Transformer partial discharge data monitoring system and method

Abstract

The invention discloses a transformer partial discharge data monitoring system and method, and relates to the technical field of power equipment monitoring, the system comprises a data acquisition unit, a data preprocessing unit, a data storage unit, a data analysis and fault diagnosis unit and a user interface; the data acquisition unit is used for acquiring various data; the data preprocessing unit is used for carrying out data cleaning, standardization and normalization processing on the data; the data storage unit is used for receiving and storing data and constructing a data set according to data content; the data analysis and fault diagnosis unit performs data analysis, fault alarm and diagnosis; and the user interface is used for providing a user interaction interface and displaying the monitoring data, the system state and the diagnosis result. The optical fiber sensor position prediction model is constructed, the optimal installation position of the optical fiber sensor can be accurately predicted, and therefore the accuracy and reliability of monitoring data are improved.

Description

Technical Field

[0001] The invention relates to the technical field of power equipment monitoring, and in particular to a transformer partial discharge data monitoring system and method. Background Art

[0002] The transformer is one of the most important equipment in the power system, and its safe and stable operation is crucial to the reliability of the power system. Transformer partial discharge monitoring is an important maintenance and safety measure in the power system. Partial discharge usually occurs in the insulation system of the transformer, forming tiny arcs in the insulation material of the transformer. These tiny arcs can gradually damage the insulation material, leading to equipment failure, aging and reduced insulation performance, and even causing serious equipment failure. In order to improve the operating reliability of the transformer and extend its service life, real-time monitoring and analysis of partial discharge is crucial.

[0003] At present, the commonly used transformer partial discharge monitoring technologies include capacitor voltage division method, high-frequency current sensing method and ultrasonic detection method. These methods are susceptible to electromagnetic interference and are difficult to monitor weak partial discharge signals. Fiber optic sensing technology can effectively detect weak partial discharge signals, is stable in high electromagnetic environments, and is suitable for use in high-temperature environments of transformers; however, the location where the fiber optic sensor is installed directly affects the accuracy and reliability of the data and the effective detection of transformer partial discharge activities. Summary of the invention

[0004] The object of the present invention is to provide a transformer partial discharge data monitoring system and method to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A transformer partial discharge data monitoring method, the method comprising the following steps:

[0007] S1, collect data, preprocess the data and construct a data set;

[0008] S2. constructing a fiber optic sensor position prediction model using a support vector regression algorithm and training the fiber optic sensor position prediction model;

[0009] S3, adjust the model parameters and optimize the model results through the root mean square error;

[0010] S4, using the trained model to predict the optimal installation position of the optical fiber sensor, and installing the optical fiber sensor according to the prediction result;

[0011] S5, preprocessing the signal collected by the optical fiber sensor;

[0012] S6, analyzing and processing the preprocessed signal, and outputting the result;

[0013] S7. Complete the monitoring of transformer partial discharge data and fault alarm.

[0014] According to the above scheme, in step S1, the data includes the signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the model of the transformer, the temperature of the transformer operating environment and the historical partial discharge data of the transformer;

[0015] The preprocessing includes cleaning the data, processing missing values ​​and outliers, and performing standardization and normalization processing; the processed data is stored and a data set is constructed;

[0016] 70% of the data in the data set is used as a training set, and 30% of the data is used as a test set.

[0017] According to the above scheme, in step S2, the data in the training set is input into the optical fiber sensor position prediction model to solve the optimization problem of support vector regression, and the formula is as follows:

[0018]

[0019] Among them, min w，b represents the minimization of the objective function; ||·|| represents the Euclidean norm; C represents the regularization parameter, which is used to balance the weight between the regularization term and the error term; L ε It is expressed as an insensitive loss function, which is used to measure the predicted value f(x i ) and the true value y i The error between; ε represents the insensitive interval threshold, which is used to define the error tolerance range; n represents the number of samples; x i It is represented as the feature vector of the i-th sample; f(x i ) is expressed as the predicted value; y i Represented as the true value of the i-th sample;

[0020] Insensitive loss function L ε , the formula is as follows:

[0021] L ε (y i ,f(x i ))=max(0,|y i -f(x i )-ε|)

[0022] Among them, L ε Expressed as an insensitive loss function; f(x i ) is expressed as the predicted value; y i Represented as the true value of the i-th sample;

[0023] The radial basis function is used as the kernel function to help the model handle nonlinear relationships. The formula is as follows:

[0024]

[0025] Among them, K(x i ,x j ) is represented as kernel function; x i and x j Represented as two samples; β represents the parameter of the kernel function; using the radial basis function as the kernel function helps the model deal with nonlinear relationships, so that the model can capture the characteristics of partial discharge more accurately;

[0026] The decision function is obtained, and the formula is as follows:

[0027]

[0028] Among them, f(x i ) represents the predicted value; w represents the weight vector; b represents the bias term; K(x i ,x) is represented as a kernel function, which is used to calculate the sample x i The inner product between the new sample x and the original sample x. In high-dimensional space, this inner product represents the similarity between the two samples.

