynamically adjust the marginal and conditional probability distributions of data using weight factors to complete fault diagnosis tasks in the target domain. Adversarial learning-based methods leverage adversarial thinking to improve the model's feature extraction capabilities, thereby obtaining domain-invariant feature representations. Yang et al. proposed a deep adversarial hybrid unsupervised domain adaptation network for fault diagnosis under different operating conditions, improving diagnostic performance. Chen et al. proposed a general framework for fault diagnosis, namely adversarial domain-invariant generalization, which improves diagnostic performance through feature normalization and adaptive weight customization strategies. However, these unsupervised domain adaptation methods do not consider the relationship between target domain samples and classification boundaries when aligning distributions, resulting in a decrease in feature discriminability and the possibility that some samples close to the classification boundary may be misclassified. Adversarial learning methods based on Maximum Classifier Discrepancy (MCD) effectively solve this problem. Saito et al. first proposed a new dual-classifier adversarial strategy to expand the decision boundary of the model by gaming the output differences between the feature extractor and the classifier. Jiao et al. added an adaptive sample selection strategy and softmax cross-entropy loss to the adversarial adaptation mechanism based on the dual-classifier difference metric to improve the accuracy of fault diagnosis. Xu et al. proposed an adversarial consistency learning method to enhance the discriminability of features by minimizing the triple loss function. The MCD-based method expands the decision boundary of the model and achieves feature alignment while maintaining feature discriminability.

[0005] However, these approaches still suffer from the following shortcomings: 1) When acquiring domain-invariant features, most unsupervised domain adaptation fault diagnosis methods, including MCD-based methods, only align the global feature distributions of source and target domain samples, ignoring the differences in feature distributions between samples of the same category in the source and target domains. This leads to misclassification of target domain samples near the classification boundary. 2) MCD-based methods leverage the differences between dual classifiers to learn transferable feature representations, often ignoring the determinism of the target domain classifier outputs, resulting in reduced feature discriminability. Summary of the Invention

[0006] The present invention aims to provide an unsupervised domain-adaptive fault diagnosis method for rolling bearings based on a two-stage joint feature alignment (MCD) network, which is used to effectively perform cross-domain fault diagnosis on rolling bearings under variable operating conditions. To reduce the difference between the source and target domain distributions, a two-stage joint feature alignment strategy is designed to align the global feature distributions of rolling bearing samples collected under different operating conditions and the feature distributions of samples within the same category. In the first stage, the maximum mean difference metric is used to reduce the global feature distribution differences between source and target domain samples, thereby improving the transferability of fault features. While ensuring the transferability of fault features in the first stage, the second stage fully utilizes the latent class information of target domain samples. The network model trained in the first stage is used to obtain target domain sample output probability vectors as pseudo-labels for target domain samples. The local maximum mean difference metric is used to reduce the feature distribution differences between samples within the same category in the source and target domains, thereby improving the discriminability of fault features. To ensure the determinism of the diagnostic results output by the dual classifiers, a fusion classifier difference metric is designed to measure the difference between the dual classifier outputs. This metric comprises a joint difference metric and an individual difference metric, which are used to ensure the consistency and certainty of the final output of the dual classifiers, respectively. The present invention is based on a two-stage joint feature alignment MCD network, which fully utilizes the potential fault category information of the target domain samples, improves the transferability and discriminability of the extracted fault features, and ensures the certainty of the diagnostic results output by the dual classifier, thereby significantly improving the diagnostic accuracy of the rolling bearing variable operating condition fault diagnosis task.

[0007] To achieve the above objectives, the present invention adopts the following technical solutions:

[0008] An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network is characterized by comprising the following steps:

[0009] A1 obtains historical operating data of rolling bearings under different working conditions (different working conditions represent different load conditions), uses the sliding window method to split the original vibration signal into data of the same length and normalizes it, constructs a labeled source domain sample set and an unlabeled target domain sample set (the source domain sample set and the target domain sample set belong to different working conditions respectively), and then divides the unlabeled target domain sample set into a training set and a test set.

