Method and system for diagnosing mechanical state of on-load tap-changer of transformer
By performing feature extraction and Volterra coefficient matrix processing on the vibration signal of the transformer on-load tap-off switch, a support vector machine model is built, which solves the problem of transformer mechanical fault diagnosis and realizes online monitoring and real-time diagnosis.
Patent Information
- Application Number
- CN202411930034.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-04-15
AI Technical Summary
The on-load tap switch of the transformer is difficult to diagnose mechanical faults under high load and frequent voltage regulation, and it is difficult for the prior art to effectively monitor and identify fault types.
By collecting the vibration signals of the transformer, extracting the vibration characteristics, obtaining the Volterra coefficient matrix, and using the singular value of the matrix as the feature quantity, a support vector machine model is built to classify and identify the mechanical faults of the transformer.
It realizes online monitoring and real-time diagnosis of the on-load tap-off switch of the transformer, significantly improves the accuracy of identifying mechanical faults, and has important engineering guidance significance.
Smart Images
Figure CN119827129A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to detection of the operating state of an on-load tap changer of a transformer, and in particular, to a method and system for diagnosing the mechanical state of an on-load tap changer of a transformer. Background Art
[0002] In the current power system construction that vigorously promotes renewable energy generation, on-load tap-changing transformers face the requirements of high operating load and wide voltage regulation range, which increases the risk of equipment defects. At the same time, the impact and intermittent characteristics of distributed power grid connection require frequent voltage adjustment to ensure the active and reactive power distribution and power supply quality of the system, which poses a severe challenge to the operational reliability of OLTC. According to transformer accident statistics, failures caused by on-load tap changers account for more than 18%, so OLTC mechanical fault diagnosis has become an important research topic at present. Summary of the invention
[0003] In order to solve the above problems, an object of the present invention is to provide a diagnosis technology for the mechanical state of the transformer on-load tap changer, which is used to monitor the switching state of the transformer on-load tap changer and analyze its fault type.
[0004] In order to achieve the above technical objectives, the present application provides a method for diagnosing the mechanical state of a transformer on-load tap changer, comprising the following steps:
[0005] Collect the vibration signal of the transformer and extract the vibration characteristics;
[0006] Based on the vibration characteristics, the Volterra coefficient matrix is obtained, and the singular values of the matrix are used as feature quantities to construct a support vector machine model for classification and identification of mechanical faults of transformers.
[0007] Preferably, in the process of collecting vibration signals, the collected vibration signals are stored through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.
[0008] Preferably, in the process of extracting vibration features, based on the EEMD algorithm, adaptive white noise is used to eliminate modal aliasing, a complete set empirical mode decomposition algorithm is formed, and features of the vibration signal are extracted to obtain vibration features.
[0009] Preferably, in the process of obtaining the Volterra coefficient matrix, a model is built based on the vibration characteristics by PCA principal component analysis to form the Volterra coefficient matrix, wherein the vibration characteristics are the vibration characteristics of the on-load tap changer, including loose moving contacts, loose stationary contacts and motor jamming.
[0010] Preferably, in the process of constructing the Volterra coefficient matrix, the mechanical vibration signal is decomposed by a complete set empirical mode decomposition algorithm to obtain the intrinsic mode function IMF of each frequency component, and the Volterra series expansion is applied to each IMF component to construct the corresponding Volterra model for each component, and the Volterra model coefficient vectors in the IMF components are arranged in rows to form a Volterra coefficient matrix.
[0011] Preferably, in the process of obtaining the singular values, singular value decomposition (SVD) is performed on the Volterra coefficient matrix to obtain the singular values of the eigenvectors.
[0012] Preferably, in the process of establishing the support vector machine model, based on the support vector machine classification model PCA-SVM, singular values are used as feature quantities to train the model and construct the support vector machine model.
[0013] Preferably, in the process of classifying and identifying mechanical faults of transformers, the vibration signal to be analyzed is collected, filtered and reconstructed, the coefficient matrix of the first 10-order modal components is constructed, and the first five-order singular values are calculated, so as to classify and identify mechanical faults through a support vector machine model.
[0014] The present invention discloses a system for diagnosing the mechanical state of an on-load tap changer of a transformer, which is used to implement the above-mentioned method for diagnosing the mechanical state of an on-load tap changer of a transformer, comprising:
[0015] A data acquisition module, used to collect vibration signals of the transformer;
[0016] A feature extraction module, used for extracting vibration features according to the vibration signal;
[0017] The classification and identification module is used to obtain the Volterra coefficient matrix based on the vibration characteristics, and use the singular values of the matrix as feature quantities to build a support vector machine model for classifying and identifying the mechanical faults of the transformer.