[0029] The support vector regression (SVR) algorithm is used to construct a fiber optic sensor position prediction model, which can accurately predict the optimal installation position of the fiber optic sensor, thereby improving the accuracy and reliability of monitoring data.

[0030] According to the above scheme, in step S3, the true value on the test set and the model prediction value are used to calculate the root mean square error, and the formula is as follows:

[0031]

[0032] Among them, RMSE represents the root mean square error; y a It is represented as the true value of the ath sample in the test set; f(x a ) represents the model's response to the a-th test sample x a The predicted value of; M represents the number of test samples;

[0033] According to the RMS error results, the model parameters are adjusted to optimize the model results.

[0034] According to the above scheme, in step S5, the signal collected by the optical fiber sensor is a high-frequency electrical pulse or optical signal, and the analog signal is converted into a digital signal through an analog-to-digital converter, and filtered, denoised and amplified;

[0035] The filtering process uses a low-pass filter, and the formula is as follows:

[0036]

[0037] Where H(f) is the frequency response function of the filter; f is the frequency; f c It is expressed as the cut-off frequency;

[0038] The denoising process uses a digital filter, and the formula is as follows:

[0039]

[0040] Among them, z k Represented as the kth output sample; x m Represents the output sample; n represents the window size of the filter;

[0041] The amplification process uses a linear amplifier, and the formula is as follows:

[0042] V out =G×V in

[0043] Among them, V out Represented as output signal; V in It represents the input signal; G represents the gain of the linear amplifier;

[0044] The signal processing technology of analog-to-digital conversion, filtering, denoising and amplification is adopted to effectively improve the signal quality, reduce noise interference and enhance the signal availability.

[0045] According to the above scheme, in step S6, a feature vector is formed based on the preprocessed signal as a feature; a support vector machine is used as a classifier and trained using the feature vector; the support vector machine model learns how to distinguish the partial discharge of the transformer based on the input feature vector through training; using the support vector machine as a classifier, different partial discharge situations of the transformer can be effectively distinguished; the support vector machine performs well on a small sample set and can improve the robustness of the classification by maximizing the decision boundary;

[0046] The root mean square error is used to adjust the model parameters and optimize the model results. The prediction accuracy and precision of the model are verified through evaluation and verification. Cross-validation is used to verify the generalization ability and stability of the model to ensure that the model can complete the monitoring of transformer partial discharge data on different data sets.

[0047] According to the above scheme, in step S7, a threshold is set according to the collected data, and the formula is as follows:

[0048] T=μ+bα

[0049] Wherein, T represents the threshold value; μ represents the average value of the historical partial discharge data of the transformer; b represents the standard deviation of the historical partial discharge data of the transformer; α represents the empirical value;

[0050] When the partial discharge data of the monitored transformer exceeds the threshold, an early warning is triggered.

[0051] A transformer partial discharge data monitoring system, the system comprising: a data acquisition unit, a data preprocessing unit, a data storage unit, a data analysis and fault diagnosis unit and a user interface;

[0052] The data acquisition unit collects the signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the model of the transformer, the temperature of the transformer operating environment and the historical partial discharge data of the transformer; and transmits the data to the data preprocessing unit;

[0053] The data preprocessing unit is used to receive the data transmitted by the data acquisition unit and preprocess the data, wherein the preprocessing includes data cleaning, processing missing data values ​​and abnormal values, and performing standardization and normalization processing; converting the high-frequency electrical pulse or optical signal collected by the optical fiber sensor from an analog signal to a digital signal, and performing filtering processing, denoising processing and amplification processing;

[0054] The data storage unit is used to receive and store the data transmitted by the data preprocessing unit and the data analysis and fault diagnosis unit, and to construct a data set according to the data content;