[0010] A2 constructs a network structure including a feature extractor and a dual classifier and initializes the network parameters. It uses the fusion classifier difference index to measure the difference between the output results of the dual classifiers and uses it to construct the network loss function.

[0011] A3 adopts a two-stage joint feature alignment strategy to reduce the difference between the source domain and target domain distributions. Each stage consists of three steps, each corresponding to a total loss function, and the overall network, classifier, and feature extractor are trained in sequence. In the first stage, labeled source domain samples and unlabeled target domain training samples are input into the network model, and the maximum mean difference is introduced to reduce the global feature distribution difference between the source and target domain samples. The three steps of this stage are trained repeatedly until the preset maximum number of iterations is reached. In the second stage, labeled source domain samples and unlabeled target domain training samples are input into the network model. Based on the first stage, the latent category information of the target domain samples is fully utilized. The target domain sample output probability vector is obtained through the network model trained in the first stage as the target domain sample pseudo-label. The local maximum mean difference is introduced to reduce the feature distribution difference between the same category samples in the source and target domains. The three steps are trained repeatedly until the maximum number of iterations of the second stage is reached, and training is stopped to generate an MCD network based on two-stage joint feature alignment with good generalization performance.

[0012] A4 uses a trained MCD network based on two-stage joint feature alignment to identify the bearing fault types of target domain test samples under variable working conditions and complete the final fault diagnosis task.

[0013] An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network is characterized in that the specific steps of A1 are as follows:

[0014] A11) Obtain actual historical operating data of the rolling bearing under different operating conditions (different operating conditions represent different load conditions). Use the sliding window method to split the original vibration signal into data of equal length and perform normalization processing. The length of each data is 1024. The normalization processing formula is as follows:

[0015]

[0016] Among them, X is the original data, X norm is the normalized data, X min and X max are the minimum and maximum values ​​in the dataset, respectively.

[0017] A12) Construct a labeled source domain sample set n s represents the number of source domain samples, represents the i-th source domain sample, for Fault type labels; Construct unlabeled target domain sample set represents the jth target domain sample, n TDenotes the number of target domain samples. The source domain sample set and the target domain sample set belong to different working conditions. 80% of the target domain sample set is randomly divided into the training set, and the remaining 20% ​​is divided into the test set.

[0018] An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network is characterized in that the specific steps of A2 are as follows:

[0019] A21) Construct a network structure including a feature extractor and a dual classifier and initialize the network parameters: the feature extractor consists of five convolutional blocks, each convolutional block consists of a convolutional layer, a batch normalization processing layer, a ReLU activation function layer and a maximum pooling layer. The convolution kernel size of the convolutional layer of the first convolutional block is 1×64, and the convolution kernel size of the convolutional layers of the remaining four convolutional blocks is 1×3. The classifier consists of two fully connected layers. The first fully connected layer uses ReLU as the activation function, and the second fully connected layer uses Softmax classifier as the activation function. The structures of the two classifiers are the same.

[0020] A22) The fusion classifier difference index is used to measure the difference between the output results of the two classifiers. The fusion classifier difference index L f Obtained by formula (2):

[0021] L f =μL u +ωL e(2)

[0022] Among them, μ and ω are equilibrium parameters, μ is 1.6, and ω is 1. u is the joint difference index, obtained by formula (3); L e is the individual difference index, which is obtained by formula (4).

[0023]

[0024] Where G represents the feature extractor, C1 and C2 represent two classifiers, c Represents the c-th element of the softmax output.

[0025]

[0026] An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network is characterized in that: in step A3, a two-stage joint feature alignment strategy is adopted to reduce the difference between the source domain and the target domain distribution. Each stage includes three steps, each step corresponds to a total loss function, and the overall network, classifier, and feature extractor are trained in sequence. The specific steps are as follows:

[0027] A31) Input source domain samples, source domain sample labels and target domain training samples, and calculate the first step total loss function. The total loss function is calculated by calculating the cross entropy loss L S and the maximum mean difference index L mmd The sum is obtained as shown in formula (5) and formula (6), and the overall network is trained.

