Intelligent diagnosis method for electromagnetic mutual inductor

Abstract

The invention relates to the technical field of power equipment fault diagnosis, in particular to an intelligent diagnosis method for an electromagnetic mutual inductor, which comprises the following steps: acquiring operation signals of the electromagnetic mutual inductor, the operation signals comprising a vibration signal, a temperature rise signal and a partial discharge signal; inputting the diagnosis features into a preset fault diagnosis model, obtaining a diagnosis result, generating a corresponding early warning signal, and carrying out early warning; and determining a maintenance strategy of the electromagnetic transformer based on the early warning signal and the historical data of the electromagnetic transformer. According to the method, multi-source operation signals such as vibration, temperature rise and partial discharge are collected, diagnosis features are extracted, and a fault diagnosis model based on a support vector machine is introduced, so that high-precision intelligent diagnosis of the degradation state and the fault of the electromagnetic transformer is realized. An early warning signal can be generated, potential hidden dangers can be found in advance, sudden failures and power failure risks are reduced, a maintenance strategy is optimized in combination with historical data, and transformation from post repair to state maintenance and predictive maintenance is achieved.

Description

Technical Field

[0001] This invention relates to the field of power equipment fault diagnosis technology, and in particular to an intelligent diagnostic method for electromagnetic transformers. Background Technology

[0002] Electromagnetic instrument transformers, as key measurement and protection devices in power systems, are widely used in the acquisition and transmission of current and voltage. Their operating status directly affects the safety and stability of the power system. With the continuous expansion of the power grid and the sustained increase in load levels, instrument transformers operate under high voltage, high current, and complex electromagnetic environments for extended periods, making them prone to degradation and faults such as inter-turn short circuits, insulation aging, local core saturation, and enhanced partial discharge. Once an instrument transformer fails, it may lead to measurement errors, abnormal operation of power protection systems, and even equipment damage and large-scale power outages. Therefore, real-time monitoring of its operating status and early fault diagnosis are of great significance.

[0003] In existing technologies, electromagnetic instrument transformer condition monitoring mainly relies on regular inspections by maintenance personnel or simple threshold judgments of single signals, making it difficult to achieve comprehensive analysis of multi-source operating signals such as vibration, temperature rise, and partial discharge. Furthermore, traditional diagnostic methods often fail to accurately characterize the mechanism of instrument transformer material degradation, and have limited ability to identify early-stage problems such as minor inter-turn short circuits and localized insulation degradation. In addition, existing technologies generally lack proactive early warning mechanisms and equipment maintenance strategy formulation methods based on diagnostic results, leading to poor predictability of fault occurrence and low efficiency in equipment maintenance and replacement. Summary of the Invention

[0004] (a) Purpose of the invention The purpose of this invention is to provide an intelligent diagnostic method for electromagnetic instrument transformers. By collecting multi-source operating signals such as vibration, temperature rise, and partial discharge, and extracting diagnostic features, a fault diagnosis model based on support vector machines is introduced to achieve high-precision intelligent diagnosis of the degradation state and faults of electromagnetic instrument transformers. It can generate early warning signals, detect potential hazards in advance, reduce the risk of sudden failures and power outages, and optimize maintenance strategies by combining historical data, realizing a shift from reactive repair to condition-based maintenance and predictive maintenance.

[0005] (II) Technical Solution To address the above problems, this invention provides an intelligent diagnostic method for electromagnetic current transformers, comprising: The operating signals of the electromagnetic current transformer are collected, including vibration signals, temperature rise signals, and partial discharge signals. Extract the features of the operating signals to obtain diagnostic features; The diagnostic features are input into a preset fault diagnosis model to obtain diagnostic results. The preset fault diagnosis model is based on the support vector machine algorithm. Based on the diagnostic results, a corresponding early warning signal is generated and an early warning is issued; Based on the warning signal and the historical data of the electromagnetic transformer, a maintenance strategy for the electromagnetic transformer is determined.

[0006] In another aspect of the present invention, preferably, the diagnostic features include vibration signal features, temperature rise signal features, and partial discharge signal features; The vibration signal characteristics include: power spectral density, dominant frequency, harmonic frequency, root mean square value, and kurtosis; The temperature rise signal features include: temperature change rate, temperature gradient, maximum temperature rise, and average temperature rise; The characteristics of the partial discharge signal include: discharge quantity, discharge frequency, discharge phase, and partial discharge pulse width.