[0018] Preferably, the system also includes a data storage module for storing the collected vibration signals through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.
[0019] The present invention discloses the following technical effects:
[0020] The present invention has important engineering guiding significance for online monitoring and real-time diagnosis of transformer on-load tap changers. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0022] Figure 1 It is a method for diagnosing the mechanical state of the transformer on-load tap changer and a system structure diagram according to the present invention;
[0023] Figure 2 This is a flow chart of the storage method of the present invention;
[0024] Figure 3 A diagram of a method for diagnosing the mechanical state of an on-load tap changer according to the present invention;
[0025] Figure 4 This is a diagram of the on-load tap changer vibration monitoring system according to the present invention;
[0026] Figure 5 It is the decomposition diagram of the vibration signal CEEMDAN of the present invention;
[0027] Figure 6 This is a diagram showing the fault diagnosis results of the on-load tap changer described in the present invention. DETAILED DESCRIPTION
[0028] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.
[0029] The present invention provides a method for diagnosing the mechanical state of a transformer on-load tap changer, comprising the following steps:
[0030] Collect the vibration signal of the transformer and extract the vibration characteristics;
[0031] Based on the vibration characteristics, the Volterra coefficient matrix is obtained, and the singular values of the matrix are used as feature quantities to construct a support vector machine model for classification and identification of mechanical faults of transformers.
[0032] Further preferably, in the process of collecting vibration signals, the diagnostic method provided by the present invention stores the collected vibration signals through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.
[0033] Further preferably, in the process of extracting vibration features, the diagnostic method provided by the present invention eliminates modal aliasing through adaptive white noise based on the EEMD algorithm, forms a complete set empirical mode decomposition algorithm, extracts features from the vibration signal, and obtains vibration features.
[0034] Further preferably, in the diagnostic method provided by the present invention, in the process of obtaining the Volterra coefficient matrix, based on the vibration characteristics, a model is formed by the PCA principal component analysis method, wherein the vibration characteristics are the vibration characteristics of the on-load tap changer, including loose moving contacts, loose stationary contacts and motor jamming.
[0035] Further preferably, in the diagnostic method provided by the present invention, in the process of constructing the Volterra coefficient matrix, the mechanical vibration signal is decomposed by a complete set empirical mode decomposition algorithm to obtain the intrinsic mode function IMF of each frequency component, and the Volterra series expansion is applied to each IMF component to construct a corresponding Volterra model for each component, and the Volterra model coefficient vectors in the IMF components are arranged in rows to form a Volterra coefficient matrix.
[0036] Further preferably, in the process of obtaining singular values, the diagnostic method provided by the present invention performs singular value decomposition (SVD) on the Volterra coefficient matrix to obtain the singular values of the eigenvectors.
[0037] Further preferably, in the process of establishing a support vector machine model, the diagnostic method provided by the present invention uses singular values as feature quantities to train the model based on the support vector machine classification model PCA-SVM to construct a support vector machine model.
[0038] Further preferably, the diagnostic method provided by the present invention, in the process of classifying and identifying the mechanical faults of the transformer, collects the vibration signal to be analyzed, performs filtering reconstruction, constructs the coefficient matrix of the first 10-order modal components, and calculates the first five-order singular values, so as to classify and identify the mechanical faults through a support vector machine model.
[0039] The present invention discloses a system for diagnosing the mechanical state of an on-load tap changer of a transformer, which is used to implement the above-mentioned method for diagnosing the mechanical state of an on-load tap changer of a transformer, comprising:
[0040] A data acquisition module, used to collect vibration signals of the transformer;
[0041] A feature extraction module, used for extracting vibration features according to the vibration signal;
[0042] The classification and identification module is used to obtain the Volterra coefficient matrix based on the vibration characteristics, and use the singular values of the matrix as feature quantities to build a support vector machine model for classifying and identifying the mechanical faults of the transformer.
[0043] Further preferably, the diagnostic system disclosed in the present invention also includes a data storage module for storing the collected vibration signals through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.
[0044] Example: Figure 1-6 As shown, the present invention provides a method for diagnosing the mechanical state of a transformer on-load tap changer, comprising the following steps:
[0045] Data measurement phase: Several LC0144S acceleration vibration sensors are arranged on the on-load tap changer, and the monitoring device containing power supply, mainboard, Bluetooth and other components is fixed on the transformer, and the vibration signal is transmitted to the server via Bluetooth.