[0055] The data analysis and fault diagnosis unit is used to train the optical fiber sensor position prediction model and the transformer partial discharge data monitoring model using the data set; predict the position of the optical fiber sensor and the transformer partial discharge data according to the trained model; and analyze and process the real-time data to identify the transformer partial discharge situation, and perform fault alarm and diagnosis; transmit the prediction results of the model to the user interface; combined with the real-time monitoring and fault alarm mechanism, it can respond to abnormal situations in a timely manner, quickly locate faults and give early warnings, thereby improving the operational safety of the transformer;

[0056] The user interface is used to provide a user interaction interface to display monitoring data, system status and diagnosis results; when the fault alarm information sent by the data analysis and fault diagnosis unit is received, the fault alarm is triggered to notify the maintenance personnel to take corresponding maintenance measures.

[0057] According to the above scheme, the optical fiber sensor position prediction model adopts support vector regression; the transformer partial discharge data monitoring model adopts support vector machine; when the optical fiber sensor position prediction model predicts the position of the optical fiber sensor and installs the optical fiber sensor according to the prediction result, the data analysis and fault diagnosis unit then trains and applies the transformer partial discharge data monitoring model.

[0058] According to the above scheme, when a fault alarm is triggered, the maintenance personnel perform maintenance measures, the user interface records the maintenance personnel's maintenance measures, and feeds back to the data analysis and fault diagnosis unit, analyzes the transformer data and maintenance measures after maintenance to learn, adjust and optimize control instructions; and transmits the data to the data storage unit to store the transformer's historical partial discharge data and fault processing; by analyzing historical data and real-time data, the system can provide data-driven decision support to help maintenance personnel formulate more effective maintenance strategies.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] 1. The present invention constructs a fiber optic sensor position prediction model, which can accurately predict the optimal installation position of the fiber optic sensor, thereby improving the accuracy and reliability of monitoring data;

[0061] 2. The present invention combines real-time monitoring and fault alarm mechanisms, can respond to abnormal situations in a timely manner, quickly locate faults and issue early warnings, and improve the operational safety of the transformer;

[0062] 3. The present invention can self-optimize based on the activities and feedback of maintenance personnel and continuously improve monitoring performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A flowchart of the steps of a transformer partial discharge data monitoring method according to the present invention;

[0064] Figure 2 It is a structural schematic diagram of a transformer partial discharge data monitoring system of the present invention;

[0065] Figure 3 The present invention is a schematic structural diagram of a fiber optic sensor position prediction model for a transformer partial discharge data monitoring method. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Example: Figure 1-Figure 3 As shown, the present invention provides a technical solution, a transformer partial discharge data monitoring method, the method comprising the following steps:

[0068] Collect 100 sets of data, each set of data includes the signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the transformer model, the temperature of the transformer operating environment and the transformer's historical partial discharge data;

[0069] Preprocessing the data includes cleaning the data, processing missing values ​​and outliers, and standardizing and normalizing the data;

[0070] The processed data are stored and a data set is constructed; 70% of the data in the data set is used as a training set, and 30% of the data in the data set is used as a test set.

[0071] The data in the training set is input into the fiber optic sensor position prediction model to solve the optimization problem of support vector regression. The formula is as follows:

[0072]

[0073] Among them, min w，b represents the minimization of the objective function; ||·|| represents the Euclidean norm; C represents the regularization parameter, which is used to balance the weight between the regularization term and the error term; L ε It is expressed as an insensitive loss function, which is used to measure the predicted value f(x i ) and the true value y i The error between; ε represents the insensitive interval threshold, which is used to define the error tolerance range; n represents the number of samples; x i It is represented as the feature vector of the i-th sample; f(x i ) is expressed as the predicted value; y i Represented as the true value of the i-th sample; set the regularization parameter C = 1 and the insensitive interval threshold ε = 0.1;

[0074] Insensitive loss function L ε , the formula is as follows:

[0075] L ε (y i ,f(x i ))=max(0,|y i -f(x i )-ε|)

[0076] Among them, L ε Expressed as an insensitive loss function; f(x i ) is expressed as the predicted value; y iRepresented as the true value of the i-th sample;

[0077] The radial basis function is used as the kernel function to help the model handle nonlinear relationships. The formula is as follows:

[0078]

[0079] Among them, K(x i ,x j ) is represented as the kernel function; x i and x j Represented as two samples; β represents the parameter of the kernel function; using the radial basis function as the kernel function helps the model deal with nonlinear relationships, so that the model can capture the characteristics of partial discharge more accurately;

[0080] The decision function is obtained, and the formula is as follows:

[0081]

[0082] Among them, f(x i ) represents the predicted value; w represents the weight vector; b represents the bias term; K(x i ,x) is represented as a kernel function, which is used to calculate the sample x i The inner product between the new sample x and the original sample x. In high-dimensional space, this inner product represents the similarity between the two samples.

