A method and system for identifying operational faults in high-voltage current transformers

Abstract

This invention discloses a method and system for fault identification in high-voltage current transformers. The method includes: acquiring high-voltage current transformer operating data; performing dimensionality reduction processing on the operating data using a principal component analysis model to obtain new principal component variables; monitoring the principal component variables using a T2 control chart model to detect the existence of outliers; if the T2 control chart model detects outliers, using a Q control chart model to determine the fault points among the outliers and providing fault identification results. Using this invention, faults can be detected quickly and accurately even with complex data and diverse fault modes, and detailed fault identification results can be provided, which is of great significance for ensuring the safe and stable operation of power systems.

Description

Technical Field

[0001] This invention belongs to the field of current transformer technology, and in particular to a method and system for identifying operational faults in high-voltage current transformers. Background Technology

[0002] High-voltage current transformers (CTs) play a crucial role in power systems. They are primarily used to convert high voltage and high current into standard low voltage and low current signals to facilitate the safe operation of measurement, protection, and control equipment. CTs are widely used in all aspects of the power system, including generation, transmission, substation, and distribution systems; they can be considered the "sensing organs" of the power system. Monitoring and ensuring the normal operation of CTs is one of the keys to guaranteeing the safe and stable operation of the power system.

[0003] Despite their sophisticated design and advanced manufacturing processes, high-voltage current transformers can still experience various faults during long-term operation. Failure to detect and address these faults promptly can not only affect the accuracy of electricity metering but also damage protection and control equipment, and in severe cases, even trigger large-scale power outages. Therefore, establishing an efficient and reliable method for diagnosing operational faults in high-voltage current transformers is of paramount importance and urgency. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for identifying faults in high-voltage current transformers, in order to overcome the shortcomings of existing technologies. This method can quickly and accurately detect faults and provide detailed fault identification results even under complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of power systems.

[0005] One embodiment of this application provides a method for identifying operational faults in a high-voltage current transformer, the method comprising:

[0006] Obtain operating data of high-voltage current transformers;

[0007] Principal component analysis model is used to reduce the dimensionality of the running data to obtain new principal component variables;

[0008] The T2 control chart model is used to monitor the principal component variables and detect the existence of outliers;

[0009] If the T2 control chart model detects outliers, the Q control chart model is used to determine the fault points among the outliers and provide fault identification results.

[0010] Optionally, the step of using a T2 control chart model to monitor principal component variables and detect the existence of outliers includes:

[0011] Obtain a new principal component variable matrix mathbf{Z}, the matrix mathbf{Z} having dimensions n*p, where n is the number of samples and p is the number of principal components;

[0012] The principal component variable matrix and each sample point are centered to obtain the centered principal component variable matrix and the centered vector of the sample points:

[0013] mathbf{Z}_{centered}=mathbf{Z}-mathbf{bar{Z}}

[0014] mathbf{z}_i_{centered}=mathbf{z}_i-·mathbf{bar{Z}}

[0015] Where, mathbf{Z}_{centered} is the centered principal component variable matrix, mathbf{z}_i^{centered} is the centered vector of the i-th sample point, and mathbf{bar{Z}} is the mean vector of each column in mathbf{Z}, where one column corresponds to one principal component and one sample point corresponds to one row;

[0016] Calculate the covariance matrix of the centered principal component variable matrix:

[0017]

[0018] Where, mathbf{Z}_{centered}^T represents the transpose of mathbf{Z}_{centered};

[0019] Calculate the T2 statistic using the centered vector and covariance matrix of the sample points:

[0020] T2_i=mathbf{z}_i_{centered}·mathbf{Sigma}_{Z}^{-1}·mathbf{z}_i_{centered}^T

[0021] Where T2_i is the T2 statistic of the i-th sample point, mathbf{Sigma}_{Z}^{-1} is the inverse matrix of mathbf{Sigma}_{Z}, and mathbf{z}_i_{centered}^T is the transpose of mathbf{z}_i_{centered}.

