Wind turbine fault diagnosis method and equipment based on CC-FOA transfer learning

Abstract

This application discloses a wind turbine fault diagnosis method and device based on CC-FOA transfer learning, which relates to the field of wind turbine fault diagnosis. The method includes: collecting vibration signal data from multiple wind turbines, dividing the vibration signal data into source domain data and target domain data, and preprocessing them; using the preprocessed source domain data to train a stacked sparse denoising autoencoder network; using a chaotic cross-domain fruit fly optimization algorithm to optimize the number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network; inputting the preprocessed source domain data and target domain data into the stacked sparse denoising autoencoder network, adding a joint probability maximum mean difference constraint to each hidden layer, and retraining the network based on the sum of errors; applying the retrained stacked sparse denoising autoencoder network to the target domain data, extracting features, and performing fault classification. This application can effectively improve the accuracy and reliability of fault identification.

Description

Technical Field

[0001] The present application relates to the field of wind turbine fault diagnosis, and in particular to a wind turbine fault diagnosis method and device based on CC-FOA transfer learning. Background Art

[0002] Identifying multi-component faults in wind turbines has become a major challenge in the wind power industry, closely linked to the safe operation, power generation efficiency, and economic benefits of wind farms. As complex electromechanical systems, wind turbines consist of multiple key components, such as blades, gearboxes, and generators. These components are prone to various faults during long-term operation. Timely and accurate identification of these faults not only prevents major accidents but also optimizes maintenance strategies, reduces downtime, and improves power generation efficiency. Therefore, identifying multi-component faults in wind turbines is crucial for ensuring stable wind farm operation, reducing maintenance costs, and extending equipment life. Currently, the main technologies for identifying multi-component faults in wind turbines include signal processing-based methods, model-based methods, and data-driven methods. Signal processing methods primarily detect anomalies by analyzing sensor data such as vibration, sound, and temperature. Model-based methods identify faults by building physical or mathematical models of the wind turbine and comparing actual operating data with model predictions. Data-driven methods leverage machine learning algorithms to learn fault characteristics from large amounts of historical data. The working principle of these methods is based on continuous monitoring and analysis of the operating status of wind turbines, and the occurrence and type of faults are determined by identifying abnormal patterns or deviations from normal operating conditions.

[0003] However, these existing technologies face several challenges. First, due to the complex and ever-changing operating environment of wind turbines, single signal processing or modeling approaches often struggle to cope with a variety of complex fault scenarios. Second, traditional data-driven approaches require large amounts of labeled data for training. However, obtaining sufficient labeled fault data in practice is often difficult and costly. Furthermore, differences in data distribution between wind farms or wind turbine types pose challenges to the generalization of fault identification models. To address these issues, researchers have proposed various improved approaches. Among them, deep learning-based methods have attracted widespread attention due to their powerful feature extraction capabilities. For example, convolutional neural networks (CNNs) or recurrent neural networks (RNNs) are used to automatically learn the time-frequency characteristics of vibration signals, thereby improving fault identification accuracy. Another approach is ensemble learning, combining the strengths of multiple algorithms to improve model robustness. Furthermore, researchers have attempted to reduce the reliance on labeled data by using unsupervised or semi-supervised learning methods. However, these approaches still have limitations. First, they typically require large amounts of labeled data, which is difficult to obtain in practice. Secondly, these methods often struggle to effectively handle distribution differences between different data domains, resulting in poor model performance on new wind farms or new types of wind turbines. Finally, traditional methods may not fully utilize existing knowledge and require training the model from scratch each time, which is not only time-consuming and labor-intensive, but may also lead to unstable model performance.

[0004] Therefore, there is an urgent need for a technical solution that can effectively utilize the knowledge of source domain data to assist the learning of target domain data and improve the accuracy and reliability of fault identification. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention provides a wind turbine fault diagnosis method and device using transfer learning based on the Chaotic Spanning Crossover Fruit Fly Optimization Algorithm (CC-FOA). This application addresses technical issues such as the low generalization ability of existing models across different fault types.

[0006] An embodiment of the present application provides a wind turbine group transfer learning fault diagnosis method based on CC-FOA, including: collecting vibration signal data of multiple wind turbines, dividing the vibration signal data into source domain data and target domain data; preprocessing the source domain data and target domain data, including signal denoising and dimensionality transformation; using the preprocessed source domain data to train a stacked sparse denoising autoencoder network; using a chaotic cross-linked fruit fly optimization algorithm (CC-FOA) to optimize the number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network; inputting the preprocessed source domain data and target domain data into the optimized stacked sparse denoising autoencoder network, adding a joint probability maximum mean difference constraint to each hidden layer, and retraining the network based on the sum of errors; applying the retrained stacked sparse denoising autoencoder network to the target domain data, extracting features and performing fault classification.

[0007] In one possible implementation, a chaotic crossover fruit fly optimization algorithm is used to optimize the number of hidden layers and the number of neurons in each layer of a stacked sparse denoising autoencoder network, including: setting the fruit fly population size, the maximum number of iterations, the error accuracy, the acceleration factor, and the vertical crossover probability; converting the initial population of fruit flies into the number of hidden layers and the number of neurons in each layer of the autoencoder for training, calculating the reconstruction error as the fitness value of the fruit flies; sorting the fitness values and individual positions, and recording the global optimal value and the corresponding optimal individual ; All fruit flies in the population are randomly combined in pairs without duplication, and cross-pollination is performed to calculate the fitness value of the fruit flies, and the fruit flies with the best fitness in the current population are retained; each dimension of the fruit flies obtained by cross-pollination is normalized, and then all dimensions are randomly paired in pairs without duplication, and cross-pollination is performed vertically, and after denormalization, the fitness values are compared, and the global optimal value is updated. and the corresponding optimal individual The algorithm iterates continuously until the maximum number of iterations is reached or the error accuracy reaches a preset threshold. The fruit fly with the best fitness is selected as the optimal solution and output. The optimal solution is the optimal number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder.