[0028]

[0029] Among them, p(y|x s ) represents the softmax output probability of the source domain sample after the classifier, k∈{1,...,K}, K represents the number of fault categories and K=10, I [k =y s ] represents the indicator vector of the source domain sample label, and E represents the mathematical expectation.

[0030]

[0031] Where H represents the reproducing kernel Hilbert space, and Φ represents the function mapped to the reproducing kernel Hilbert space.

[0032] A32) Input source domain samples, source domain sample labels and target domain training samples, and calculate the second step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the fusion classifier difference index L f The sum is obtained and two classifiers are trained.

[0033] A33) Input the target domain training sample and calculate the total loss function of the third step. The total loss function of this step is calculated by calculating the difference index of the fusion classifier - L f Obtain and train the feature extractor.

[0034] A34) Repeat steps A31) to A33) until the maximum number of iterations of the first stage is reached, which is 15,000.

[0035] A35) Input source domain samples, source domain sample labels and target domain training samples, and calculate the first step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the local maximum mean difference index L lmmd (x s ,x T ) is obtained by summing up, as shown in formula (5) and formula (7), and the feature overall network is trained.

[0036]

[0037] in, and represent source domain samples and target domain samples respectively, and represent the weights of the k-th fault category in the source domain and the target domain respectively. The calculation is shown in formula (8):

[0038]

[0039] in, represents the probability that the sample belongs to fault category k, the source domain sample The kth item of the corresponding one-hot vector label is taken as The target domain samples are collected from unlabeled working conditions, so the kth item of the probability vector output by the network model is used as the target domain sample. of

[0040] A36) Input source domain samples, source domain sample labels and target domain training samples, and calculate the second step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the fusion classifier difference index L f The sum is obtained and two classifiers are trained.

[0041] A37) Input the target domain training sample and calculate the total loss function of the third step. The total loss function of this step is calculated by calculating the difference index of the fusion classifier - L f Obtain and train the feature extractor.

[0042] A38) Repeat steps A35) to A37 until the maximum number of iterations of the second stage is reached, which is 5000. Stop training and generate an MCD network based on two-stage joint feature alignment with good generalization performance.

[0043] An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network is provided, wherein the specific steps of A4 are as follows:

[0044] A41) Inputting the target domain test sample into the trained MCD network based on two-stage joint feature alignment, obtaining the output results of the two classifiers, and averaging the two results to obtain the final target domain test sample diagnosis result.

[0045] Beneficial effects

[0046] The present invention discloses an unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment (MCD) network. A two-stage joint feature alignment strategy is designed to align the global feature distribution of rolling bearing samples collected under different operating conditions with the feature distribution between samples of the same category, while improving the transferability and discriminability of the extracted fault features. In the first stage, the maximum mean difference index is used to reduce the global feature distribution difference between source and target domain samples, thereby improving the transferability of the fault features. Based on the first stage's guarantee of fault feature transferability, the second stage fully utilizes the latent category information of the target domain samples. The network model trained in the first stage is used to obtain the target domain sample output probability vector as the target domain sample pseudo-label. The local maximum mean difference index is then used to reduce the feature distribution difference between the source and target domain samples of the same category, thereby improving the discriminability of the fault features. A fusion classifier difference index is designed to measure the difference between the classifier output results. This index includes a joint difference index and an individual difference index, which are used to ensure the consistency and certainty of the final output results of the dual classifiers, respectively. The present invention makes full use of the latent category information of target domain samples, which not only improves the transferability of the extracted fault features, but also improves the discriminability of the extracted fault features, while ensuring the certainty of the diagnostic results output by the dual classifier. It has the advantages of strong feature extraction ability, strong generalization ability and high diagnostic accuracy, and is of great significance for the unsupervised domain adaptive fault diagnosis of rolling bearings under variable working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Shown is a flow chart of the method of the present invention;

[0048] Figure 2 The figure shows the rolling bearing unsupervised domain adaptation fault diagnosis method based on the two-stage joint feature alignment MCD network of the present invention;