[0022] The T2 control limits are calculated as follows:

[0023] T2_{lim}=chi^2_{p,alpha}

[0024] Where T2_{lim} is the T2 control limit, chi^2 represents the chi-square distribution, chi^2_{p,alpha} represents the critical value of the chi^2 distribution with cumulative probability 1-alpha when the degrees of freedom are p, and alpha is the confidence level;

[0025] For each sample point, the T2 statistic T2_i is compared with the control limit T2_{lim}. Sample points whose T2 statistic is greater than the T2 control limit are identified as outliers.

[0026] Optionally, the step of using the Q-control chart model to determine the fault point among outliers includes:

[0027] For each outlier point mathbf{z}_j, calculate the reconstructed value for that outlier point, which is then reconstructed using the retained principal components:

[0028] hat{mathbf{z}_j}

[0029] =mathbf{P}·mathbf{P}^T·(·mathbf{z}_j

[0030] -mathbf{bar{Z}})+mathbf{bar{z}}

[0031] Where, mathbf{P} is the load matrix with principal components, and mathbf{P}^T is the transpose of the load matrix;

[0032] Calculate the residual vector for each outlier:

[0033] mathbf{e}_j=mathbf{z}_j-hat{mathbf{z}_j}

[0034] Based on the residual vector, calculate the Q statistic for each outlier:

[0035] Q_j=mathbf{e}_j^T·mathbf{e}_j

[0036] Where, mathbf{e}_j^T is the transpose of the residual vector mathbf{e}_j;

[0037] The Q control limits are determined as follows:

[0038]

[0039] Where z_{beta} is the critical value of the beta level of the standard normal distribution, and h is:

[0040]

[0041] in: lambda_q is the qth eigenvalue of the covariance matrix mathbf{Sigma}_{Z}, and m is the total number of variables;

[0042] For each outlier, the Q statistic Q_j is compared with the Q control limit Q_{lim}. Outliers with Q statistic greater than the Q control limit are identified as fault points.

[0043] For a fault point, analyze each component of the residual vector mathbf{e}_j of the fault point. If there is a component whose Q statistic is greater than the Q control limit, then determine the fault variable from the components of the principal component variable matrix corresponding to that component.

[0044] Optionally, providing the fault determination result includes:

[0045] The detected fault variables are compared with the features of each fault mode in the fault mode library to determine the best matching fault mode.

[0046] Another embodiment of this application provides a high-voltage current transformer operation fault identification system, the system comprising:

[0047] The acquisition module is used to acquire operating data of the high-voltage current transformer;

[0048] The dimensionality reduction module is used to reduce the dimensionality of the running data using the principal component analysis model to obtain new principal component variables;

[0049] The detection module is used to monitor the principal component variables using the T2 control chart model and detect the presence of outliers.

[0050] The determination module is used to identify fault points among outliers when outliers are detected by the T2 control chart model, and to provide fault identification results.

[0051] Another embodiment of this application provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described in any of the preceding claims when running.

[0052] Another embodiment of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described in any of the preceding claims.

[0053] Compared with existing technologies, this invention provides a method for fault identification in high-voltage current transformers. It acquires high-voltage current transformer operating data; uses a principal component analysis model to reduce the dimensionality of the operating data, obtaining new principal component variables; monitors the principal component variables using a T2 control chart model to detect outliers; if the T2 control chart model detects outliers, it uses a Q control chart model to determine the fault points within the outliers and provides fault identification results. This method can quickly and accurately detect faults and provide detailed fault identification results even with complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of power systems. Attached Figure Description

[0054] Figure 1 Hardware structure block diagram of a computer terminal for a high-voltage current transformer operation fault identification method provided in an embodiment of the present invention;

[0055] Figure 2 A flowchart illustrating a method for identifying operational faults in a high-voltage current transformer, provided in an embodiment of the present invention.

[0056] Figure 3 This is a schematic diagram of a high-voltage current transformer operation fault identification system provided in an embodiment of the present invention. Detailed Implementation

[0057] 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.

[0058] This invention first provides a method for identifying operational faults in high-voltage current transformers. This method can be applied to electronic devices, such as computer terminals, specifically ordinary computers.