[0008] In a possible implementation, the preprocessed source domain data and target domain data are input into the optimized stacked sparse denoising autoencoder network, a joint probability maximum mean difference constraint is added to each hidden layer, and the network is retrained based on the sum of errors, including: inputting the preprocessed source domain data and target domain data into the stacked sparse denoising autoencoder network; using the maximum mean difference to measure the distribution difference between the source domain and the target domain, and constructing a joint probability maximum mean difference constraint including marginal distribution and conditional distribution; embedding the joint probability maximum mean difference into the hidden layer of each sparse denoising autoencoder to migrate the model layer by layer; and training the stacked sparse denoising autoencoder network by minimizing the loss function including reconstruction error and joint probability maximum mean difference to obtain the optimal parameter set.

[0009] In a possible implementation, the maximum mean difference is used to measure the distribution difference between the source domain and the target domain, and a joint probability maximum mean difference constraint including marginal distribution and conditional distribution is constructed, including: establishing a joint probability difference model between the source domain and the target domain, including a same-class transferability measure and a different-class distinguishability measure to define the maximum mean difference of the discriminant joint probability; constructing a label matrix of the source domain data and a pseudo-label matrix of the target domain data, and calculating the first and second normalized matrices; generating an extended label matrix of the source domain and the target domain to calculate the third and fourth normalized matrices; constructing an optimization objective function for the maximum mean difference of the discriminant joint probability based on the first, second, third and fourth normalized matrices to solve the optimal linear mapping.

[0010] In a possible implementation, the maximum mean difference of the joint probability is embedded in the hidden layer of each sparse denoising autoencoder to perform layer-by-layer migration of the model, including: embedding the maximum mean difference of the joint probability into the hidden layer of each sparse denoising autoencoder, and extracting features of the source domain data and the target domain data by minimizing the loss function of the first sparse denoising autoencoder to align the feature space; using the output of the first sparse denoising autoencoder as the input of the second sparse denoising autoencoder, and reducing the distribution difference of the source domain data and the target domain data in the new feature space by minimizing the loss function of the second sparse denoising autoencoder; using the output of the second sparse denoising autoencoder as the input of the third sparse denoising autoencoder, and achieving full alignment of the source domain data and the target domain data in the common feature space by minimizing the loss function of the third sparse denoising autoencoder to migrate the model.

[0011] In a possible implementation method, a joint probability difference model of the source domain and the target domain is established, including a homogeneous transferability metric and a heterogeneous distinguishability metric, to define the maximum mean difference of the discriminant joint probability, including: establishing a joint probability difference model of the source domain and the target domain including a homogeneous transferability metric and a heterogeneous distinguishability metric, wherein the homogeneous transferability metric uses the maximum mean difference (MMD) to calculate the transferability from the source domain to the target domain; based on an empirical formula, obtaining the product formula of the conditional expectation and prior probability of the source domain and the target domain, substituting it into the maximum mean difference (MMD) formula to obtain the final source domain to target domain transferability metric formula; establishing a heterogeneous distinguishability metric model, using the maximum mean difference (MMD) to calculate the distinguishability between different fault categories in the source domain and the target domain, to obtain the final source domain and target domain heterogeneous distinguishability metric formula; substituting the obtained homogeneous transferability metric formula and heterogeneous distinguishability metric formula into the established joint probability difference model to obtain a complete definition of the maximum mean difference of the discriminant joint probability.

[0012] In one possible implementation, the portability metric includes:

[0013] ,

[0014] in, represents the transferability metric; C represents the total number of categories, that is, the number of fault types; c represents the current category index, ranging from 1 to C; Represents the total number of samples of source domain data; represents the number of samples belonging to category c in the source domain; Represents the total number of samples of target domain data; represents the number of samples belonging to category c in the target domain; Represents the transpose of the linear mapping matrix A, which is used to map data from the original feature space to the new feature space; represents the ii-th sample in the source domain belonging to category c; represents the jjth sample in the target domain belonging to category c.

[0015] In one possible implementation, the discriminative metric includes:

[0016] ,

[0017] in, Represents a discriminative metric, which is used to measure the distribution difference between the source domain and the target domain between different categories; Represents another category index, ranging from 1 to C, and ; Indicates that the target domain belongs to the category The number of samples; Indicates that the jj-th item in the target domain belongs to the category Sample.

[0018] An embodiment of the present application also provides a wind turbine group transfer learning fault diagnosis device based on CC-FOA, including: a processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method described in the above embodiment.

[0019] In the method and device for wind turbine transfer learning fault diagnosis based on CC-FOA provided above, the embodiment of the present application collects vibration signal data of the wind turbine, trains the stacked sparse denoising autoencoder network, and optimizes the number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network in combination with the chaotic vertical and horizontal cross fruit fly optimization algorithm, and uses the optimized stacked sparse denoising autoencoder network and the joint probability maximum mean difference constraint to achieve domain adaptation, which can effectively utilize the knowledge of the source domain data to assist the learning of the target domain data, thereby improving the accuracy and reliability of fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 A flow chart of a wind turbine fault diagnosis method based on CC-FOA transfer learning provided in an embodiment of the present application;

[0022] Figure 2 A flowchart of a maximum mean difference constraint method provided in an embodiment of the present application;

[0023] Figure 3 A schematic diagram of a process for a layer-by-layer migration method provided in an embodiment of the present application;

[0024] Figure 4 A schematic diagram of a chaotic cross-cutting fruit fly optimization algorithm flow provided in an embodiment of the present application;

[0025] Figure 5 A schematic diagram of the overall process of a migration identification method provided in an embodiment of the present application;

[0026] Figure 6A schematic diagram of a migration identification model provided in an embodiment of the present application;

[0027] Figure 7 A schematic diagram of migration recognition performance provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present disclosure.