[0049] Figure 3 Shown is a two-stage joint feature alignment strategy. (a) First stage; (b) Second stage;

[0050] Figure 4 Shown is a schematic diagram of the bearing test rig at Case Western Reserve University (CWRU);

[0051] Figure 5 Precision evaluation metrics for different models in the CWRU dataset transfer learning task 3-0;

[0052] Figure 6 Recall evaluation metrics for different models in the CWRU dataset transfer learning task 3-0;

[0053] Figure 7 F1 score evaluation indicators of different models in the CWRU dataset transfer learning task 3-0. DETAILED DESCRIPTION

[0054] In response to the shortcomings of the existing technology, the present invention provides an unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network. This method can effectively align the feature distribution of samples collected under different working conditions of rolling bearings, thereby improving the accuracy of rolling bearing fault diagnosis under variable working conditions.

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0056] Reference Figure 2 The present invention provides an unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment MCD network, which overcomes the problem of difficulty in achieving accurate rolling bearing fault diagnosis under variable working conditions. The proposed two-stage joint feature alignment strategy can better reduce the difference between the source domain and the target domain distribution, such as Figure 3 As shown in Figure 2, a classifier difference metric, called the fusion classifier difference metric, is designed to measure the differences between classifiers and used to construct a loss function based on a two-stage joint feature alignment MCD network. Ultimately, this approach extracts highly transferable and discriminative features from both source and target domain samples, maintaining the accuracy of rolling bearing fault diagnosis under varying operating conditions.

[0057] The Case Western Reserve University (CWRU) public bearing dataset is widely used to verify the performance of fault diagnosis. Figure 4 The CWRU bearing test rig is shown. It consists of two electric motors, a torque sensor, a power meter, and other control devices. The dataset contains four health conditions: normal (N), roller element fault (RF), inner raceway fault (IF), and outer raceway fault (OF). Furthermore, each fault type has three different damage levels on the bearing: 0.007 inches, 0.014 inches, and 0.021 inches. In this experiment, a drive-end bearing dataset with a sampling frequency of 12 kHz was selected. The experimental platform operated under four different load conditions: 0 HP, 1 HP, 2 HP, and 3 HP. Ten fault types were selected. The four different operating conditions are shown in Table 1. For each experiment, two of the four conditions were selected as the source and target domains. For example, Task 1-2 indicates that the source and target domain samples were collected under operating conditions 1 and 2, respectively. To reduce the impact of experimental randomness, each task was repeated 10 times to obtain the average diagnostic accuracy.

[0058] For hyperparameter optimization, the SGD optimizer was used. The batch size of the training model was set to 1024, the initial learning rate was set to 0.01, and a dynamic learning rate update technique was employed. The model had 15,000 iterations in the first phase and 5,000 in the second phase. Regarding the model structure, a network architecture consisting of a feature extractor and a dual classifier was constructed. The feature extractor consists of five convolutional blocks, each consisting of a convolutional layer, a batch normalization layer, a ReLU activation function layer, and a max pooling layer. The classifier consists of two fully connected layers. The first fully connected layer uses ReLU as the activation function, and the second fully connected layer uses Softmax as the activation function. Both classifiers have the same structure.

[0059] Based on the above description and in accordance with the invention, the specific implementation process is as follows:

[0060] 1) Obtain the actual operating history data of rolling bearings under different working conditions (different working conditions represent different load conditions), use the sliding window method to split the original vibration signal into data of the same length and perform normalization processing to construct a labeled source domain sample set n s represents the number of source domain samples, represents the i-th source data sample, for Fault labels; Construct unlabeled target domain sample set represents the jth target domain sample, n T Denotes the number of target domain samples. The source domain sample set and the target domain sample set belong to different working conditions. 80% of the target domain sample set is randomly divided into the training set, and the remaining 20% ​​is divided into the test set.

[0061] 2) Input source domain samples, source domain sample labels and target domain training samples, and calculate the total loss function L of the network at this time using formula (5) and formula (6): s (x s ,y s )+L mmd (x s ,x T ), train the overall network.