[0059] The following detailed explanation uses a computer terminal as an example. Figure 1 This is a hardware structure block diagram of a computer terminal for a high-voltage current transformer operation fault diagnosis method provided in an embodiment of the present invention. Figure 1 As shown, a computer terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. Optionally, the computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the computer terminal described above. For example, the computer terminal may also include components that are more complex than those described above. Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0060] The memory 104 can be used to store software programs and modules for application software, such as the program instructions / modules corresponding to the high-voltage current transformer operation fault detection method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0061] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider for the computer terminal. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0062] See Figure 2 The present invention provides a method for identifying operational faults in high-voltage current transformers, which may include the following steps:

[0063] S201, Obtain high-voltage current transformer operating data; specifically, one implementation method may include:

[0064] Step 1: Design and Deployment of the Data Acquisition System

[0065] 1. Install smart sensors at key nodes of high-voltage current transformers. These sensors can capture data in multiple dimensions such as current, voltage, and temperature.

[0066] 2. The data collected by the sensors is transmitted to the centralized control system in real time via a wireless transmission module (such as ZigBee, LoRa or 5G).

[0067] Step 2: Data Preprocessing

[0068] 1. Noise Removal: Wavelet transform or Kalman filtering are used to remove noise from sensor data to ensure data stability and accuracy.

[0069] 2. Data calibration: The sensor is calibrated using a known standard signal source to eliminate the sensor's own deviations and errors.

[0070] 3. Data interpolation: For missing data points, Lagrange interpolation or Kriging interpolation is used to interpolate the data to ensure data integrity.

[0071] Step 3: Data storage

[0072] 1. Data storage format: Store the pre-processed data in a distributed database (such as HBase) to ensure high availability and scalability of the data.

[0073] 2. Data compression: Time series data compression algorithms (such as the Gorilla algorithm) are used to compress the data, saving storage space.

[0074] Step 4: Data Quality Supervision

[0075] 1. Data quality monitoring: Real-time monitoring of data quality, including data integrity, accuracy, and consistency.

[0076] 2. Anomaly Reporting Mechanism: If a data quality problem is detected, a report is generated in real time and the operation and maintenance personnel are notified through the alarm system.

[0077] Through the above steps, the operating data of the high-voltage current transformer will be acquired in a high-quality and multi-dimensional format, providing a solid data foundation for subsequent principal component analysis and fault diagnosis. This data acquisition method not only improves the reliability and accuracy of the data but also captures more potential anomalies, greatly enhancing the accuracy and efficiency of fault diagnosis.

[0078] S202, using the principal component analysis model to reduce the dimensionality of the running data to obtain new principal component variables;

[0079] Specifically, Principal Component Analysis (PCA) is a classic dimensionality reduction technique, typically used to transform high-dimensional data into low-dimensional data while preserving as much of the key information as possible. Traditional PCA methods may neglect the temporal and spatial characteristics of the data. Therefore, this application proposes an improved PCA method that combines time series decomposition extraction, making it more robust and effective in processing high-voltage current transformer operating data. One implementation may include:

[0080] Step 1: Data Preprocessing

[0081] 1. Data Standardization: The operating data of the high-voltage current transformer is standardized so that the mean of each variable is 0 and the variance is 1. This step can eliminate the influence of dimensions and improve the effect of principal component analysis.

[0082]

[0083] Where, mathbf{X}_{standardized} is the standardized data, mathbf{X} is the original data matrix of the running data, and mu_{mathbf{X}} and sigma_{mathbf{X}} are the mean and standard deviation of each column, respectively.

[0084] Step 2: Time Series Decomposition

[0085] 1. Data Denoising: Wavelet transform is used to denoise the standardized data, resulting in smoother data.

[0086] mathbf{X}_{denoised}

[0087] =wavelet_transform_denoise(marhbf{X}_{standardized})

[0088] Where, mathbf{X}_{denoised} is the denoised data.

[0089] 2. Time Series Decomposition: The Empirical Mode Decomposition (EMD) technique is used to decompose the time series data into a series containing several intrinsic mode functions (IMFs) and a remainder term.

[0090] mathbf{X}_{EMD}=EMD(mathbf{X}_{denoised}

[0091] Where mathbf{X}_{EMD} is time series data.

[0092] Step 3: Principal Component Analysis

[0093] 1. Covariance matrix calculation: Calculate the covariance matrix mathbf{Sigma} of the time series data after denoising and time series decomposition.

[0094] 2. Eigenvalue and Eigenvector Decomposition: The covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues ​​and corresponding eigenvectors.