[0029] Those skilled in the art will understand that the terms "first", "second" and the like in the embodiments of the present disclosure are only used to distinguish different steps, devices or modules, etc., and do not represent any specific technical meaning, nor do they represent the necessary logical order between them. It should also be understood that in the embodiments of the present disclosure, "multiple" may refer to two or more, and "at least one" may refer to one, two or more. It should also be understood that any component, data or structure mentioned in the embodiments of the present disclosure can generally be understood as one or more, unless explicitly defined or given a contrary revelation in the context. In addition, the term "and / or" in the present disclosure is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present disclosure generally indicates that the associated objects before and after are in an "or" relationship. It should also be understood that the description of each embodiment in the present disclosure emphasizes the differences between the embodiments, and the same or similar aspects thereof can be referenced to each other. For the sake of brevity, they will not be described one by one.

[0030] At the same time, it should be understood that for ease of description, the dimensions of the various parts shown in the drawings are not drawn according to the actual proportional relationship. The following description of at least one exemplary embodiment is actually only illustrative and is in no way intended to limit the present disclosure and its application or use. Technologies, methods and equipment known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods and equipment should be considered part of the specification. It should be noted that similar numbers and letters represent similar items in the following figures, so once an item is defined in one figure, it does not need to be further discussed in subsequent figures.

[0031] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0032] Figure 1 This is a flow chart of a CC-FOA-based wind turbine fault diagnosis method 100 for transfer learning, provided in an embodiment of the present application. This method aims to address various fault issues that may arise during the actual operation of wind turbines. By collecting vibration signal data from the wind turbine, a stacked sparse denoising autoencoder network is trained. The number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network are optimized using the chaotic cross-correlation fruit fly optimization algorithm (CC-FOA). Transfer learning techniques are then used to process and analyze the data, enabling fault diagnosis of multiple components of the wind turbine.

[0033] like Figure 1 As shown, in step S101, vibration signal data from multiple wind turbines is collected and divided into source domain data and target domain data. Specifically, the method first collects vibration signal data from multiple wind turbines and divides the collected vibration signal data into source domain data and target domain data. Source domain data typically refers to historical data of known fault types, while target domain data refers to real-time data of faults to be identified or data of unknown fault types. This division of source domain data into target domain data provides a foundation for subsequent transfer learning.

[0034] At step S102, the source and target domain data are preprocessed, including signal denoising and dimensionality transformation. After obtaining the source and target domain data, the method preprocesses them, including signal denoising and dimensionality transformation. Signal denoising removes noise interference from vibration signals to improve signal quality and reliability. This can be achieved through various signal processing techniques, such as wavelet transform and Fourier transform. Dimensionality transformation converts the original high-dimensional vibration signal data into a more compact and efficient low-dimensional representation to reduce computational complexity and improve feature extraction efficiency. Common dimensionality transformation methods include principal component analysis (PCA) and independent component analysis (ICA).

[0035] In step S103, the preprocessed source domain data is used to train a stacked sparse denoising autoencoder network. A stacked sparse denoising autoencoder is a deep learning model composed of multiple stacked sparse denoising autoencoders. Sparse denoising autoencoders are unsupervised learning algorithms that can learn high-level feature representations of data while also exhibiting denoising and sparsification properties. By introducing sparsity constraints in each hidden layer of the sparse denoising autoencoder, the model can learn more compact and discriminative feature representations.

[0036] In step S104, the number of hidden layers and the number of neurons per layer of the stacked sparse denoising autoencoder network are optimized using the chaotic cross-linked fruit fly optimization algorithm (CC-FOA). To optimize the model's hyperparameters, the chaotic cross-linked fruit fly optimization algorithm is combined with the CC-FOA to optimize the number of hidden layers and the number of neurons per layer of the stacked sparse denoising autoencoder network. The CC-FOA is an improved fruit fly optimization algorithm that combines chaos theory with a cross-linked strategy to enhance its global search capability and convergence accuracy. The CC-FOA algorithm's execution process includes initializing the population with chaos, calculating the fitness value of each individual, updating the fruit fly's position based on a chaotic map, and optimizing individual positions using a cross-linked strategy. The algorithm iterates until a stopping condition is met, ultimately outputting a global optimal solution.

[0037] Figure 4 A flow chart of a chaotic cross-cutting fruit fly optimization algorithm provided in the embodiment of the present application. This method aims to solve the hyperparameter optimization problem of stacked sparse denoising autoencoders. Figure 4 As shown, we first initialize the parameters, set the fruit fly population size to 30, the maximum number of iterations to 200, and the vertical and horizontal crossover probability to 0.6. Then set the number of iterations to Set to 1. Convert the initial population of fruit flies into the number of hidden layers and neurons in each layer of the stacked sparse denoising autoencoder for training. Calculate the reconstruction error using the following formula, which is the fitness value of the fruit fly.