[0007] In another aspect of the present invention, preferably, the fault diagnosis model includes a kernel function, which is a linear kernel function, a polynomial kernel function, or a radial basis function, and the kernel function is used to map the diagnostic features to a high-dimensional space.

[0008] In another aspect of the present invention, preferably, the kernel function of the fault diagnosis model is expressed using the following formula: in, This represents the output value of the kernel function, i.e., the similarity between sample x and sample x′ in high-dimensional space. σ represents the square of the Euclidean distance between two samples, and σ represents the width parameter of the kernel function.

[0009] In another aspect of the present invention, preferably, the preset fault diagnosis model is represented by the following formula: Where f(x) represents the diagnostic result, α i Denotes Lagrange multipliers, y i The label representing sample number i. denoted by , b represents the kernel function, and b represents the bias term.

[0010] In another aspect of the present invention, preferably, the fault diagnosis model is trained, the training comprising: Initialize the penalty parameter and the width parameter of the kernel function, wherein the penalty parameter represents the degree of penalty for misclassification of the fault diagnosis model; Construct an optimization objective function; The objective function is solved using a preset optimization algorithm to obtain the optimal value; Based on the optimal value, the range of supported samples is determined; Calculate the bias term based on any one of the support samples within the range of support samples; Based on the optimal value and bias term, the trained fault diagnosis model is obtained.

[0011] In another aspect of the present invention, preferably, the optimization objective function is expressed using the following formula: Where U represents the objective function; This means optimizing over all possible α vectors to find an optimal α vector that minimizes the objective function; the α vector contains the Lagrange multipliers corresponding to all samples. i and α j y represents the components of vector α, and y represents the Lagrange multiplier corresponding to each component; i The label y represents the sample numbered i. j K(x) represents the label of sample j; i ,x j ) represents the kernel function.

[0012] In another aspect of the invention, preferably, the bias term is calculated using the following formula: Where b represents the bias term, α i Represents the Lagrange multiplier, y i The label representing sample number i. Represents the kernel function. Indicates supporting sample, y s Labels that indicate support for the sample.

[0013] In another aspect of the present invention, preferably, the step of generating a corresponding early warning signal based on the diagnostic results and issuing an early warning includes: When the diagnostic result is a typical fault, a corresponding early warning signal is generated. The typical faults include inter-turn short circuit, insulation degradation and local core saturation. The early warning signal includes the fault type, suspected fault location, fault development trend and fault severity. The warning signal will be sent to the maintenance personnel in real time.

[0014] In another aspect of the present invention, preferably, determining the maintenance strategy for the electromagnetic transformer based on the early warning signal and the historical data of the electromagnetic transformer includes: Based on the warning signal and the historical data of the electromagnetic transformer, the health status and remaining lifespan of the electromagnetic transformer are assessed, and the assessment results are obtained. Based on the evaluation results, a maintenance strategy for the electromagnetic transformer is determined.

[0015] (III) Beneficial Effects The above-described technical solution of the present invention has the following beneficial technical effects: This invention simultaneously collects multiple types of signals, including vibration, temperature rise, and partial discharge, comprehensively reflecting the operational health of instrument transformers from multiple dimensions such as structure, thermal characteristics, and insulation status. Compared with traditional single-signal monitoring methods, it significantly improves the ability to identify early degradation characteristics. By inputting diagnostic features into a fault diagnosis model built based on a support vector machine algorithm, high-precision classification of typical fault modes such as inter-turn short circuits and insulation aging can be achieved, improving the accuracy and stability of diagnosis. Early warning signals generated based on the diagnostic results can provide advance warning of potential equipment risks, enabling maintenance personnel to take timely measures and effectively reduce the probability of fault escalation. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall process of one embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0018] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0019] In the description of this invention, it should be noted that the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0020] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0021] Example 1 A smart diagnostic method for electromagnetic current transformers. Figure 1 A schematic diagram of the overall process of an embodiment of the present invention is shown, as follows: Figure 1 As shown, it includes: The electromagnetic instrument transformer's operating signals are collected, including vibration signals, temperature rise signals, and partial discharge signals. Vibration signals are used to reflect problems such as loose core and abnormal mechanical structure. High-precision accelerometers are installed at key parts of the electromagnetic instrument transformer to collect vibration signals. The sampling frequency of the accelerometer should be determined based on the operating frequency and vibration characteristics of the electromagnetic instrument transformer, and set to at least 1000 times per second to ensure the capture of high-frequency vibration signals. Temperature rise signals are used to indicate potential problems such as winding overheating and poor contact. Thermocouples or infrared thermal imagers are installed at key parts of the electromagnetic instrument transformer, such as the windings and core, to collect temperature rise signals. Thermocouples should have a measurement accuracy of at least 0.1℃, and infrared thermal imagers should have a resolution of at least 0.1℃ to accurately reflect the temperature distribution and changes of the transformer. Partial discharge signals are used to characterize discharge risks such as insulation degradation and insulation defects. High-frequency current sensors or ultrasonic sensors are installed at key parts of the electromagnetic instrument transformer, such as the windings and insulation materials, to collect partial discharge signals. The high-frequency current sensor has a bandwidth of 1MHz-100MHz, and the ultrasonic sensor has a bandwidth of 20kHz-200kHz, to accurately capture high-frequency current pulses and ultrasonic signals generated by partial discharge. Data storage and preprocessing are performed through field monitoring devices or online monitoring systems.

