Improved DBSCAN dry-type transformer vibration signal fault early warning method and system

Abstract

The invention provides an improved DBSCAN dry-type transformer vibration signal fault early warning method and system, and belongs to the technical field of dry-type transformer fault diagnosis, and the method comprises the steps: obtaining a traction rectifier transformer vibration signal of a to-be-detected urban rail power supply system; processing the acquired vibration signal by using a pre-trained fault diagnosis model to obtain a fault diagnosis result; and if the number of the abnormal points in the fault diagnosis result exceeds a set threshold value, performing early warning. According to the method, the DBSCAN parameters are automatically optimized through multi-objective optimization and a simulated annealing algorithm, so that the judgment of abnormal points reaches a most accurate state, and the workload is reduced.

Description

Technical Field

[0001] This invention relates to the field of dry-type transformer fault diagnosis technology, specifically to an improved DBSCAN (density-based clustering algorithm) method and system for early warning of vibration signals in dry-type transformers. Background Technology

[0002] In urban rail power supply systems, dry-type transformers are key components, and their stability and safety are crucial for ensuring the normal operation of urban rail trains and the safety of passengers. However, due to the complexity and unique characteristics of urban rail power supply systems, traction rectifier transformers are susceptible to various factors during operation, leading to potential failure risks. Traditional fault diagnosis methods often rely on experience-based judgment or threshold-based monitoring, which are often insufficient to accurately assess the transformer's health status in complex environments.

[0003] To promptly detect and prevent faults in dry-type transformers in urban rail power supply systems, vibration signal-based fault early warning methods have become an important technical means. Neural network technology, which has emerged in recent years, possesses strong signal processing capabilities and has been widely applied by scholars both domestically and internationally in the vibration signal fault diagnosis of electrical equipment. However, neural networks require a large amount of data of different types for model training. Since the actual collected vibration signals are mostly normal signals and the fault signal samples are small, this affects the accuracy of model training, thus impacting the final precision.

[0004] To address the imbalance in data acquisition during real-world fault diagnosis, anomaly detection algorithms can be employed to diagnose faults using the actual acquired data. First, vibration signals from the traction rectifier transformers of an actual urban rail power supply system are acquired and preprocessed, including filtering, key feature extraction, and data standardization, to ensure the effectiveness of DBSCAN's distance metric. Second, using the global profile coefficient and DBI as objective functions, simulated annealing is selected as the optimization algorithm, and a parameter optimization model is built to optimize the DBSCAN parameters. Then, DBSCAN is used to cluster the data. Normal data samples may form one or more high-density clusters, while faulty samples are marked as anomalies or form independent clusters. Finally, the DBSCAN output is examined to identify anomalies and independent clusters, uncovering potential fault modes. The core concept of DBSCAN is clustering based on the density of data points. It can identify sparse regions in the data as noise, which is very useful for discovering uncommon fault patterns. When fault samples are scarce, DBSCAN can help identify anomalous data points that differ significantly from normal samples, i.e., potential fault samples. However, in current research, the identification of outliers and independent clusters often requires secondary manual determination based on the actual situation, which is a semi-automatic process. Summary of the Invention

[0005] The purpose of this invention is to provide an improved DBSCAN method and system for early warning of vibration signals in dry-type transformers, so as to solve at least one of the technical problems existing in the background art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides an improved DBSCAN method for early warning of vibration signals in dry-type transformers, comprising:

[0008] Obtain the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested;

[0009] The acquired vibration signals are processed using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer of the actual urban rail power supply system and establishing an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis.

[0010] If the number of abnormal points in the fault diagnosis results exceeds the set threshold, an early warning will be issued.

[0011] As a further limitation of the first aspect of the present invention, the imbalanced dataset is collected during the normal operation of the urban rail transit system, and most of the data are normal signals with very few abnormal samples.