[0038] ;

[0039] Mean Squared Error (MSE) is used to measure the difference between the model's predicted value and the true value. In the present invention, MSE is used to calculate the reconstruction error of the stacked sparse denoising autoencoder network, which is used as the fitness value of individual fruit flies in the fruit fly optimization algorithm to evaluate the performance of the stacked sparse denoising autoencoder network. is the true value (target value), which represents the true label or true output of the t-th sample. In the present invention, Represents the true value (or original input data) of the wind turbine vibration signal; is the predicted value (model output), which represents the model predicted value of the t-th sample. In the invention, represents the reconstructed value of the input data by the stacked sparse denoising autoencoder network (i.e., the model's predicted output for the input data); K is the total number of samples, indicating the total number of samples in the dataset; m is an index variable, indicating the error calculation for the mth sample, from 1 to N. By minimizing the MSE, the number of hidden layers and the number of neurons in each layer of the autoencoder can be optimized, thereby improving the model's ability to reconstruct wind turbine vibration signals and the accuracy of fault diagnosis. The calculation of MSE is one of the core indicators of model training and optimization, directly affecting the model's performance and generalization ability. Sort the fitness values and individual positions, and record the global optimal value. and the corresponding optimal individual and individual optimal values ; When the maximum number of iterations is not reached, the fruit fly population enters the optimization process according to the vertical and horizontal crossover algorithm generated by the following formula;

[0040] ;

[0041] Among them: Chaos fruit flies In the fruit fly population, for is the population size of the crossover algorithm The dimension of the variables included for Drosophila; That is ; Randomly generate chaotic initial fruit flies ; u is the control parameter of the chaotic map, with a value of 2; s is the perturbation intensity (or mixing coefficient), which is used to control the influence of the chaotic map and the global optimal value on the position update of the fruit fly. s is a parameter between 0 and 1, which is used to balance the chaotic map ( ) and the global optimal value ( ) contributes to the update of the fruit fly's position. Preferably, the value of s is 0.4; For the The updated value of the position of the fruit fly, in the chaotic cross-cutting fruit fly optimization algorithm, represents the position of the fruit fly after being updated by the chaotic mapping and cross-cutting strategy; For the The position of a chaotic fruit fly represents the position of the fruit fly generated by the chaotic mapping. It is the result of the chaotic mapping and is used to introduce chaotic characteristics to enhance the global search ability of the algorithm. is the position of the current global optimal solution, indicating the optimal fruit fly position found by the algorithm during the iteration process, and is used to guide the fruit fly population to search in the direction of the global optimal solution to improve the convergence of the algorithm; For the In the iteration The chaotic variable value of dimension 2 is the input value of the chaotic map, which is used to generate the next generation of chaotic variables. .

[0042] according to Different chaotic mapping formulas are used to update the positive and negative cases of , to ensure the dynamics and diversity of chaotic variables.

[0043] By using chaotic mapping and a cross-cutting strategy, the algorithm can more effectively explore the global optimal solution in the search space, avoiding being trapped in a local optimum. In the invention, the above formula is used to optimize the number of hidden layers and the number of neurons per layer in a stacked sparse denoising autoencoder network, thereby improving the performance of the wind turbine fault diagnosis model.

[0044] All fruit flies in the population are randomly combined in pairs without repetition, and crossover is performed according to the following formula to calculate the fitness value of the fruit flies, and the fruit flies with the smallest fitness value are retained;

[0045] ;

[0046] ;

[0047] ;

[0048] in: A random number or weight coefficient in [0,1] used to control and Contribution ratio in mixing operations; A random number in the range [-1,1] used to control the difference term or In the mixed operation, the difference term reflects the difference between the two individuals on dimension d; M is the size of the fruit fly, that is, the total number of individuals; is the variable dimension, that is, the total number of dimensions of the problem; 、 Parent Drosophila and No. dimension 、 They are and The first dimensional offspring; i, j represent the index of the individual; N represents a set of natural numbers, N(1,M) represents a set of natural numbers from 1 to the total number of individuals M; and N(1,D) represents a set of natural numbers from 1 to the number of variable dimensions D.

[0049] Normalize each dimension of the fruit flies obtained by horizontal crossover, then randomly pair all dimensions without duplication, perform vertical crossover according to the following formula, compare the fitness values after denormalization, and update the fruit flies;

[0050] ;

[0051] ;

[0052] in, Indicated in dimension Above, the result of some mixing or crossover operation of individual i; Indicates that individual i is in dimension The eigenvalue or function value on , Represents individual i in dimension The eigenvalues or function values on the two dimensions are the attributes or states of individual i in different dimensions. r is a random number or weight coefficient with a value range of [0,1]. It is used to control and The contribution ratio in the mixing operation. When r is close to 1, The contribution of is greater; when r is close to 0, The contribution is greater.

[0053] The algorithm continues to iterate until the maximum number of iterations is reached or the error accuracy reaches a preset threshold. The final global optimal solution is the optimal number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder.

[0054] At step S105, the pre-processed source domain data and target domain data are input into the optimized stacked sparse denoising autoencoder network, a joint probability maximum mean difference constraint is added to each hidden layer, and the network is retrained based on the sum of errors. Specifically, the steps include: inputting the pre-processed source domain data and target domain data into the stacked sparse denoising autoencoder network; using the maximum mean difference to measure the distribution difference between the source domain and the target domain, and constructing a joint probability maximum mean difference constraint (such as ) that includes marginal distribution and conditional distribution. Figure 2 As shown); embed the joint probability maximum mean difference into the hidden layer of each sparse denoising autoencoder to transfer the model layer by layer (as shown Figure 3 The stacked sparse denoising autoencoder network is trained to obtain the optimal parameter set by minimizing the loss function including the reconstruction error and the maximum mean difference of the joint probability.

[0055] The maximum mean difference in joint probability is a metric that measures the difference in distribution between the source and target domains, taking into account both marginal and conditional distributions. By embedding this constraint into the hidden layers of each sparse denoising autoencoder, we can achieve layer-by-layer model transfer, gradually aligning the feature distributions of the source and target domains and reducing inter-domain differences. This transfer learning approach effectively leverages knowledge from the source domain to aid learning in the target domain, improving the model's generalization and adaptability.

[0056] After adding the joint probability maximum mean difference constraint, this method constructs a loss function based on the reconstruction error and the joint probability maximum mean difference, and retrains the stacked sparse denoising autoencoder network. The reconstruction error measures the model's ability to reconstruct the input data, while the joint probability maximum mean difference measures the distribution difference between the source and target domains. By minimizing the loss function that includes these two terms, the model can effectively narrow the gap between the source and target domains while maintaining good reconstruction capabilities, achieving domain adaptation.