[0062] 3) Input source domain samples, source domain sample labels and target domain training samples, and calculate the total loss function L of the network at this time using formula (5) and formula (2) s (x s ,y s )+L f , train two classifiers.

[0063] 4) Input the target domain training sample and calculate the total loss function of the network at this time -L by formula (2) f , train the feature extractor.

[0064] 5) Repeat steps 2) to 4) until the maximum number of iterations of the first stage is reached. The number of iterations of the first stage is 15,000.

[0065] 6) Input source domain samples, source domain sample labels and target domain training samples, and calculate the total loss function L of the network at this time using formula (5) and formula (7) s (x s ,y s )+L lmmd (x s ,x T ), train the overall network.

[0066] 7) Input source domain samples, source domain sample labels and target domain training samples, and calculate the total loss function L of the network at this time using formula (5) and formula (2) s (x s ,y s )+L f , train two classifiers.

[0067] 8) Input the target domain training sample and calculate the total loss function of the network at this time -L by formula (2) f , train the feature extractor.

[0068] 9) Repeat steps 6) to 8) until the maximum number of iterations of the second stage is reached, which is 5000. Stop training and generate an MCD network based on two-stage joint feature alignment with good generalization performance.

[0069] 10) Input the target domain test sample into the MCD network based on two-stage joint feature alignment to obtain the output results of the two classifiers. The two results are averaged to obtain the final target domain test sample diagnosis result.

[0070] To fully demonstrate the effectiveness of the proposed model, five domain adaptation methods were selected for comparison. For methods based on dissimilarity metrics, the Deep Domain Adaptation Network (DAN), which uses the maximum mean difference metric, and the CORAL method, which uses correlation matching, were selected. For methods based on domain adversarial learning, the Deep Adversarial Network (DANN) was selected. For methods based on dual-classifier adversarial learning, the MCD method was selected. Furthermore, the local alignment-based domain adaptation method CMUTLN was selected.

[0071] In order to comprehensively evaluate the performance of the proposed fault diagnosis method, accuracy, precision, recall and F1-score are used as indicators to evaluate the fault diagnosis performance. These four indicators are defined as:

[0072]

[0073]

[0074]

[0075]

[0076] Among them, TP, FP, FN and TN represent the positive samples predicted by the model as positive, the negative samples predicted by the model as positive, the positive samples predicted by the model as negative and the negative samples predicted by the model as negative, respectively.

[0077] A total of 12 groups of transfer learning tasks were set up in this experiment, and the diagnosis results are shown in Table 2. The present invention has achieved relatively high diagnostic accuracy in almost all transfer learning tasks. Given 12 groups of transfer learning tasks, the average diagnostic accuracy reached 98.78%, and fault diagnosis under variable working conditions has a high diagnostic accuracy. This is mainly due to the reduction of the feature distribution differences between samples of the same category in the source domain and the target domain, thereby making the extracted features more discriminative. However, for methods such as CORAL, DAN, DANN and MCD that only perform global feature distribution alignment, the diagnostic results are not ideal because they do not take into account the feature distribution differences between samples of the same category in the source domain and the target domain. In addition, the diagnostic result of the worst-performing method on the most difficult task was 88.54%, while that of the present invention was 95.76%.

[0078] In order to further demonstrate the superiority of the present invention, the performance of the proposed model and the comparison model was evaluated using standard indicators such as precision, recall and F1 score. The precision, recall and F1 evaluation indicators of different models in task 3-0 on the CWRU dataset are as follows: Figure 5 、 Figure 6 and Figure 7 shown.