[0012] As a further limitation of the first aspect of the present invention, the filtering process needs to remove unnecessary interference in the environment. Since the actual acquisition environment is complex, there may be interference from train noise, human voices and other electrical equipment vibrations. Therefore, the db8 wavelet is used to perform wavelet threshold denoising on the signal.

[0013] As a further limitation of the first aspect of this invention, the key feature extraction specifically includes eight key features: standard deviation, kurtosis, spectral mean, spectral root mean square value, frequency centroid, linear frequency axis, power spectral entropy, and energy entropy, forming a (1000*8) feature matrix. The first two are time-domain features, the next four are frequency-domain spectral features, and the last two are entropy features.

[0014] As a further limitation of the first aspect of the present invention, the improved multi-objective parameter optimization mainly optimizes the values ​​of the two main parameters of DBSCAN, Eps and Minpts, and uses the global contour coefficient and DBI of clustering as the objective function for optimization, and the optimization algorithm adopts the simulated annealing algorithm.

[0015] As a further limitation of the first aspect of the present invention, the clustering analysis in S4 is mainly based on the density of data points, and is an unsupervised learning algorithm that does not rely on label information for training. If transformer fault samples form sparse density regions in the data space, while normal samples are in dense regions, DBSCAN can help identify these sparse regions, thereby marking possible fault modes.

[0016] Furthermore, the evaluation method in S5 uses silhouette coefficient and DBI as evaluation parameters, which can measure the cluster compactness and separation, respectively. Finally, the number of outliers is output, the results are observed, and a threshold is set for early warning.

[0017] Secondly, the present invention provides an improved DBSCAN dry-type transformer vibration signal fault early warning system, comprising:

[0018] The acquisition module is used to acquire the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested;

[0019] The processing module is used to process the acquired vibration signals using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals from the traction rectifier transformer of an actual urban rail power supply system to establish an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective parameter optimization algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis.

[0020] The early warning module is used to issue an early warning if the number of abnormal points in the fault diagnosis results exceeds the set threshold.

[0021] Thirdly, the present invention provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in the first aspect.

[0022] Fourthly, the present invention provides a computer device including a memory and a processor, wherein the processor and the memory communicate with each other, the memory stores program instructions executable by the processor, and the processor invokes the program instructions to execute the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in the first aspect.

[0023] Fifthly, the present invention provides an electronic device comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in the first aspect.

[0024] The beneficial effects of this invention are: by automatically optimizing DBSCAN parameters through multi-objective optimization and simulated annealing algorithm, the determination of outliers can be made to the most accurate state, thereby reducing the workload.

[0025] The advantages of additional aspects of the invention will be set forth more clearly in the following description or will be learned by practice of the invention. Attached Figure Description

[0026] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart of the improved DBSCAN dry-type transformer vibration signal fault diagnosis method according to an embodiment of the present invention.

[0028] Figure 2 This is a flowchart of the improved multi-objective optimization parameter algorithm described in an embodiment of the present invention.

[0029] Figure 3 This is a diagram of the DBSCAN clustering structure described in an embodiment of the present invention.

[0030] Figure 4 This is a flowchart of the improved DBSCAN algorithm described in an embodiment of the present invention. Detailed Implementation

[0031] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0032] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0033] It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as here.

[0034] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, and / or groups thereof.

[0035] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0036] To facilitate understanding of the present invention, the present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments. However, the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0037] Those skilled in the art should understand that the accompanying drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.

[0038] Example 1

[0039] In this embodiment 1, an improved DBSCAN dry-type transformer vibration signal fault early warning system is provided, including: an acquisition module for acquiring vibration signals of the traction rectifier transformer of the urban rail power supply system to be detected; a processing module for processing the acquired vibration signals using a pre-trained fault diagnosis model to obtain fault diagnosis results; and an early warning module for issuing an early warning if the number of abnormal points in the fault diagnosis results exceeds a set threshold.