[0083] The support vector regression (SVR) algorithm is used to construct a fiber optic sensor position prediction model, which can accurately predict the optimal installation position of the fiber optic sensor, thereby improving the accuracy and reliability of monitoring data.

[0084] The root mean square error is calculated using the true value on the test set and the model prediction value. The formula is as follows:

[0085]

[0086] Among them, RMSE represents the root mean square error; y a It is represented as the true value of the ath sample in the test set; f(x a ) represents the model's response to the a-th test sample x a The predicted value; M represents the number of test samples; according to the result of the root mean square error, the model parameters are adjusted to optimize the model results;

[0087] The test set has 30 true values ​​of samples, namely y 1 ,y 2 , ..., y 30 The model prediction values ​​are f(x 1 ), f(x 2 ), ..., f(x 30 );

[0088] The root mean square error is

[0089] The trained model is used to predict the optimal installation location of the fiber optic sensor, and the fiber optic sensor is installed according to the prediction results.

[0090] The signals collected by the optical fiber sensor are high-frequency electrical pulses or optical signals. The analog signals are converted into digital signals through an analog-to-digital converter, and then filtered, denoised and amplified.

[0091] The filtering process uses a low-pass filter, and the formula is as follows:

[0092]

[0093] Where H(f) is the frequency response function of the filter; f is the frequency; f c It is expressed as the cut-off frequency;

[0094] The denoising process uses a digital filter, and the formula is as follows:

[0095]

[0096] Among them, z k Represented as the kth output sample; x m Represents the output sample; n represents the window size of the filter;

[0097] The amplification process uses a linear amplifier, and the formula is as follows:

[0098] V out =G×V in

[0099] Among them, V out Represented as output signal; V in It represents the input signal; G represents the gain of the linear amplifier;

[0100] The signal processing technology of analog-to-digital conversion, filtering, denoising and amplification is adopted to effectively improve the signal quality, reduce noise interference and enhance the signal availability.

[0101] The preprocessed signal is used as a feature to form a feature vector; the support vector machine is used as a classifier and trained using the feature vector; the support vector machine model learns how to distinguish the partial discharge of the transformer based on the input feature vector through training; using the support vector machine as a classifier, different partial discharge situations of the transformer can be effectively distinguished; the support vector machine performs well on small sample sets and can improve the robustness of classification by maximizing the decision boundary;

[0102] The root mean square error is used to adjust the model parameters and optimize the model results. The prediction accuracy and precision of the model are verified through evaluation and verification. Cross-validation is used to verify the generalization ability and stability of the model to ensure that the model can complete the monitoring of transformer partial discharge data on different data sets.

[0103] The threshold is set according to the collected data. The formula is as follows:

[0104] T=μ+bα

[0105] Wherein, T represents the threshold value; μ represents the average value of the historical partial discharge data of the transformer; b represents the standard deviation of the historical partial discharge data of the transformer; α represents the empirical value;

[0106] The threshold T is set to 0.05. When the value of a partial discharge event is detected to be 0.07>0.05, a fault alarm is triggered.

[0107] The present invention provides another technical solution, a transformer partial discharge data monitoring system, the system comprising: a data acquisition unit, a data preprocessing unit, a data storage unit, a data analysis and fault diagnosis unit and a user interface;

[0108] The data acquisition unit collects the signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the model of the transformer, the temperature of the transformer operating environment and the historical partial discharge data of the transformer; and transmits the collected data to the data preprocessing unit;

[0109] A data preprocessing unit, used for receiving the data transmitted by the data acquisition unit and preprocessing the data;

[0110] Preprocessing includes data cleaning, filling missing values ​​with interpolation methods, identifying and replacing outliers with the IQR method, standardizing numerical features, and one-hot encoding categorical features;

[0111] The high-frequency electrical pulses or optical signals collected by the optical fiber sensor are converted from analog signals to digital signals, filtered using a low-pass filter to remove high-frequency noise, denoised using a sliding average filter, and amplified using a linear amplifier;