[0079] Across different fault types, CORAL and DAN generally achieve low F1 scores. This is because unsupervised domain adaptation fault diagnosis methods based on difference metrics focus solely on the global feature distribution differences between source and target domain samples, resulting in poor discriminability of the extracted domain-invariant features. For adversarial unsupervised domain adaptation fault diagnosis methods such as DANN and MCD, although DANN adds a domain classifier and MCD uses dual classifiers through a special network structure to better extract domain-invariant features, they also focus solely on the global feature distribution differences between source and target domain samples, ignoring the potential fault category information of target domain samples. This results in poor discriminability of the extracted fault features and, in turn, incorrect fault diagnosis results. CMUTLN uses a local alignment method for domain adaptation, but because the classifier difference metric cannot ensure the determinism of the classifier's final output, some target domain samples near the classification boundary will be misclassified. In contrast, the method proposed in this paper not only reduces the feature distribution differences between the same-category samples in the source and target domains, but also ensures the determinism of the final output, demonstrating good performance under the F1 metric.

[0080] Finally, it can be concluded that the present invention makes full use of the potential category information of the target domain samples, which not only improves the transferability of the extracted fault features, but also improves the discriminability of the extracted fault features, while ensuring the certainty of the diagnostic results output by the classifier, thereby significantly improving the diagnostic performance indicators of the rolling bearing variable working condition fault diagnosis task, and has a great beneficial effect on rolling bearing fault diagnosis.

[0081] Table 1 Four different working conditions on the CWRU bearing dataset

[0082]

[0083] Table 2 Diagnostic accuracy of each model on the CWRU bearing dataset

[0084]

Claims

1. An unsupervised domain adaptation fault diagnosis method for rolling bearings based on a two-stage joint feature alignment maximum classifier difference network, characterized by: The following steps are involved: A1 obtains historical operating data of rolling bearings under different operating conditions, which represent different load conditions. It uses a sliding window method to segment the original vibration signal into data of equal length and normalizes it. It then constructs a labeled source domain sample set and an unlabeled target domain sample set. The source domain sample set and the target domain sample set each belong to different operating conditions. The unlabeled target domain sample set is then divided into a training set and a test set. A2 builds a network structure including a feature extractor and a dual classifier and initializes the network parameters. It uses the fusion classifier difference index to measure the difference between the dual classifier output results and to construct the network loss function. A3 uses a two-stage joint feature alignment strategy to reduce the difference between the source and target domain distributions. Each stage consists of three steps, each corresponding to a total loss function, and sequentially trains the overall network, classifier, and feature extractor. In the first stage, labeled source domain samples and unlabeled target domain training samples are fed into the network model, and the maximum mean difference is introduced to reduce the global feature distribution difference between the source and target domain samples. The three steps of this stage will be trained repeatedly until the preset maximum number of iterations is reached; in the second stage, the labeled source domain samples and the unlabeled target domain training samples are input into the network model. Based on the first stage, the potential category information of the target domain samples is fully utilized. The target domain sample output probability vector is obtained through the network model trained in the first stage as the target domain sample pseudo label. The local maximum mean difference is introduced to reduce the feature distribution difference between the same category samples in the source and target domains. The three steps are trained repeatedly until the maximum number of iterations of the second stage is reached, and the training is stopped to generate a maximum classifier difference network based on the two-stage joint feature alignment; A4 uses a trained two-stage joint feature alignment based maximum classifier difference network to identify the bearing fault type of the target domain test samples under variable working conditions and complete the final fault diagnosis task.

2. The rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier discrepancy network according to claim 1 is characterized by: The specific steps of A1 are as follows: A11) Obtain actual historical operating data of the rolling bearing under different operating conditions (different operating conditions represent different load conditions). Use the sliding window method to split the original vibration signal into data of equal length and perform normalization processing. The length of each data is 1024. The normalization processing formula is as follows: Among them, X is the original data, X norm is the normalized data, X min and X max are the minimum and maximum values ​​in the data set, respectively; A12) Construct a labeled source domain sample set n s represents the number of source domain samples, represents the i-th source domain sample, for Fault type labels; Construct unlabeled target domain sample set represents the jth target domain sample, n T represents the number of target domain samples; the source domain sample set and the target domain sample set belong to different working conditions; the target domain sample set is randomly divided into a training set and a test set.