[0057] In step S106, the retrained stacked sparse denoising autoencoder network is applied to the target domain data to extract features and perform fault classification. Finally, the method applies the retrained stacked sparse denoising autoencoder network to the target domain data to extract features and perform fault classification. By inputting the target domain data into the trained network, a high-level feature representation of the target domain data is obtained. These feature representations contain key information about the target domain data and can effectively distinguish different fault types. Based on the extracted features, various classification algorithms, such as support vector machines (SVMs) and random forests, can be used to identify faults in multiple components of wind turbines.

[0058] In summary, the CC-FOA-based wind turbine transfer learning fault diagnosis method provided in the embodiment of the present application collects vibration signal data of the wind turbine, trains a stacked sparse denoising autoencoder network, and optimizes the number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network by combining the chaotic cross-linking fruit fly optimization algorithm. The optimized stacked sparse denoising autoencoder network and the joint probability maximum mean difference constraint are used to achieve domain adaptation. The method can effectively utilize the knowledge of the source domain data to assist the learning of the target domain data, thereby improving the accuracy and reliability of fault identification. This method has important practical application value for the operation and maintenance and fault diagnosis of wind turbines, and can reduce the downtime of wind turbines, improve power generation efficiency, and reduce maintenance costs.

[0059] Figure 2A flowchart of a maximum mean difference constraint method provided for an embodiment of the present application. This method aims to solve the feature migration and classification tasks in the domain adaptation problem. Domain adaptation refers to the process of migrating knowledge learned in one domain (source domain) to another related but different domain (target domain). In practical applications, there is usually a difference in data distribution between the source domain and the target domain. This difference will lead to performance degradation when the model trained in the source domain is directly applied to the target domain. Therefore, this method effectively reduces the distribution difference between the source domain and the target domain by establishing a discriminant joint probability maximum mean difference model, while maintaining the distinguishability between different categories.

[0060] like Figure 2 As shown, at step S201, a joint probability difference model of the source domain and the target domain is established, including a homogeneous transferability metric and a heterogeneous distinguishability metric, to define the maximum mean difference of the discriminant joint probability. Specifically, it includes: establishing a joint probability difference model of the source domain and the target domain including a homogeneous transferability metric and a heterogeneous distinguishability metric, wherein the homogeneous transferability metric uses the maximum mean difference (MMD) to calculate the transferability from the source domain to the target domain; based on the empirical formula, obtaining the product formula of the conditional expectation and prior probability of the source and target domains, substituting it into the maximum mean difference (MMD) formula to obtain the final transferability metric formula from the source domain to the target domain; establishing a heterogeneous distinguishability metric model, using the maximum mean difference (MMD) to calculate the distinguishability between different fault categories in the source domain and the target domain, to obtain the final heterogeneous distinguishability metric formula of the source domain and the target domain; substituting the obtained homogeneous transferability metric formula and heterogeneous distinguishability metric formula into the established joint probability difference model to obtain a complete definition of the maximum mean difference of the discriminant joint probability.

[0061] Among them, the transferability metrics include:

[0062] ,

[0063] in, represents the transferability metric, which is used to measure the distribution difference between the source domain and the target domain under the same category; C represents the total number of categories, that is, the number of fault types; c represents the current category index, which ranges from 1 to C; Represents the total number of samples of source domain data; represents the number of samples belonging to category c in the source domain; Represents the total number of samples of target domain data; represents the number of samples belonging to category c in the target domain; A represents the linear mapping matrix, which is used to map the original feature space to the new feature space in order to reduce the distribution difference between the source domain and the target domain. Represents the transpose of the linear mapping matrix A, which is used to map data from the original feature space to the new feature space; represents the ii-th sample in the source domain belonging to category c; represents the jj-th sample belonging to category c in the target domain; Represents the square of the L2 norm, which is used to calculate the square of the Euclidean distance between two vectors.

[0064] Among them, the discriminative metrics include:

[0065] ,

[0066] in, Represents a discriminative metric, which is used to measure the distribution difference between the source domain and the target domain between different categories; Represents another category index, ranging from 1 to C, and ; Indicates that the target domain belongs to the category The number of samples; Indicates that the jj-th item in the target domain belongs to the category Sample.

[0067] At step S202, a label matrix for the source domain data and a pseudo-label matrix for the target domain data are constructed, and the first and second normalized matrices are calculated. A label matrix is a matrix used to represent data category information, where each row corresponds to a sample and each column corresponds to a category. For source domain data, the label matrix can be directly constructed using the actual label information. However, for target domain data, since actual label information is generally unavailable, a pseudo-label matrix must be constructed. Pseudo-labels can be obtained by predicting the target domain data using an initial classifier (e.g., a classifier trained on the source domain).

[0068] The purpose of calculating the normalization matrix is to standardize the data so that scale differences between different features do not adversely affect the subsequent optimization process. The first normalization matrix is calculated based on the label matrix of the source domain data, while the second normalization matrix is calculated based on the pseudo-label matrix of the target domain data. These two normalization matrices will play an important role in the subsequent optimization process.

[0069] At step S203, an extended label matrix of the source domain and the target domain is generated to calculate the third and fourth normalized matrices. The extended label matrix is an extension of the original label matrix, which may contain more information, such as the similarity between samples or the relationship between categories. By generating the extended label matrix, more prior knowledge or structural information can be introduced, thereby improving the effect of domain adaptation. The third and fourth normalized matrices are calculated based on the extended label matrix. The introduction of these matrices can help better capture the intrinsic structure and category information of the data, thereby better maintaining the discriminability of the data during the optimization process.

[0070] At step S204, an optimization objective function for the maximum mean difference of the discriminant joint probabilities is constructed based on the first, second, third, and fourth normalized matrices to solve the optimal linear mapping. This optimization objective function is the core of the entire method, comprehensively considering both in-class transferability and heterogeneous distinguishability, while leveraging label information (including pseudo-labels) from both the source and target domains.