[0112] A data storage unit, used to receive and store data transmitted by the data preprocessing unit and the data analysis and fault diagnosis unit, and to construct a data set according to the data content;

[0113] Data analysis and fault diagnosis unit, used to train the optical fiber sensor position prediction model and transformer partial discharge data monitoring model using data sets; the optical fiber sensor position prediction model uses support vector regression; the transformer partial discharge data monitoring model uses support vector machine;

[0114] Predicting the position of the optical fiber sensor according to the trained model. When the optical fiber sensor position prediction model predicts the position of the optical fiber sensor and the optical fiber sensor is installed according to the prediction result, the system then trains and uses the transformer partial discharge data monitoring model to monitor the transformer partial discharge data.

[0115] According to the real-time data, analysis and processing are performed to identify the partial discharge of the transformer, and fault alarm and diagnosis are performed; the prediction results of the model are transmitted to the user interface; combined with the real-time monitoring and fault alarm mechanism, it is possible to respond to abnormal situations in a timely manner, quickly locate faults and issue warnings, thereby improving the operational safety of the transformer.

[0116] The user interface is used to provide a user interaction interface to display monitoring data, system status and diagnosis results; when receiving the fault alarm information sent by the data analysis and fault diagnosis unit, it triggers the fault alarm and notifies the maintenance personnel to take corresponding maintenance measures;

[0117] When a fault alarm is triggered, the maintenance personnel take maintenance measures, and the user interface records the maintenance personnel's maintenance measures and feeds them back to the data analysis and fault diagnosis unit, which analyzes the transformer data and maintenance measures after maintenance to learn, adjust and optimize control instructions;

[0118] The data is transmitted to the data storage unit for storage of historical partial discharge data and fault processing of the transformer; by analyzing historical data and real-time data, the system can provide data-driven decision support to help maintenance personnel develop more effective maintenance strategies.

[0119] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A transformer partial discharge data monitoring method, characterized in that: The method comprises the following steps: S1, collect data, preprocess the data and construct a data set; S2. constructing a fiber optic sensor position prediction model using a support vector regression algorithm and training the fiber optic sensor position prediction model; S3, adjust the model parameters and optimize the model results through the root mean square error; S4, using the trained model to predict the optimal installation position of the optical fiber sensor, and installing the optical fiber sensor according to the prediction result; S5, preprocessing the signal collected by the optical fiber sensor; S6, analyzing and processing the preprocessed signal, and outputting the result; S7. Complete the monitoring of transformer partial discharge data and fault alarm.

2. A transformer partial discharge data monitoring method according to claim 1, characterized in that: In step S1, the data includes signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the model of the transformer, the temperature of the transformer operating environment, and historical partial discharge data of the transformer; The preprocessing includes cleaning the data, processing missing values ​​and outliers, and performing standardization and normalization processing; the processed data is stored and a data set is constructed; 70% of the data in the data set is used as a training set, and 30% of the data is used as a test set.

3. A transformer partial discharge data monitoring method according to claim 2, characterized in that: In step S2, the data in the training set is input into the optical fiber sensor position prediction model to solve the optimization problem of support vector regression. The formula is as follows: Among them, min w，b represents the minimization of the objective function; ||·|| represents the Euclidean norm; C represents the regularization parameter, which is used to balance the weight between the regularization term and the error term; L ε It is expressed as an insensitive loss function, which is used to measure the predicted value f(x i ) and the true value y i The error between; ε represents the insensitive interval threshold, which is used to define the error tolerance range; n represents the number of samples; x i It is represented as the feature vector of the i-th sample; f(x i ) is expressed as the predicted value; y i Represented as the true value of the i-th sample; Insensitive loss function L ε , the formula is as follows: L ε (y i ,f(x i ))=max(0,|y i -f(x i )-ε|) Among them, L ε Expressed as an insensitive loss function; f(x i ) is expressed as the predicted value; y i Represented as the true value of the i-th sample; The radial basis function is used as the kernel function to help the model handle nonlinear relationships. The formula is as follows: Among them, K(x i ,x j ) is represented as kernel function; x i and x j Represented as two samples; β represents the parameter of the kernel function; The decision function is obtained, and the formula is as follows: Among them, f(x i ) represents the predicted value; w represents the weight vector; b represents the bias term; K(x i ,x) is represented as a kernel function, which is used to calculate the sample x i The inner product between the new sample x and the original sample x. In high-dimensional space, this inner product represents the similarity between the two samples.