[0022] Preprocessing includes data cleaning, which cleans the acquired signal data to remove outliers and noise. Outliers can be detected and removed using statistical analysis methods. The mean and standard deviation of the data are calculated, and data points exceeding the mean plus or minus three times the standard deviation are considered outliers. Noise data is processed using filtering algorithms. A low-pass filter is used to remove high-frequency noise, and a median filter is used to remove impulse noise.

[0023] Preprocessing also includes data normalization: converting data of different dimensions into a uniform range using a 0-1 normalization method. For each feature data x, its normalized value x′ is calculated: Here, min(x) and max(x) are the minimum and maximum values ​​of the feature data, respectively.

[0024] The operating signals are extracted to obtain diagnostic features. These diagnostic features include vibration signal features, temperature rise signal features, and partial discharge signal features. The vibration signal features include: power spectral density, obtained by performing a Fast Fourier Transform (FFT) on the vibration signal to obtain the frequency domain energy distribution. Power spectral density reflects the distribution characteristics of vibration energy at different frequencies, helping to identify characteristic frequencies such as core loosening and component collisions; dominant frequency, which is the dominant frequency where energy is most concentrated in the vibration signal. If the dominant frequency drifts or abnormal high-frequency components appear, it usually indicates an operational abnormality in the internal mechanical structure; harmonic frequencies, including integer multiples of the dominant frequency, are used to determine whether the transformer has problems such as periodic mechanical vibration or electromagnetic excitation frequency coupling; root mean square value, reflecting the overall intensity of the vibration signal, is an important parameter for judging the trend of equipment vibration energy changes; kurtosis, used to characterize the sharpness of the signal peak, can sensitively capture sudden vibration events such as short-term impacts and loosening collisions. These vibration features can comprehensively reflect the mechanical stability and structural health of the transformer. The temperature rise signal features include: temperature change rate, i.e., the rate at which temperature rises over time. If the rate of temperature change suddenly increases, there may be hidden dangers such as increased contact resistance and local overheating; temperature gradient, the temperature difference between different measuring points, can be used to determine whether the heat distribution is uniform. If the temperature of a certain part is significantly higher than that of other parts, there may be a local overheating fault; maximum temperature rise, reflecting the peak temperature reached by the equipment in a certain operating cycle, is used to determine whether it exceeds the design limit and average temperature rise, the average temperature level within a certain time window, can be used to determine the thermal stability of long-term operation. The characteristics of the partial discharge signal include: discharge quantity, referring to the amount of charge released in each discharge, which is the core feature for judging the severity of insulation defects; discharge frequency, the number of discharge pulses occurring per unit time, the higher the frequency, the more severe the insulation damage; discharge phase, based on the discharge distribution characteristics of the power frequency phase, can be used to identify different types of insulation defects, such as surface discharge, floating potential discharge, internal air gap discharge, etc.; and partial discharge pulse width, used to characterize the duration of the discharge pulse, which is closely related to the type and scale of the insulation defect. By extracting multi-dimensional features from vibration, temperature rise, and partial discharge signals, a complete diagnostic feature set was constructed, enabling the quantitative expression of the transformer's operating state and providing accurate, sufficient, and interpretable feature inputs for subsequent fault diagnosis models based on support vector machines.