[0095] mathbf{Sigma}·mathbf{v}_i=lambda_i·mathbf{v}_i

[0096] Where lambda_i and mathbf{v}_i are the i-th eigenvalue and eigenvector of the covariance matrix, respectively.

[0097] 3. Eigenvalue selection: Select the eigenvectors corresponding to the p largest eigenvalues ​​to construct the principal component loading matrix mathbf{P}:

[0098] mathbf{P}=[mathbf{v}_1,mathbf{v}_2,...,mathbf{v}_p]

[0099] Step 4: Calculation of principal component variables

[0100] 1. Principal component variable matrix: Projecting the time series data X_EMD onto the principal component loading matrix P, we obtain the principal component variable matrix Z:

[0101] mathbf{Z}=mathbf{X}_{EMD}mathbf{P}

[0102] Where, mathbf{Z} is the new principal component variable matrix.

[0103] Through the above steps, not only is the traditional PCA method utilized, but time series decomposition and spatial feature extraction are also combined, making the dimensionality reduction process more comprehensive and accurate, thus obtaining more meaningful principal component variables. This method can better capture the underlying structure and relationships in the data, laying the foundation for subsequent fault identification.

[0104] S203 uses the T2 control chart model to monitor principal component variables and detect the existence of outliers;

[0105] Specifically, a new principal component variable matrix mathbf{Z} can be obtained, wherein the dimension of the matrix mathbf{Z} is n*p, where n is the number of samples and p is the number of principal components;

[0106] Principal Component Analysis (PCA) transforms original high-dimensional data into a new principal component variable matrix, `mathbf{Z}`, through dimensionality reduction. In this matrix, each row represents a sample, and each column represents a principal component. The dimensionality-reduced matrix retains most of the information from the data while reducing its complexity and dimensionality.

[0107] The principal component variable matrix and each sample point are centered to obtain the centered principal component variable matrix and the centered vector of the sample points:

[0108] mathbf{Z}_{centered}=mathbf{Z}-mathbf{bar{Z}}

[0109] mathbf{z}_i_{centered}=mathbf{z}_i-mathbf{bar{Z}}

[0110] Where, mathbf{Z}_{centered} is the centered principal component variable matrix, mathbf{z}_i^{centered} is the centered vector of the i-th sample point, and mathbf{bar{Z}} is the mean vector of each column in mathbf{Z}, where one column corresponds to one principal component and one sample point corresponds to one row;

[0111] Centering is performed to eliminate mean differences between different samples, making the data more comparable. By subtracting the mean vector `mathbf{bar{Z}}` from each principal component variable, the mean of each column in the resulting centered matrix `mathbf{Z}_{centered}` is zero. The same process is applied to each sample point `mathbf{z}_i`, resulting in a centered vector `mathbf{z}_i^{centered}`, which helps avoid bias in subsequent calculations.

[0112] Calculate the covariance matrix of the centered principal component variable matrix:

[0113]

[0114] Where, mathbf{Z}_{centered}^T represents the transpose of mathbf{Z}_{centered};

[0115] The covariance matrix (mathbf{Sigma}_Z) describes the linear relationship and variability among the different principal components. Calculating the covariance matrix using the centered principal component variable matrix (mathbf{Z}_{centered}) reveals the internal structure of the data. The covariance matrix is ​​then used in subsequent steps to calculate the T2 statistic, which reflects the covariance information among the principal components.

[0116] Calculate the T2 statistic using the centered vector and covariance matrix of the sample points:

[0117] T2_i=mathbf{z}_i_{centered}mathbf{Sigma}_{Z}^{-1}

[0118] ·mathbf{z}_i_{centered}^T

[0119] Where T2_i is the T2 statistic of the i-th sample point, mathbf{Sigma}_{Z}^{-1} is the inverse matrix of mathbf{Sigma}_{Z}, and mathbf{z}_i_{centered}^T is the transpose of mathbf{z}_i_{centered}.

[0120] The T² statistic is calculated using the inverse of the centered vector (mathbf{z}_i^{centered}) and the covariance matrix (mathbf{Sigma}_Z) of the sample points. This statistic measures the degree of deviation of the sample points from the mean in a multidimensional space. Specifically, the T² statistic T²_i represents the distance of a sample point from the population mean, taking into account the covariance information among the principal components. A larger T² statistic indicates a greater deviation of the sample points from the mean.