[0040] In this embodiment, an improved DBSCAN dry-type transformer vibration signal fault early warning method is implemented using the above-described system. The method includes: acquiring the vibration signal of the traction rectifier transformer in the urban rail power supply system to be detected; processing the acquired vibration signal using a pre-trained fault diagnosis model to obtain fault diagnosis results; wherein, the training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer in the actual urban rail power supply system to establish an unbalanced dataset; performing filtering and key feature extraction preprocessing on the dataset; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis. If the number of abnormal points in the fault diagnosis results exceeds a set threshold, an early warning is issued.

[0041] The signal was denoised using wavelet thresholding with the db8 wavelet. Key feature extraction included eight features: standard deviation, kurtosis, spectral mean, spectral root mean square value, frequency centroid, linear frequency axis, power spectral entropy, and energy entropy, forming a 1000*8 feature matrix. Multi-objective parameter optimization was improved, including optimizing the values ​​of the two main DBSCAN parameters, Eps and Minpts, using the global silhouette coefficient and DBI of clustering as objective functions. The optimization algorithm employed simulated annealing. Cluster analysis included clustering based on data point density. If transformer fault samples formed sparse density regions in the data space, while normal samples were in dense regions, density-based clustering was used to identify these sparse regions and label the fault modes. The evaluation parameters of the fault diagnosis model were the silhouette coefficient and DBI, which respectively measured the cluster compactness and separation.

[0042] Example 2

[0043] like Figure 1 As shown in the figure, this embodiment 2 provides an improved DBSCAN method for early warning of vibration signals in dry-type transformers.

[0044] First, the acquired signals are extracted into an imbalanced dataset with an advance interval of 1000 and a default number of rows of 50, which can be increased based on the amount of data acquired. For filtering, db8 wavelets are used for denoising to remove unnecessary interference.

[0045] Secondly, key features are extracted from the dataset. Due to the unique characteristics of urban rail transit, signals vary with the periodic changes in load, so selecting appropriate features can help improve diagnostic accuracy. This invention combines multiple features for comprehensive analysis, selecting eight features to form a feature matrix: standard deviation, kurtosis, spectral mean, spectral root mean square value, frequency centroid, linear frequency axis, power spectral entropy, and energy entropy. The first two are time-domain features, the last four are frequency-domain features, and the last two are entropy features. The calculation formulas are listed below.

[0046]

[0047] Among them, X i Let n be the i-th sample point, and n be the total number of samples. S(f) is the mean of the signal, N is the total number of frequency components, and S(f) is the mean of the signal. i Let f be the power spectral density value of the i-th frequency component. max For the highest frequency, f min For the lowest frequency, P(f) i E is the normalized power spectral density value of the i-th frequency component. i Let be the energy of the i-th time window or frequency band.

[0048] Figure 2 The flowchart illustrates the improved multi-objective optimization parameter algorithm provided by this invention. Multi-objective optimization aims to simultaneously optimize multiple objective functions, further refining the parameters of DBSCAN.

[0049] First, we define two objective functions: F(1) global silhouette coefficient and F(2) Davies-Bouldin Index (DBI), which measure the cluster compactness and separation, respectively, and combine them into a comprehensive objective function.

[0050] The global silhouette coefficient is used to evaluate the quality of clustering results, measuring the difference between the similarity of each sample to other samples in the same cluster and the similarity to its nearest neighbor cluster. The silhouette coefficient value ranges from -1 to 1. A value close to 1 indicates that the sample is well clustered in the same cluster, a value close to 0 indicates that the sample is located on the boundary between two clusters, and a value close to 1 indicates that the sample is incorrectly clustered into other clusters. Its calculation formula is as follows:

[0051]

[0052] Where i is the number of samples, and its cluster is defined as C. i a(i) is the average distance from sample i to other samples in the same cluster (cluster density), and b(i) is the distance from sample i to its nearest cluster C. i The average distance (inter-cluster separation) of all samples within the cluster, where n is the total number of samples.