[0071] The process of constructing the optimization objective function may involve a weighted combination of multiple sub-objectives. For example, this may include minimizing the MMD distance between the source and target domains (to improve transferability), maximizing the MMD distance between different categories (to improve distinguishability), and minimizing the classification error in the source domain. These sub-objectives are balanced using appropriate weight coefficients to form the final optimization objective.

[0072] Finding the optimal linear mapping is the ultimate goal of this method. A linear mapping can be viewed as a feature transformation that maps data points from the original feature space to a new feature space. In this new feature space, the distribution difference between the source and target domains is minimized while the distinguishability between different classes is maximized. This optimal linear mapping can be found using optimization algorithms such as gradient descent.

[0073] It is worth noting that although this method uses linear mapping as a means of feature transformation, it does not mean that it is only applicable to linearly separable problems. Through the kernel trick, this method can be extended to nonlinear cases, thereby handling more complex domain adaptation problems.

[0074] In summary, the maximum mean difference constraint method provided by the embodiments of this application constructs an effective domain adaptation framework by cleverly combining a homogeneous transferability metric with a heterogeneous distinguishability metric. This method not only considers the distribution differences between the source and target domains but also fully utilizes category information, thereby achieving effective knowledge transfer while maintaining category discriminability.

[0075] Figure 3 A flow chart of a layer-by-layer migration method provided in an embodiment of the present application. This method aims to solve the problem of inconsistent data distribution between the source domain and the target domain in cross-domain learning. The layer-by-layer migration method refers to gradually reducing the distribution difference between the source domain data and the target domain data through multiple consecutive processing steps, and finally achieving effective migration of the model. The core idea of this method is to use a multi-layer sparse denoising autoencoder (SDAE) structure, combined with the joint maximum mean discrepancy (JMMD) technology, to align the feature distribution of the source domain and the target domain layer by layer.

[0076] Specifically, the sparse denoising autoencoder is an unsupervised learning algorithm that discovers useful features in the data by learning a compressed representation of the input data. "Sparse" means that in the hidden layer, the activation values of most neurons are close to zero, with only a few neurons being activated. This helps learn more robust and meaningful feature representations. "Denoising" refers to adding noise to the input data during training and then requiring the network to reconstruct the original noise-free data. This method improves the model's robustness to noise and learns more stable features.

[0077] Joint Maximum Mean Difference (JMMD) is a method for measuring the difference between two probability distributions. In cross-domain learning, JMMD can be used to measure the difference between the source and target domain data distributions and achieve domain adaptation by minimizing this difference. Embedding JMMD into the hidden layers of SDAE allows for feature extraction while considering the consistency of the source and target domain data distributions, thereby better aligning the feature space.

[0078] This method uses a three-layer sparse denoising autoencoder structure, where each layer performs a different task and works together to achieve the goal of gradual migration. This multi-layer structure allows the model to gradually refine and optimize feature representations, thereby more effectively handling complex cross-domain learning problems.

[0079] like Figure 3 As shown, at step S301, the maximum mean difference of the joint probability is embedded in the hidden layer of each sparse denoising autoencoder, and features of the source domain data and the target domain data are extracted by minimizing the loss function of the first sparse denoising autoencoder to align the feature space.

[0080] Specifically, the maximum mean difference of joint probabilities is embedded into the hidden layer of the first sparse denoising autoencoder. The purpose of this step is to consider the consistency of the data distribution of the source and target domains while performing feature extraction. By minimizing the loss function of the first sparse denoising autoencoder, preliminary feature extraction and alignment of the source and target domain data can be achieved. The loss function here may include reconstruction error, sparsity constraint, and JMMD term. The reconstruction error ensures that the autoencoder can effectively reconstruct the input data, the sparsity constraint forces the hidden layer to learn sparse feature representation, and the JMMD term is used to reduce the difference in feature distribution between the source and target domains.

[0081] In step S302, the output of the first sparse denoising autoencoder is used as the input of the second sparse denoising autoencoder, and the distribution difference between the source domain data and the target domain data in the new feature space is reduced by minimizing the loss function of the second sparse denoising autoencoder.

[0082] This method further reduces the distribution difference between source and target domain data in the new feature space. Specifically, the output of the first sparse denoising autoencoder is used as the input to the second sparse denoising autoencoder. This cascaded structure allows the model to further extract and optimize features based on the features learned in the first layer. By minimizing the loss function of the second sparse denoising autoencoder, the distribution difference between source and target domain data can be further reduced in the new feature space. The loss function in this step may be similar to that in the first step, but the weights of each item may be adjusted to place greater emphasis on reducing inter-domain differences.

[0083] In step S303, the output of the second sparse denoising autoencoder is used as the input of the third sparse denoising autoencoder, and the source domain data and the target domain data are fully aligned in the common feature space by minimizing the loss function of the third sparse denoising autoencoder to perform model migration.

[0084] Specifically, the output of the second sparse denoising autoencoder is used as the input to the third sparse denoising autoencoder. By minimizing the loss function of the third sparse denoising autoencoder, sufficient alignment of the source and target domain data can be achieved in the final feature space. The loss function of this step may further increase the weight of domain alignment to ensure that the data distribution of the source and target domains in the final feature space is as consistent as possible.

[0085] It's worth noting that the loss function for each layer of a sparse denoising autoencoder may contain multiple components, such as reconstruction error, sparsity constraint, regularization term, and domain alignment term (such as JMMD). The specific forms and weights of these components may be adjusted based on the specific application scenario and data characteristics. For example, the reconstruction error can take the form of mean squared error or cross entropy, the sparsity constraint can be implemented using KL divergence, the regularization term can use the L1 or L2 norm, and the domain alignment term can use JMMD or other domain adaptation techniques.