[0121] The T2 control limits are calculated as follows:

[0122] T2_{lim}=chi^2_{p,alpha}

[0123] Where T2_{lim} is the T2 control limit, chi^2 represents the chi-square distribution, chi^2_{p,alpha} represents the critical value of the chi^2 distribution with cumulative probability 1-alpha when the degrees of freedom are p, and alpha is the confidence level;

[0124] The T2 control limit, T2_{lim}, is calculated based on the critical value of the chi-square distribution and is used to determine whether the T2 statistic is within the normal range. The critical value of the chi-square distribution depends on the set confidence level and the number of principal components. In practical applications, an appropriate confidence level (such as 95% or 99%) is selected to determine the control limit, thereby distinguishing normal points from outliers. Setting reasonable control limits can effectively improve the accuracy of outlier detection.

[0125] For each sample point, the T2 statistic T2_i is compared with the control limit T2_{lim}. Sample points whose T2 statistic is greater than the T2 control limit are identified as outliers.

[0126] We determine whether a sample point is an outlier by comparing its T2 statistic with the T2 control limit. If T2_i exceeds T2_{lim}, the sample point is considered an outlier, indicating a significant deviation from the mean. Outliers may indicate potential anomalies or malfunctions, requiring further analysis and processing.

[0127] The explanations of each step above provide a comprehensive understanding of their specific role in the fault monitoring process. This method ensures the detection of anomalies in the principal component variables during the operation of high-voltage current transformers, enabling effective identification and monitoring of outliers.

[0128] S204. If the T2 control chart model detects outliers, the Q control chart model is used to determine the fault points among the outliers and to provide fault identification results.

[0129] Specifically, for each outlier point mathbf{z}_j, the reconstructed value of that outlier point is calculated, and the reconstructed value is rebuilt using the retained principal components:

[0130] hat{mathbf{z}_j}

[0131] =mathbf{P}·mathbf{P}^T·(mathbf{z}_j

[0132] -mathbf{bar{Z}})+mathbf{bar{Z}}

[0133] Where, mathbf{P} is the load matrix with principal components, and mathbf{P}^T is the transpose of the load matrix;

[0134] The reconstructed values ​​of outliers are obtained by reconstructing the principal components. Here, the loading matrix of the principal components, `mathbf{P}`, is used for reconstruction. First, the mean `mathbf{bar{Z}}` is subtracted from the outlier `mathbf{z}_j` for centering. Then, the centered vector is transformed using the principal component loading matrix `mathbf{P}` and its transpose. Finally, the mean `mathbf{bar{Z}}` is added back to obtain the reconstructed value `hat{mathbf{z}_j}`. This process maps the outlier back to the principal component space by preserving the main information, thus obtaining an estimate corresponding to the normal state.

[0135] Calculate the residual vector for each outlier:

[0136] mathbf{e}_j=mathbf{z}_j-hat{mathbf{z}_j}

[0137] The residual vector `mathbf{e}_j` represents the difference between an outlier `mathbf{z}_j` and its reconstructed value `hat{mathbf{z}_j}`. These residuals reflect the degree of deviation of the outlier from the normal state; larger residuals indicate a higher degree of outlier anomaly. By calculating the residual vector, we can better understand which components or variables in the outlier contribute to the anomaly.

[0138] Based on the residual vector, calculate the Q statistic for each outlier:

[0139] Q_j=mathbf{e}_j^T·mathbf{e}_j

[0140] Where, mathbf{e}_j^T is the transpose of the residual vector mathbf{e}_j;

[0141] The Q statistic quantifies the distance between an outlier and its reconstructed value, and is a standard measure for anomaly detection. A higher Q statistic indicates that the outlier deviates significantly from the normal value. Here, the Q statistic is used to further identify fault points among the outliers.