[0025] The diagnostic features are input into a preset fault diagnosis model to obtain diagnostic results. The preset fault diagnosis model is based on the Support Vector Machine (SVM) algorithm. In this embodiment, the SVM algorithm is used to construct the diagnostic model, leveraging its good classification performance under small sample conditions. The model is trained using historical labeled data and incorporates linear kernel functions, polynomial kernel functions, or radial basis function kernel functions. The kernel function is used to map the diagnostic features to a high-dimensional space to accurately distinguish between multiple types of transformer fault modes. The diagnostic results output by the model include various state categories such as normal state, slight degradation, insulation abnormality, and structural loosening. Furthermore, in this embodiment, the kernel function of the fault diagnosis model is expressed using the following formula: in, This represents the output value of the kernel function, i.e., the similarity between sample x and sample x′ in high-dimensional space. σ represents the square of the Euclidean distance between two samples, and σ represents the width parameter of the kernel function.

[0026] The preset fault diagnosis model is represented by the following formula: Where f(x) represents the diagnostic result, α i Denotes Lagrange multipliers, y i The label representing sample number i. denoted by , b represents the kernel function, and b represents the bias term.

[0027] Based on the diagnostic results, corresponding early warning signals are generated and issued. Early warning signals are generated according to different fault categories. These signals can be divided into three categories: general early warning, key early warning, and emergency early warning, corresponding to minor equipment anomalies, increased potential risks, and increased likelihood of major faults, respectively. Early warning signals can be pushed to maintenance personnel through the monitoring platform or linked with the substation automation system to achieve status visualization and real-time alarms. Further, in this embodiment, generating corresponding early warning signals based on the diagnostic results includes: When the diagnostic result is a typical fault, a corresponding early warning signal is generated. The typical faults include inter-turn short circuit, insulation degradation and local core saturation. The early warning signal includes the fault type, suspected fault location, fault development trend and fault severity. The warning signal will be sent to maintenance personnel in real time. The warning signal will be sent to the power system's maintenance personnel via SMS, email, or system alarms to promptly notify relevant personnel so that measures can be taken in advance to prevent the fault from escalating further.

[0028] Based on the warning signal and historical data of the electromagnetic instrument transformer, a maintenance strategy for the electromagnetic instrument transformer is determined. This includes determining a maintenance strategy for the current equipment status based on the warning signal and historical operating data of the electromagnetic instrument transformer. Historical data includes long-term operating temperature, vibration trends, insulation discharge trends, and past maintenance records. The system generates a maintenance plan, including maintenance time suggestions, key component inspections, insulation treatment measures, and structural reinforcement measures, through rule-based analysis or data-driven strategy analysis. If multiple indicators of the equipment are abnormal, a comprehensive health index can be further determined using a weighted model, and graded maintenance recommendations can be given based on the index's changing trend. In this embodiment, determining the maintenance strategy for the electromagnetic instrument transformer based on the warning signal and historical data includes: Based on the warning signal and historical data of the electromagnetic transformer, the health status and remaining lifespan of the electromagnetic transformer are assessed to obtain the assessment results. The warning signal and the historical data of the electromagnetic transformer are input into the equipment lifespan assessment model to assess the remaining lifespan of the equipment. The warning signal reflects the type and severity of the anomalies currently detected by the equipment, such as minor vibration anomalies, excessive temperature rise, and increased partial discharge, and is an important basis for the real-time status of the equipment. Historical data includes multiple data records accumulated since the equipment started operating, such as long-term temperature curves, vibration trends, partial discharge evolution trajectories, overload operation records, changes in environmental conditions, and past fault and maintenance history, which together constitute the health record of the equipment throughout its entire life cycle. By inputting the warning signal and historical data into the preset equipment lifespan assessment model, the current health index of the equipment can be calculated, and its remaining lifespan can be predicted. The lifespan assessment model can be a prediction model based on machine learning, such as a support vector regression model (SVR) or a long short-term memory network (LSTM) model based on degradation trend fitting, or an insulation aging model and a thermal aging model based on physical mechanisms.