[0086] In general, the layer-by-layer migration method provided in the embodiment of the present application achieves a gradual alignment of the data distribution of the source domain and the target domain by cleverly combining the sparse denoising autoencoder and the joint probability maximum mean difference. This method can not only effectively extract the feature representation of the data, but also consider the differences between domains in the process of feature extraction, thereby achieving more effective knowledge transfer. The layer-by-layer design of this method allows the model to gradually refine and optimize the feature representation, which is of great significance for dealing with complex cross-domain learning problems. Through this method, when the target domain data is limited, the knowledge of the source domain can be fully utilized to improve the performance of the model in the target domain.

[0087] Figure 5A schematic diagram of the overall process of a migration identification method provided in an embodiment of the present application. First, the source domain data and target domain data are destroyed with the same noise level. Then, the destroyed source domain data is used to train the AE with sparsity constraints added. The trained AE can better handle data with this noise level and has higher robustness and analysis capabilities. Next, the destroyed source domain data and target domain data are input into the trained AE to obtain the hidden features of the source domain and target domain respectively. Then, DJPMMD is used to measure the difference between the hidden features of the source domain and the target domain, and the AE is trained again until DJPMMD converges. Finally, the noise level is reduced, the original source domain data and target domain data are destroyed again, and the AE is retrained until the noise level reaches the set value. The hidden layer input of the trained AE is used as the input of the next AE, and the above process is repeated until all AEs are trained, and the trained multiple AEs are stacked to obtain the final TL-SSMDAE.

[0088] Multi-level noise training can capture more detailed information from different scales hidden in the original data, which is beneficial for fault classification. In addition, data corruption using multiple noise levels actually generates artificial data to expand the data size, which can alleviate the problem of small datasets to some extent.

[0089] Figure 6 This is a schematic diagram of a migration identification model provided in an embodiment of the present application. Figure 6 The model shown collects vibration signals from different components of each wind turbine from the SCADA system. The vibration signals are initially processed and appropriately transformed. The source domain vibration signal data is fed into the SSMDAE for training. Both the source domain and target domain vibration signal data are simultaneously fed into the trained SSMDAE. DJPMMD is added to the hidden layer of each SSMDAE. The sum of the error between DJPMMD and SSMDAE is used as the new error for training the SSMDAE. The retrained SSMDAE is applied to the target domain, and the final output is classified.

[0090] It should be understood that the systems depicted in the diagrams are illustrative and non-restrictive. This means that the system architectures depicted are not limited to a specific form or design, but are presented as examples. In other words, the architectures shown in the diagrams can be viewed as a means of expressing clearly the relevant concepts and relationships, and do not exclude other forms of architecture. Therefore, when interpreting the architectures depicted in the diagrams, it should be understood that the model is flexible and diverse, and its purpose is to provide an illustrative description, not a restrictive stipulation of a specific form.

[0091] Figure 7A schematic diagram illustrating the transfer recognition performance provided by an embodiment of the present application. The experimental testing of the model's generalization capabilities is divided into two main parts: the first is the generalization capability between datasets measured at different locations, which primarily analyzes the model's generalization capability across different components. The second is the generalization capability between datasets of different faults measured at the same location, which primarily analyzes the model's generalization capability across different faults.

[0092] First, we analyzed the model's generalization ability across datasets collected at different locations. The specific research content is shown in Table 1. The DE, FE, and BA datasets were each divided into a 70% training set and a 30% test set. The model was trained using the training set from one domain and its classification performance was tested using the test set from another domain, thereby measuring the model's generalization ability.

[0093] Table 1 Migration experiment information

[0094]

[0095] To avoid contingency, all experiments were repeated 10 times, the maximum and minimum values were removed and the average was taken. The classification results of SSMDAE without DJMMMD are shown in the figure. Figure 7 As shown in Figure 2, since the BA side does not collect normal data sets, experiments 7, 8, and 9 only use 9 types of fault data.

[0096] from Figure 7 As can be seen from (A) in the figure, when the source domain and the target domain are from the same dataset, that is, experiments 1, 5, and 9, the classification accuracy of these experiments is above 94%. However, when the source domain and the target domain are from different datasets, the classification accuracy does not exceed 60%. This shows that the model has poor generalization ability between datasets obtained at different locations.

[0097] Figure 7 (B) in the figure is the classification accuracy of the proposed method in experiments 1-9. Figure 7 As can be seen in (B), the classification accuracy of the model across different domains has increased by more than 35%. Among them, after the model trained on BA was migrated to DE, it achieved a classification accuracy of 94.88%, which is comparable to BA itself. This shows that the proposed method architecture can significantly improve the generalization ability of the model between datasets collected at different locations and enhance the robustness of the model.

[0098] Furthermore, an embodiment of the present application also provides a wind turbine group transfer learning fault diagnosis device based on CC-FOA, including: a processor, a memory, and a system bus; the processor and the memory are connected through the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes any of the above methods.

[0099] Furthermore, an embodiment of the present application also provides a computer program product, which, when running on a terminal device, enables the terminal device to execute any of the above methods.

[0100] Through the description of the above embodiments, it can be seen that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in each embodiment of the present application or certain parts of the embodiments.

[0101] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the methods.