3. The rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier discrepancy network according to claim 1 is characterized by: The specific steps of A2 are as follows: A21) Construct a network structure consisting of a feature extractor and a dual classifier and initialize the network parameters: the feature extractor consists of five convolutional blocks, each of which consists of a convolutional layer, a batch normalization layer, a ReLU activation function layer, and a max pooling layer. The convolution kernel size of the convolutional layer of the first convolutional block is 1×64, and the convolution kernel size of the convolutional layers of the remaining four convolutional blocks is 1×3. The classifier consists of two fully connected layers. The first fully connected layer uses ReLU as the activation function, and the second fully connected layer uses Softmax classifier as the activation function. The two classifiers have the same structure; A22) The fusion classifier difference index is used to measure the difference between the output results of the two classifiers. The fusion classifier difference index L f Obtained by formula (2): L f =μL u +ωL e (2) Among them, μ and ω are equilibrium parameters, μ is 1.6, ω is 1; L u is the joint difference index, obtained by formula (3); L e is the individual difference index, obtained by formula (4); Where G represents the feature extractor, C1 and C2 represent two classifiers, c Represents the cth element of the softmax output; 4. The rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier discrepancy network according to claim 1, characterized in that: In step A3, a two-stage joint feature alignment strategy is used to reduce the difference between the source and target domain distributions. Each stage consists of three steps, each corresponding to a total loss function. The overall network, classifier, and feature extractor are trained in sequence. The specific steps are as follows: A31) Input source domain samples, source domain sample labels and target domain training samples, and calculate the first step total loss function. The total loss function is calculated by calculating the cross entropy loss L S and the maximum mean difference index L mmd The sum is obtained and the overall network is trained; A32) Input source domain samples, source domain sample labels and target domain training samples, and calculate the second step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the fusion classifier difference index L f The sum is obtained and two classifiers are trained; A33) Input the target domain training sample and calculate the total loss function of the third step. The total loss function of this step is calculated by calculating the difference index of the fusion classifier - L f Obtain and train a feature extractor; A34) repeatedly performing steps A31) to A33) until the maximum number of iterations of the first stage is reached, where the number of iterations of the first stage is 15,000; A35) Input source domain samples, source domain sample labels and target domain training samples, and calculate the first step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the local maximum mean difference index L lmmd (x s ,x T ) and obtain the sum, and train the overall network; A36) Input source domain samples, source domain sample labels and target domain training samples, and calculate the second step total loss function. The total loss function is calculated by calculating the cross entropy loss L s (x s ,y s ) and the fusion classifier difference index L f The sum is obtained and two classifiers are trained; A37) Input the target domain training sample and calculate the total loss function of the third step. The total loss function of this step is calculated by calculating the difference index of the fusion classifier - L f Obtain and train a feature extractor; A38) Repeat steps A35) to A37 until the maximum number of iterations of the second stage is reached, which is 5000. Stop training and generate a maximum classifier difference network based on the two-stage joint feature alignment.

5. The rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier discrepancy network according to claim 4, characterized in that: In step A31, the total loss function is calculated by L S Cross entropy loss and L mmd The sum of the maximum mean difference index is obtained, as shown in formula (5) and formula (6): Among them, p(y|x s ) represents the softmax output probability of the source domain sample after the classifier, k∈{1,...,K}, K represents the number of fault categories and K=10, I [k =y s ] represents the indicator vector of the source domain sample label, and E represents the mathematical expectation; Where H represents the reproducing kernel Hilbert space, and Φ represents the function mapped to the reproducing kernel Hilbert space.

6. The rolling bearing unsupervised domain adaptation fault diagnosis method based on a two-stage joint feature alignment maximum classifier discrepancy network according to claim 4, characterized in that: In step A35, the local maximum mean difference index L in the overall loss function lmmd Obtained by formula (7): in, and represent source domain samples and target domain samples respectively, and Represents the weight of the k-th fault category in the source domain and the target domain respectively; The calculation formula is as follows: in, represents the probability that the sample belongs to fault category k, the source domain sample The kth item of the corresponding one-hot vector label is taken as The target domain samples are collected from unlabeled working conditions, so the kth item of the probability vector output by the network model is used as the target domain sample. of

Images