[0046] Data triggering stage: When the current sensor collects the motor driving current, it starts to trigger the data acquisition module to collect the vibration signal.
[0047] Data storage stage: Figure 2 As shown, a circular buffer can be used in combination with multi-threading technology to ensure the continuity of data collection and reduce data loss.
[0048] Data transfer phase: During the operation of the ring buffer, multithreading technology is used to parallelize the acquisition and reading of data. The acquisition thread continuously writes new data into the buffer, while the reading thread extracts data from the buffer and transfers it to the memory area set by the program.
[0049] Data processing stage: Figure 3 As shown in the figure, first, the complete set empirical mode decomposition technique is applied to the vibration signal of the on-load tap changer to extract the vibration characteristics of the OLTC. Then, the Volterra model is established for each intrinsic mode function IMF component obtained by decomposition, and the Volterra coefficient matrix is generated. The singular values of the matrix are calculated as the characteristic quantities of the OLTC mechanical fault. Based on these characteristic quantities, a support vector machine model is constructed to realize the classification and identification of OLTC mechanical faults.
[0050] Result verification stage: After the system is built, it is assembled on the on-load tap-changer. After a routine fault occurs in the on-load tap-changer, the detection system is tested to see whether it can correctly determine the fault, and the measurement accuracy of the fault monitoring system based on the decomposition and extraction of the transformer on-load tap-changer vibration signal is verified.
[0051] When the current sensor collects the motor drive current, it starts to trigger the data acquisition module to collect vibration signals, thereby reducing the power consumption of the equipment and increasing the service life of the equipment.
[0052] Combined with the actual field application, in order to minimize the interference of transformer vibration signals in actual operation and facilitate installation, three sensors are arranged on the top and two side surfaces of the transformer oil tank wall to obtain mechanical vibration signals in three different directions perpendicular to each other during the switching process of the on-load tap changer.
[0053] A vibration monitoring system for on-load tap changer has been developed to obtain vibration signals during its switching process, such as Figure 4 shown.
[0054] Although traditional empirical mode decomposition (EMD) is used for vibration signal processing, it has the problem of modal aliasing when processing nonlinear and non-stationary signals. Therefore, the CEEMDAN algorithm is optimized on the basis of EEMD, and adaptive white noise is added to eliminate modal aliasing, which improves the reliability of the decomposition results. The vibration signal is decomposed by CEEMDAN, and each IMF component is as follows: Figure 5 shown.
[0055] This paper uses the PCA principal component analysis method to model the vibration characteristics of the on-load tap changer, and takes the five singular values of the Volterra coefficient matrix as the input sequence. The faults of the on-load tap changer include loose moving contacts, loose static contacts, and motor jams. The data set contains a total of 60 samples, including 6 normal samples and 54 faulty samples.
[0056] The method for extracting OLTC mechanical fault characteristics mainly includes the following steps: First, the mechanical vibration signal of OLTC during the switching process is decomposed by CEEMDAN to obtain the intrinsic mode function (IMF) of each frequency component. Then, the Volterra series expansion is applied to each IMF component to establish a coefficient vector matrix, and the singular value is calculated as the classification basis to identify the type of mechanical fault of OLTC. The specific steps are as follows:
[0057] 1. Establish a Volterra model for each IMF component: After obtaining each IMF component, build a corresponding Volterra model for each component. The Volterra model is a nonlinear prediction model that is suitable for processing complex and non-stationary signals. The nonlinear characteristics in the IMF components can be extracted through this model. In order to make the prediction more accurate and applicable, the normalized least mean square adaptive algorithm is used in the model to solve the coefficient vector.
[0058] 2. Form the Volterra coefficient matrix: Arrange the Volterra model coefficient vectors in the IMF components in rows to form a multidimensional matrix, namely the Volterra coefficient matrix. This method can effectively aggregate the characteristic information of all IMF components and provide rich signal details for further analysis and identification.
[0059] 3. Calculate the singular values of the matrix: By performing singular value decomposition (SVD) on the Volterra coefficient matrix, the singular values of the eigenvector can be obtained. Since the singular values can represent the energy distribution and signal characteristics of the system, they can be effectively used to quantify the characteristics of mechanical faults. Under different fault conditions, the distribution of singular values will be different, which can provide clear fault feature input for subsequent classifiers.
[0060] 4. Apply fault classifier for fault identification: After extracting singular values as features, input them into the support vector machine (PCA-SVM) classification model, such as Figure 6 As shown in the figure, through data learning and verification of the training set and the test set, various typical mechanical faults in OLTC, such as loose moving contacts, loose static contacts, and motor jamming, can be effectively identified, significantly improving the accuracy of fault diagnosis.