[0102] It should also be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0103] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily 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 application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A wind turbine fault diagnosis method based on CC-FOA transfer learning, characterized in that: include: Collecting vibration signal data of a plurality of wind turbine generator sets, and dividing the vibration signal data into source domain data and target domain data; Preprocessing the source domain data and the target domain data, including signal noise reduction and dimensionality transformation; Use the preprocessed source domain data to train a stacked sparse denoising autoencoder network; The chaotic cross-pollination fruit fly optimization algorithm is used to optimize the number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder network, including: Set the fruit fly population size, maximum number of iterations, error accuracy, acceleration factor, and vertical crossover probability; convert the initial population of fruit flies into the number of hidden layers and neurons in each layer of the autoencoder for training, and calculate the reconstruction error as the fitness value of the fruit flies; Sort the fitness values and individual positions and record the global optimal value and the corresponding optimal individual ; All fruit flies in the population are randomly combined in pairs without duplication, and crossover is performed to calculate the fitness value of the fruit flies, and the fruit flies with the best fitness in the current population are retained; Normalize each dimension of the fruit flies obtained by horizontal crossover, then randomly pair all dimensions without duplication, perform vertical crossover operation, compare fitness values after denormalization, and update the global optimal value and the corresponding optimal individual ; The algorithm iterates continuously until the maximum number of iterations is reached or the error accuracy reaches a preset threshold. The fruit fly with the best fitness is selected as the optimal solution and output. The optimal solution is the optimal number of hidden layers and the number of neurons in each layer of the stacked sparse denoising autoencoder. Input the preprocessed source domain data and target domain data into an optimized stacked sparse denoising autoencoder network, add a joint probability maximum mean difference constraint to each hidden layer, and retrain the network based on the sum of errors; The retrained stacked sparse denoising autoencoder network is applied to the target domain data to extract features and perform fault classification.

2. The method according to claim 1, characterized in that in, Inputting the preprocessed source domain data and target domain data into the optimized stacked sparse denoising autoencoder network, adding a joint probability maximum mean difference constraint to each hidden layer, and retraining the network based on the sum of errors, including: Input the preprocessed source domain data and target domain data into the stacked sparse denoising autoencoder network; The maximum mean difference is used to measure the distribution difference between the source domain and the target domain, and the maximum mean difference constraint of the joint probability including marginal distribution and conditional distribution is constructed; Embedding the maximum mean difference of the joint probability into the hidden layer of each sparse denoising autoencoder to perform layer-by-layer migration of the model; The stacked sparse denoising autoencoder network is trained to obtain the optimal parameter set by minimizing a loss function that includes reconstruction error and maximum mean difference of joint probabilities.

3. The method according to claim 2, characterized in that in, The maximum mean difference is used to measure the distribution difference between the source domain and the target domain, and a joint probability maximum mean difference constraint containing marginal distribution and conditional distribution is constructed, including: Establish a joint probability difference model between the source domain and the target domain, including the same-class transferability measure and the different-class distinguishability measure to define the maximum mean difference of the discriminant joint probability; Construct a label matrix for source domain data and a pseudo label matrix for target domain data, and calculate the first and second normalization matrices; Generate the expanded label matrices of the source and target domains to calculate the third and fourth normalization matrices; An optimization objective function for determining the maximum mean difference of the joint probability of discrimination is constructed based on the first, second, third and fourth normalized matrices to solve the optimal linear mapping.

4. The method according to claim 2, characterized in that in, The maximum mean difference of the joint probability is embedded into the hidden layer of each sparse denoising autoencoder to perform layer-by-layer migration of the model, including: The maximum mean difference of joint probabilities is embedded into the hidden layer of each sparse denoising autoencoder. By minimizing the loss function of the first sparse denoising autoencoder, features of the source domain data and the target domain data are extracted to align the feature space. The output of the first sparse denoising autoencoder is used as the input of the second sparse denoising autoencoder. By minimizing the loss function of the second sparse denoising autoencoder, the distribution difference between the source domain data and the target domain data in the new feature space is reduced. The output of the second sparse denoising autoencoder is used as the input of the third sparse denoising autoencoder. By minimizing the loss function of the third sparse denoising autoencoder, the source domain data and the target domain data are fully aligned in the common feature space to perform model migration.

5. The method according to claim 3, characterized in that in, Establish a joint probability difference model between the source domain and the target domain, including the same-class transferability measure and the different-class distinguishability measure to define the maximum mean difference of the discriminant joint probability, including: A joint probability difference model of the source and target domains is established, which includes the same-class transferability metric and the heterogeneous distinguishability metric. The same-class transferability metric uses the maximum mean difference (MMD) to calculate the transferability from the source domain to the target domain. Based on the empirical formula, we obtain the product formula of the conditional expectation and prior probability of the source and target domains, and substitute it into the maximum mean difference (MMD) formula to obtain the final transferability measurement formula from the source domain to the target domain. A heterogeneous distinguishability measurement model is established, and the maximum mean difference (MMD) is used to calculate the distinguishability between different fault categories in the source and target domains, and the final heterogeneous distinguishability measurement formula for the source and target domains is obtained; The obtained transferability measurement formula and heterogeneous discriminability measurement formula are substituted into the established joint probability difference model to obtain the complete definition of the maximum mean difference of the discriminant joint probability.

6. The method according to claim 5, characterized in that in, Transferability metrics, including: ， in, represents the transferability metric; C represents the total number of categories, that is, the number of fault types; c represents the current category index, ranging from 1 to C; Represents the total number of samples of source domain data; represents the number of samples belonging to category c in the source domain; Represents the total number of samples of target domain data; represents the number of samples belonging to category c in the target domain; Represents the transpose of the linear mapping matrix A, which is used to map data from the original feature space to the new feature space; represents the ii-th sample belonging to category c in the source domain; represents the jjth sample in the target domain belonging to category c.

7. The method according to claim 6, characterized in that in, Discriminative measures, including: ， in, represents a discriminative measure; Represents another category index, ranging from 1 to C, and ; Indicates that the target domain belongs to the category The number of samples; Indicates that the jjth item in the target domain belongs to the category Sample.

8. A wind turbine fault diagnosis device based on CC-FOA transfer learning, characterized in that: include: A processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; The memory is configured to store one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by the processor, the processor is enabled to perform the method according to any one of claims 1 to 7.

Images