[0142] The Q control limits are determined as follows:

[0143]

[0144] Where z_{beta} is the critical value of the beta level of the standard normal distribution, and h is:

[0145]

[0146] in: lambda_q is the qth eigenvalue of the covariance matrix mathbf{Sigma}_{Z}, and m is the total number of variables;

[0147] The Q control limit, Q_{lim}, is calculated based on the eigenvalues ​​lambda of the covariance matrix mathbf{Sigma}_Z. The calculation of the control limit comprehensively considers the first, second, and third moments theta_1, theta_2, and theta_3 of the eigenvalues ​​to accurately determine the degree of anomaly among outliers. The critical value z_beta of the standard normal distribution is used to set the confidence level, thereby separating normal points from outliers.

[0148] For each outlier, the Q statistic Q_j is compared with the Q control limit Q_{lim}. Outliers with Q statistic greater than the Q control limit are identified as fault points.

[0149] By comparing the Q-statistic Q_j of each outlier with the pre-calculated Q-control limit Q_{lim}, it is possible to determine which outliers exhibit significant anomalies. If Q_j is greater than Q_{lim}, the outlier is identified as a fault point, indicating that its outlier state is sufficiently abnormal and may foreshadow equipment malfunction.

[0150] For a fault point, analyze each component of the residual vector mathbf{e}_j of the fault point. If there is a component whose Q statistic is greater than the Q control limit, then determine the fault variable from the components of the principal component variable matrix corresponding to that component.

[0151] After identifying the fault point, further analysis can be performed on the components of the residual vector `mathbf{e}_j` at that point. This step, by meticulously examining the Q-statistic of each component, identifies the specific components of the anomaly. If the Q-statistic of a residual component exceeds the Q-control limit `Q_{lim}`, then the component of the principal component variable matrix corresponding to that residual component is identified as the fault variable. This step helps to locate and understand which specific variables caused the fault, thus providing more accurate information for fault diagnosis and repair.

[0152] The above steps, explained in detail, provide a comprehensive understanding of each step and its specific role in the fault diagnosis process. This method ensures accurate detection of faults in high-voltage current transformers and effectively distinguishes between normal and abnormal states.

[0153] Specifically, the detected fault variables can be compared with the characteristics of each fault mode in the fault mode library to determine the best matching fault mode.

[0154] To effectively identify fault types, these fault variables can be compared with known fault modes pre-stored in a fault mode library. Each mode in the fault mode library is built using historical data or existing experience and knowledge, and contains features of different fault types. Each fault mode is represented by a feature vector.

[0155] The similarity between the detected fault variables and each feature vector in the fault mode library can be calculated using Euclidean distance. By comparing the Euclidean distances of different fault modes, the mode with the smallest distance can be selected as the final fault discrimination result.

[0156] By following the steps described above, the detected combinations of fault variables can be scientifically compared with a fault mode library to ultimately determine the best-matching fault mode, thus providing accurate fault diagnosis results. This method ensures the reliability and effectiveness of fault diagnosis.

[0157] It is evident that acquiring high-voltage current transformer operating data; using principal component analysis (PCA) to reduce the dimensionality of the operating data and obtain new principal component variables; monitoring these principal component variables using a T2 control chart model to detect outliers; and if the T2 control chart model detects outliers, using a Q control chart model to determine the fault points within the outliers and provide fault identification results—this approach enables rapid and accurate fault detection and provides detailed fault identification results even with complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of the power system.

[0158] Another embodiment of the present invention provides a fault detection system for high-voltage current transformers, see [link to relevant documentation]. Figure 3 The system may include:

[0159] The acquisition module 301 is used to acquire the operating data of the high-voltage current transformer;

[0160] Dimensionality reduction module 302 is used to reduce the dimensionality of the running data using the principal component analysis model to obtain new principal component variables;

[0161] The detection module 303 is used to monitor the principal component variables using the T2 control chart model and detect the existence of outliers.

[0162] The determination module 304 is used to determine the fault point among the outliers when the T2 control chart model detects outliers, and to provide the fault identification result.

[0163] It is evident that acquiring high-voltage current transformer operating data; using principal component analysis (PCA) to reduce the dimensionality of the operating data and obtain new principal component variables; monitoring these principal component variables using a T2 control chart model to detect outliers; and if the T2 control chart model detects outliers, using a Q control chart model to determine the fault points within the outliers and provide fault identification results—this approach enables rapid and accurate fault detection and provides detailed fault identification results even with complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of the power system.