[0053] Dissimilarity in Clusters (DBI) is used to evaluate the separation and compactness of clusters. It calculates the similarity between each pair of clusters and takes the highest similarity value among all clusters as the metric. A lower DBI value indicates better clustering results, as it indicates better separation between clusters and higher compactness within clusters. The calculation formula is as follows:

[0054]

[0055] Among them, S i u is the average distance from the samples within the cluster to the cluster center. i As the center of the cluster, |C i | represents the number of samples in the cluster, ||μ i -μ j The distance between clusters i and j was calculated, where N is the total number of clusters. The average of all clusters was calculated by finding the maximum value of the Davies-Bouldin ratio for each pair of clusters.

[0056] Secondly, an optimization algorithm for simulated annealing was selected, and initial parameters were set, such as initial temperature, temperature decay rate, and initial solution. Detailed parameter setting ranges are shown in Table 1. The initial solutions Eps' and Minpts' were obtained through the algorithm.

[0057] Table 1. Specific initial parameter settings for the improved algorithm

[0058]

[0059]

[0060] Finally, the objective function is recalculated to obtain F(1)' and F(2)', which determines whether to accept the new solution and updates the current new solution and temperature. DBSCAN is retrained using the optimized parameters, and the quality of the clustering results is verified.

[0061] Figure 3 This is a DBSCAN clustering structure diagram provided in an embodiment of the present invention.

[0062] The DBSCAN algorithm, as a density-based unsupervised training algorithm, can adaptively determine the number of splits without setting a specific number of splits to achieve classification results. It performs well in handling imbalanced datasets. Two parameters for optimization are chosen: neighborhood and minimum number of neighborhood points.

[0063] Neighborhood (Eps): Defines the radius of a point's neighborhood, indicating that points within this radius will be considered neighborhood points.

[0064] Minimum number of neighboring points (Minpts): Defines the minimum number of neighboring points required for a point to become a core point.

[0065] Because the datasets differ, the parameters will also differ. Using the dataset provided in this invention, and after parameter optimization of the improved algorithm, the DBSCAN clustering result is as follows:

[0066] Initial DBSCAN parameters: eps = 1, min_samples = 8

[0067] Initial clustering results evaluation metrics: Silhouette Score = 0.29, Davies-Bouldin Index = 1.19

[0068] Optimized DBSCAN parameters: eps = 1.61, min_samples = 5.05

[0069] The optimized clustering results were evaluated using the Silhouette Score (Score) and Davies-Bouldin Index (DBI) of 0.63. The silhouette coefficient was closer to 1, indicating a higher density of samples. The lower DBI indicated a higher degree of separation between clusters and a higher density within clusters, suggesting a good clustering effect.

[0070] Figure 4 A flowchart of the improved DBSCAN algorithm provided for embodiments of the present invention.

[0071] The optimized parameters obtained from the above steps are then input into DBSCAN for cluster analysis. To evaluate the anomaly detection performance, two sets of signals were collected for comparison: one set was the normal signal D1, and the other set was the deliberately disturbed signal D2. Improved clustering processing was performed on both sets of signals, and the results are shown in Table 2.

[0072] Table 2 Improved DBSCAN clustering results

[0073]

[0074] The results show that the number of abnormal points in a normal signal generally does not exceed 5, while the number of abnormal points will significantly exceed 10 when there is a knock. Therefore, the alarm threshold can be set to 5, that is, when the number of abnormal points exceeds 5, a fault warning signal will be released to achieve the effect of automatic fault warning.

[0075] Example 3

[0076] This embodiment 3 provides a non-transitory computer-readable storage medium for storing computer instructions. When these computer instructions are executed by a processor, they implement the improved DBSCAN dry-type transformer vibration signal fault early warning method described above. The method includes:

[0077] Obtain the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested;

[0078] The acquired vibration signals are processed using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer of the actual urban rail power supply system and establishing an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis.

[0079] If the number of abnormal points in the fault diagnosis results exceeds the set threshold, an early warning will be issued.