[0029] Based on the assessment results, a maintenance strategy for the electromagnetic instrument transformer is determined. The maintenance strategy is automatically generated and categorized according to the severity of the health status and the remaining lifespan, including the following: Routine maintenance strategy: When the equipment is in good health and has a long remaining lifespan, it is recommended to continue routine inspections according to the regular inspection cycle and continuously record key monitoring indicators. Preventive maintenance strategy: When the assessment results show that the equipment has minor abnormalities or the remaining lifespan is declining, preventive maintenance is recommended, such as checking wiring contacts, tightening mechanical parts, and cleaning insulation surfaces to slow down the degradation process. Key monitoring and short-term maintenance strategy: When the equipment has moderate abnormalities or a significantly shortened remaining lifespan, the system will recommend shortening the inspection cycle, increasing monitoring frequency, and arranging targeted repairs in advance, such as strengthening insulation on abnormal hot spots and correcting loose parts. Emergency maintenance or replacement strategy: When the equipment is in a severe abnormality or the remaining lifespan is close to its limit, the system will trigger an emergency maintenance recommendation, prompting maintenance personnel to immediately take measures such as shutdown for repair, component replacement, or replacement with backup equipment to prevent equipment failure from damaging the system.

[0030] Furthermore, in this embodiment, the fault diagnosis model is trained, and the training includes: Initialize the penalty parameter and the width parameter of the kernel function. The penalty parameter represents the degree of penalty for misclassification by the fault diagnosis model. Select the optimal penalty parameter C and kernel function parameter σ using methods such as grid search and cross-validation. Set the value range of C to 0.1, 1, 10, 100, and the value range of σ to 0.1, 1, 10, 100. Select the optimal parameter combination through cross-validation to improve the accuracy and generalization ability of the model.

[0031] Construct an optimization objective function; the optimization objective function is expressed by the following formula: Where U represents the objective function; This means optimizing over all possible α vectors to find an optimal α vector that minimizes the objective function; the α vector contains the Lagrange multipliers corresponding to all samples. i and α j y represents the components of vector α, and y represents the Lagrange multiplier corresponding to each component; i The label y represents the sample numbered i. j K(x) represents the label of sample j; i ,x j The kernel function is represented by . In the optimization problem of Support Vector Machines (SVM), the objective is to minimize the objective function.

[0032] The objective function is solved using a preset optimization algorithm to obtain the optimal value; the Sequence Minimum Optimization (SMO) algorithm or other optimization algorithms are used to solve the above optimization problem to obtain the optimal value, which is the optimal Lagrange multiplier.

[0033] Based on the optimal value, the range of supporting samples is determined, i.e., satisfying α. i Samples >0; The bias term is calculated based on any one of the support samples within the support sample range; the bias term is calculated using the following formula: Where b represents the bias term, α i Represents the Lagrange multiplier, y i The label representing sample number i. Represents the kernel function. Indicates supporting sample, y s Labels that indicate support for the sample.

[0034] Based on the optimal value and bias term, the trained fault diagnosis model is obtained.

[0035] Furthermore, the trained fault diagnosis model is validated using test set data to evaluate metrics such as accuracy, recall, and F1 score. The specific steps are as follows: Performance Evaluation: The model is used to make predictions using test set data, and metrics such as accuracy, recall, and F1 score are calculated. Accuracy represents the proportion of correctly classified samples out of the total number of samples; recall represents the proportion of correctly classified positive samples out of the actual number of positive samples; and the F1 score is the harmonic mean of accuracy and recall, comprehensively reflecting the model's classification performance.

[0036] Results Analysis: Based on the model's performance evaluation results, analyze the model's strengths and weaknesses. If the model's performance does not meet the requirements, it is necessary to readjust the model's parameters or select a different kernel function, and retrain and validate until the model's performance reaches the expected target.

[0037] This invention simultaneously collects multiple types of signals, including vibration, temperature rise, and partial discharge, comprehensively reflecting the operational health of instrument transformers from multiple dimensions such as structure, thermal characteristics, and insulation status. Compared with traditional single-signal monitoring methods, it significantly improves the ability to identify early degradation characteristics. By inputting diagnostic features into a fault diagnosis model built based on a support vector machine algorithm, high-precision classification of typical fault modes such as inter-turn short circuits and insulation aging can be achieved, improving the accuracy and stability of diagnosis. Early warning signals generated based on the diagnostic results can provide advance warning of potential equipment risks, enabling maintenance personnel to take timely measures and effectively reduce the probability of fault escalation.

[0038] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

[0039] The present invention has been described above with reference to embodiments thereof. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. The scope of the invention is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.

[0040] Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions, and modifications can be made to the embodiments of the present invention without departing from the spirit and scope of the invention.