[0061] according to Figure 5As shown in the figure, in the process of processing the vibration signal, through the analysis of the decomposition results, it is found that the main characteristic information of the signal is concentrated in the intrinsic mode function components IMF1 to IMF10. The characteristics of the jamming fault are mainly distributed in IMF1-IMF7, the characteristics of the loosening fault of the moving contact are mainly distributed in IMF1-IMF8, and the characteristics of the loosening fault of the static contact are mainly distributed in IMF1-IMF10. Based on this, the signal is filtered and reconstructed to construct the coefficient matrix of the first 10 modal components. Based on this observation, the first ten IMF components are selected to construct the Volterra coefficient matrix to fully retain the key information of the signal. After further calculating the singular values of the matrix, it is found that the main features are concentrated in the first five singular values, so the first five singular values are used as the key characteristic values to characterize the fault of the on-load tap changer. This processing strategy significantly simplifies the complexity of the model. On the one hand, by focusing on representative singular values, the data dimension is reduced, thus simplifying the subsequent feature extraction and model input process; on the other hand, this feature selection method improves the computational efficiency while effectively improving the PCA-SVM model's speed and accuracy in distinguishing OLTC faults. This not only optimizes the diagnostic process, but also provides theoretical support and practical value for real-time fault detection.
[0062] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0063] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0064] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for diagnosing the mechanical state of a transformer on-load tap changer, characterized in that: The following steps are involved: Collect the vibration signal of the transformer and extract the vibration characteristics; Based on the vibration characteristics, a Volterra coefficient matrix is obtained, and the singular values of the matrix are used as feature quantities to construct a support vector machine model for classifying and identifying mechanical faults of the transformer.
2. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 1, characterized in that: In the process of collecting vibration signals, the collected vibration signals are stored through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.
3. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 2, characterized in that: In the process of extracting vibration features, based on the EEMD algorithm, adaptive white noise is used to eliminate modal aliasing, forming a complete set empirical mode decomposition algorithm, performing feature extraction on the vibration signal, and obtaining the vibration features.
4. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 3, characterized in that: In the process of obtaining the Volterra coefficient matrix, based on the vibration characteristics, a model is built by the PCA principal component analysis method to form the Volterra coefficient matrix, wherein the vibration characteristics are vibration characteristics of the on-load tap changer, including loose moving contacts, loose stationary contacts and motor jamming.
5. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 4, characterized in that: In the process of constructing the Volterra coefficient matrix, the mechanical vibration signal is decomposed by the complete set empirical mode decomposition algorithm to obtain the intrinsic mode function IMF of each frequency component, and the Volterra series expansion is applied to each IMF component to construct the corresponding Volterra model for each component, and the Volterra model coefficient vectors in the IMF component are arranged in rows to form the Volterra coefficient matrix.
6. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 5, characterized in that: In the process of obtaining the singular values, singular value decomposition (SVD) is performed on the Volterra coefficient matrix to obtain the singular values of the eigenvectors.
7. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 6, characterized in that: In the process of establishing the support vector machine model, based on the support vector machine classification model PCA-SVM, singular values are used as feature quantities to train the model to construct the support vector machine model.
8. A method for diagnosing the mechanical state of a transformer on-load tap changer according to claim 7, characterized in that: In the process of classifying and identifying mechanical faults of transformers, the vibration signal to be analyzed is collected, filtered and reconstructed, the coefficient matrix of the first 10-order modal components is constructed, and the first five-order singular values are calculated, so as to classify and identify mechanical faults through the support vector machine model.
9. A system for diagnosing the mechanical state of an on-load tap changer of a transformer, used to implement a method for diagnosing the mechanical state of an on-load tap changer of a transformer as claimed in any one of claims 1 to 8, characterized in that: include: A data acquisition module, used to collect vibration signals of the transformer; A feature extraction module, used to extract vibration features according to the vibration signal; The classification and identification module is used to obtain the Volterra coefficient matrix based on the vibration characteristics, and use the singular values of the matrix as feature quantities to construct a support vector machine model for classifying and identifying the mechanical faults of the transformer.
10. A diagnostic system for the mechanical state of a transformer on-load tap changer according to claim 9, characterized in that: The system also includes a data storage module, which is used to store the collected vibration signal through a ring buffer combined with multi-threading technology to ensure the continuity of data collection and reduce data loss.