[0080] Example 4

[0081] This embodiment 4 provides a computer device, including a memory and a processor, wherein the processor and the memory communicate with each other, and the memory stores program instructions that can be executed by the processor. The processor calls the program instructions to execute the improved DBSCAN dry-type transformer vibration signal fault early warning method as described above, the method including:

[0082] Obtain the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested;

[0083] The acquired vibration signals are processed using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer of the actual urban rail power supply system and establishing an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis.

[0084] If the number of abnormal points in the fault diagnosis results exceeds the set threshold, an early warning will be issued.

[0085] Example 5

[0086] This embodiment 5 provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions to implement the improved DBSCAN dry-type transformer vibration signal fault early warning method as described above, the method including:

[0087] Obtain the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested;

[0088] The acquired vibration signals are processed using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer of the actual urban rail power supply system and establishing an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis.

[0089] If the number of abnormal points in the fault diagnosis results exceeds the set threshold, an early warning will be issued.

[0090] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0091] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0092] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0093] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0094] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or variations that can be made by those skilled in the art without creative effort should be included within the scope of protection of the present invention.

Claims

1. An improved DBSCAN method for early warning of vibration signals in dry-type transformers, characterized in that, include: Obtain the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested; The acquired vibration signals are processed using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals of the traction rectifier transformer of the actual urban rail power supply system and establishing an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective optimization parameter algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis. If the number of abnormal points in the fault diagnosis results exceeds the set threshold, an early warning will be issued.

2. The improved DBSCAN dry-type transformer vibration signal fault early warning method according to claim 1, characterized in that, The signal is denoised using wavelet thresholding with db8 wavelet.

3. The improved DBSCAN dry-type transformer vibration signal fault early warning method according to claim 1, characterized in that, Key feature extraction includes eight key features: standard deviation, kurtosis, spectral mean, spectral root mean square value, frequency centroid, linear frequency axis, power spectral entropy, and energy entropy, which are used to establish a 1000*8 feature matrix.

4. The improved DBSCAN dry-type transformer vibration signal fault early warning method according to claim 1, characterized in that, Improved multi-objective parameter optimization includes optimizing the values ​​of the two main DBSCAN parameters, Eps and Minpts, and using the global silhouette coefficient and DBI of clustering as objective functions for optimization. The optimization algorithm adopted is the simulated annealing algorithm.

5. The improved DBSCAN dry-type transformer vibration signal fault early warning method according to claim 1, characterized in that, Cluster analysis includes clustering based on the density of data points. If transformer fault samples form sparse density regions in the data space, while normal samples are in dense regions, density-based clustering algorithms can be used to identify these sparse regions and label the fault patterns.

6. The improved DBSCAN dry-type transformer vibration signal fault early warning method according to claim 1, characterized in that, The evaluation parameters for the fault diagnosis model are the silhouette coefficient and DBI, which measure the compactness and separation of clusters, respectively.

7. An improved DBSCAN dry-type transformer vibration signal fault early warning system, characterized in that, include: The acquisition module is used to acquire the vibration signal of the traction rectifier transformer of the urban rail power supply system to be tested; The processing module is used to process the acquired vibration signals using a pre-trained fault diagnosis model to obtain fault diagnosis results. The training of the fault diagnosis model includes: collecting vibration signals from the traction rectifier transformer of an actual urban rail power supply system to establish an unbalanced dataset; preprocessing the dataset by filtering and extracting key features; building an improved multi-objective parameter optimization algorithm model to optimize the DBSCAN parameters; and inputting the optimized parameters into DBSCAN for cluster analysis. The early warning module is used to issue an early warning if the number of abnormal points in the fault diagnosis results exceeds the set threshold.

8. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in any one of claims 1-6.

9. A computer device, characterized in that, The method includes a memory and a processor, the processor and the memory communicating with each other, the memory storing program instructions executable by the processor, and the processor calling the program instructions to execute the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in any one of claims 1-6.

10. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions to implement the improved DBSCAN dry-type transformer vibration signal fault early warning method as described in any one of claims 1-6.

Images