[0164] This invention also provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0165] Specifically, in this embodiment, the storage medium can be configured to store a computer program for performing the following steps:

[0166] S201, Obtain high-voltage current transformer operating data;

[0167] S202, using the principal component analysis model to reduce the dimensionality of the running data to obtain new principal component variables;

[0168] S203 uses the T2 control chart model to monitor principal component variables and detect the existence of outliers;

[0169] S204. If the T2 control chart model detects outliers, the Q control chart model is used to determine the fault points among the outliers and to provide fault identification results.

[0170] It is evident that acquiring high-voltage current transformer operating data; using principal component analysis (PCA) to reduce the dimensionality of the operating data and obtain new principal component variables; monitoring these principal component variables using a T2 control chart model to detect outliers; and if the T2 control chart model detects outliers, using a Q control chart model to determine the fault points within the outliers and provide fault identification results—this approach enables rapid and accurate fault detection and provides detailed fault identification results even with complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of the power system.

[0171] This invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0172] Specifically, the aforementioned electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the aforementioned processor, and the input / output device is connected to the aforementioned processor.

[0173] Specifically, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0174] S201, Obtain high-voltage current transformer operating data;

[0175] S202, using the principal component analysis model to reduce the dimensionality of the running data to obtain new principal component variables;

[0176] S203 uses the T2 control chart model to monitor principal component variables and detect the existence of outliers;

[0177] S204. If the T2 control chart model detects outliers, the Q control chart model is used to determine the fault points among the outliers and to provide fault identification results.

[0178] It is evident that acquiring high-voltage current transformer operating data; using principal component analysis (PCA) to reduce the dimensionality of the operating data and obtain new principal component variables; monitoring these principal component variables using a T2 control chart model to detect outliers; and if the T2 control chart model detects outliers, using a Q control chart model to determine the fault points within the outliers and provide fault identification results—this approach enables rapid and accurate fault detection and provides detailed fault identification results even with complex data and diverse fault modes, which is of great significance for ensuring the safe and stable operation of the power system.

[0179] The above description, based on the embodiments shown in the figures, details the structure, features, and effects of the present invention. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for identifying operational faults in a high-voltage current transformer, characterized in that, The method includes: Obtain operating data of high-voltage current transformers; Principal component analysis model is used to reduce the dimensionality of the running data to obtain new principal component variables; The T2 control chart model is used to monitor the principal component variables and detect outliers. A new principal component variable matrix, `mathbf{Z}`, is obtained, with dimensions n*p, where n is the number of samples and p is the number of principal components. The principal component variable matrix and each sample point are then centered to obtain the centered principal component variable matrix and the centered vector of the sample points. Where, `mathbf{Z}_{centered}` is the centered principal component variable matrix, `mathbf{z}_i_{centered}` is the centered vector of the i-th sample point, and `mathbf{bar{Z}}` is the mean vector of each column in `mathbf{Z}`, where one column corresponds to one principal component and one sample point corresponds to one row; calculate the covariance matrix of the centered principal component variable matrix: Where, mathbf{Z}_{centered}^T represents the transpose of mathbf{Z}_{centered}; using the vector and covariance matrix after centering the sample points, the T2 statistic is calculated: Where T2_i is the T2 statistic of the i-th sample point, mathbf{Sigma}_{Z}^{-1} is the inverse matrix of mathbf{Sigma}_{Z}, and mathbf{z}_i_{centered}^T is the transpose of mathbf{z}_i_{centered}; the T2 control limits are calculated as follows: in, T2 is the control limit, and chi^2 represents the chi-square distribution. This indicates that when the degrees of freedom are p, the cumulative probability is 1- The critical value of the chi^2 distribution, The confidence level; for each sample point, the T2 statistic T2_i, and the control limit By comparison, sample points whose T2 statistic is greater than the T2 control limit are identified as outliers; If the T2 control chart model detects outliers, the Q control chart model is used to determine the fault points among the outliers and provide fault identification results.