[0041] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A smart diagnostic method for electromagnetic current transformers, characterized in that, include: The operating signals of the electromagnetic current transformer are collected, including vibration signals, temperature rise signals, and partial discharge signals. Extract the features of the operating signals to obtain diagnostic features; The diagnostic features are input into a preset fault diagnosis model to obtain diagnostic results. The preset fault diagnosis model is based on the support vector machine algorithm. Based on the diagnostic results, a corresponding early warning signal is generated and an early warning is issued; Based on the warning signal and the historical data of the electromagnetic transformer, a maintenance strategy for the electromagnetic transformer is determined.

2. The intelligent diagnostic method for electromagnetic transformers according to claim 1, characterized in that, The diagnostic features include vibration signal features, temperature rise signal features, and partial discharge signal features; The vibration signal characteristics include: power spectral density, dominant frequency, harmonic frequency, root mean square value, and kurtosis; The temperature rise signal features include: temperature change rate, temperature gradient, maximum temperature rise, and average temperature rise; The characteristics of the partial discharge signal include: discharge quantity, discharge frequency, discharge phase, and partial discharge pulse width.

3. The intelligent diagnostic method for electromagnetic transformers according to claim 1, characterized in that, The fault diagnosis model includes a kernel function, which can be a linear kernel function, a polynomial kernel function, or a radial basis function. The kernel function is used to map the diagnostic features to a high-dimensional space.

4. The intelligent diagnostic method for electromagnetic current transformers according to claim 3, characterized in that, The kernel function of the fault diagnosis model is expressed by the following formula: in, This represents the output value of the kernel function, i.e., the similarity between sample x and sample x′ in high-dimensional space. σ represents the square of the Euclidean distance between two samples, and σ represents the width parameter of the kernel function.

5. The intelligent diagnostic method for electromagnetic current transformers according to claim 1, characterized in that, The preset fault diagnosis model is represented by the following formula: Where f(x) represents the diagnostic result, α i Denotes Lagrange multipliers, y i The label representing sample number i. denoted by , b represents the kernel function, and b represents the bias term.

6. The intelligent diagnostic method for electromagnetic current transformers according to claim 5, characterized in that, The fault diagnosis model is trained, and the training includes: Initialize the penalty parameter and the width parameter of the kernel function, wherein the penalty parameter represents the degree of penalty for misclassification of the fault diagnosis model; Construct an optimization objective function; The objective function is solved using a preset optimization algorithm to obtain the optimal value; Based on the optimal value, the range of supported samples is determined; Calculate the bias term based on any one of the support samples within the range of support samples; Based on the optimal value and bias term, the trained fault diagnosis model is obtained.

7. The intelligent diagnostic method for electromagnetic transformers according to claim 6, characterized in that, The optimization objective function is expressed by the following formula: Where U represents the objective function; This means optimizing over all possible α vectors to find an optimal α vector that minimizes the objective function; the α vector contains the Lagrange multipliers corresponding to all samples. i and α j y represents the components of vector α, and y represents the Lagrange multiplier corresponding to each component; i The label y represents the sample numbered i. j K(x) represents the label of sample j; i ,x j ) represents the kernel function.

8. The intelligent diagnostic method for electromagnetic current transformers according to claim 6, characterized in that, The bias term is calculated using the following formula: Where b represents the bias term, α i Represents the Lagrange multiplier, y i The label representing sample number i. Represents the kernel function. Indicates supporting sample, y s Labels that indicate support for the sample.

9. The intelligent diagnostic method for electromagnetic transformers according to claim 1, characterized in that, The step of generating a corresponding early warning signal based on the diagnostic results and issuing an early warning includes: When the diagnostic result is a typical fault, a corresponding early warning signal is generated. The typical faults include inter-turn short circuit, insulation degradation and local core saturation. The early warning signal includes the fault type, suspected fault location, fault development trend and fault severity. The warning signal will be sent to the maintenance personnel in real time.

10. The intelligent diagnostic method for electromagnetic transformers according to claim 1, characterized in that, The process of determining the maintenance strategy for the electromagnetic instrument transformer based on the early warning signal and historical data of the electromagnetic instrument transformer includes: Based on the warning signal and the historical data of the electromagnetic transformer, the health status and remaining lifespan of the electromagnetic transformer are assessed, and the assessment results are obtained. Based on the evaluation results, a maintenance strategy for the electromagnetic transformer is determined.

Images