2. The method according to claim 1, characterized in that, The method of using the Q-control chart model to determine the fault points among outliers includes: For each outlier point mathbf{z}_j, calculate the reconstructed value of that outlier point, which is then reconstructed using the retained principal components: Where, mathbf{P} is the load matrix with principal components, and mathbf{P}^T is the transpose of the load matrix; Calculate the residual vector for each outlier: Based on the residual vector, calculate the Q statistic for each outlier: Where, mathbf{e}_j^T is the transpose of the residual vector mathbf{e}_j; The Q control limits are determined as follows: Where z_{beta} is the critical beta level of the standard normal distribution, and h is: in: k=1,2,3; Let q be the eigenvalue of the covariance matrix mathbf{Sigma}_{Z}, and m be the total number of variables; For each outlier, the Q-statistic Q_j and the Q-control limit By comparing the results, outliers whose Q-statistics are greater than the Q control limit are identified as fault points. For a fault point, analyze each component of the residual vector mathbf{e}_j of the fault point. If there is a component whose Q statistic is greater than the Q control limit, then determine the fault variable from the component of the principal component variable matrix corresponding to that component.

3. The method according to claim 2, characterized in that, The fault diagnosis results provided include: The detected fault variables are compared with the features of each fault mode in the fault mode library to determine the best matching fault mode.

4. A fault detection system for high-voltage current transformers, characterized in that, The system includes: The acquisition module is used to acquire operating data of the high-voltage current transformer; The dimensionality reduction module is used to reduce the dimensionality of the running data using the principal component analysis model to obtain new principal component variables; The detection module monitors the principal component variables using a T2 control chart model to detect outliers. Specifically, it obtains a new principal component variable matrix `mathbf{Z}`, where the matrix has dimensions `n*p`, where `n` is the number of samples and `p` is the number of principal components. The module then centers both the principal component variable matrix and each sample point, resulting in a centered principal component variable matrix and a centered vector of sample points. Where, `mathbf{Z}_{centered}` is the centered principal component variable matrix, `mathbf{z}_i_{centered}` is the centered vector of the i-th sample point, and `mathbf{bar{Z}}` is the mean vector of each column in `mathbf{Z}`, where one column corresponds to one principal component and one sample point corresponds to one row; calculate the covariance matrix of the centered principal component variable matrix: Where, mathbf{Z}_{centered}^T represents the transpose of mathbf{Z}_{centered}; using the vector and covariance matrix after centering the sample points, the T2 statistic is calculated: Where T2_i is the T2 statistic of the i-th sample point, mathbf{Sigma}_{Z}^{-1} is the inverse matrix of mathbf{Sigma}_{Z}, and mathbf{z}_i_{centered}^T is the transpose of mathbf{z}_i_{centered}; the T2 control limits are calculated as follows: in, T2 is the control limit, and chi^2 represents the chi-square distribution. This indicates that when the degrees of freedom are p, the cumulative probability is 1- The critical value of the chi^2 distribution, The confidence level; for each sample point, the T2 statistic T2_i, and the control limit By comparison, sample points whose T2 statistic is greater than the T2 control limit are identified as outliers; The determination module is used to identify fault points among outliers when outliers are detected by the T2 control chart model, and to provide fault identification results.

5. The system according to claim 4, characterized in that, The determining module is specifically used for: For each outlier point mathbf{z}_j, calculate the reconstructed value of that outlier point, which is then reconstructed using the retained principal components: Where, mathbf{P} is the load matrix with principal components, and mathbf{P}^T is the transpose of the load matrix; Calculate the residual vector for each outlier: Based on the residual vector, calculate the Q statistic for each outlier: Where, mathbf{e}_j^T is the transpose of the residual vector mathbf{e}_j; The Q control limits are determined as follows: Where z_{beta} is the critical beta level of the standard normal distribution, and h is: in: k=1,2,3; Let q be the eigenvalue of the covariance matrix mathbf{Sigma}_{Z}, and m be the total number of variables; For each outlier, the Q-statistic Q_j and the Q-control limit By comparing the results, outliers whose Q-statistics are greater than the Q control limit are identified as fault points. For a fault point, analyze each component of the residual vector mathbf{e}_j of the fault point. If there is a component whose Q statistic is greater than the Q control limit, then determine the fault variable from the component of the principal component variable matrix corresponding to that component.

6. The system according to claim 5, characterized in that, The determining module is specifically used for: The detected fault variables are compared with the features of each fault mode in the fault mode library to determine the best matching fault mode.

7. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method of any one of claims 1-3 when it is run.

8. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-3.

Images