4. A transformer partial discharge data monitoring method according to claim 1, characterized in that: In step S3, the root mean square error is calculated using the true value on the test set and the model prediction value. The formula is as follows: Among them, RMSE represents the root mean square error; y a It is represented as the true value of the ath sample in the test set; f(x a ) represents the model's response to the a-th test sample x a The predicted value of; M represents the number of test samples; According to the RMS error results, the model parameters are adjusted to optimize the model results.

5. A transformer partial discharge data monitoring method according to claim 1, characterized in that: In step S5, the signal collected by the optical fiber sensor is a high-frequency electrical pulse or optical signal, and the analog signal is converted into a digital signal through an analog-to-digital converter, and filtered, denoised and amplified; The filtering process uses a low-pass filter, and the formula is as follows: Where H(f) is the frequency response function of the filter; f is the frequency; f c It is expressed as the cut-off frequency; The denoising process uses a digital filter, and the formula is as follows: Among them, z k Represented as the kth output sample; x m Represents the output sample; n represents the window size of the filter; The amplification process uses a linear amplifier, and the formula is as follows: In out =G×H in Among them, V out Represented as output signal; V in Represents the input signal; G represents the amplification factor of the linear amplifier.

6. A transformer partial discharge data monitoring method according to claim 1, characterized in that: In step S6, a feature vector is formed based on the preprocessed signal as a feature; a support vector machine is used as a classifier and trained using the feature vector; the support vector machine model learns through training how to distinguish the partial discharge of the transformer based on the input feature vector; the model parameters are adjusted using the root mean square error to optimize the model results; the prediction accuracy and precision of the model are verified through evaluation and verification; cross-validation is used to verify the generalization ability and stability of the model to ensure that the model can complete the monitoring of the partial discharge data of the transformer on different data sets.

7. A transformer partial discharge data monitoring method according to claim 2, characterized in that: In step S7, a threshold is set according to the collected data, and the formula is as follows: T=μ+bα Wherein, T represents the threshold value; μ represents the average value of the historical partial discharge data of the transformer; b represents the standard deviation of the historical partial discharge data of the transformer; α represents the empirical value; When the partial discharge data of the monitored transformer exceeds the threshold, an early warning is triggered.

8. A transformer partial discharge data monitoring system, characterized in that: The system includes: a data acquisition unit, a data preprocessing unit, a data storage unit, a data analysis and fault diagnosis unit and a user interface; The data acquisition unit collects the signal strength and frequency characteristics captured by the optical fiber sensor at different positions, the model of the transformer, the temperature of the transformer operating environment and the historical partial discharge data of the transformer; and transmits the data to the data preprocessing unit; The data preprocessing unit is used to receive the data transmitted by the data acquisition unit and preprocess the data, wherein the preprocessing includes data cleaning, processing missing data values ​​and abnormal values, and performing standardization and normalization processing; converting the high-frequency electrical pulse or optical signal collected by the optical fiber sensor from an analog signal to a digital signal, and performing filtering processing, denoising processing and amplification processing; The data storage unit is used to receive and store the data transmitted by the data preprocessing unit and the data analysis and fault diagnosis unit, and to construct a data set according to the data content; The data analysis and fault diagnosis unit is used to train the optical fiber sensor position prediction model and the transformer partial discharge data monitoring model using the data set; predict the position of the optical fiber sensor and the transformer partial discharge data according to the trained model; and analyze and process the real-time data to identify the transformer partial discharge situation, and perform fault alarm and diagnosis; and transmit the prediction results of the model to the user interface; The user interface is used to provide a user interaction interface to display monitoring data, system status and diagnosis results; when the fault alarm information sent by the data analysis and fault diagnosis unit is received, the fault alarm is triggered to notify the maintenance personnel to take corresponding maintenance measures.

9. A transformer partial discharge data monitoring system according to claim 8, characterized in that: The optical fiber sensor position prediction model adopts support vector regression; the transformer partial discharge data monitoring model adopts support vector machine; when the optical fiber sensor position prediction model predicts the position of the optical fiber sensor and the optical fiber sensor is installed according to the prediction result, the data analysis and fault diagnosis unit then trains and uses the transformer partial discharge data monitoring model.

10. A transformer partial discharge data monitoring system according to claim 8, characterized in that: When a fault alarm is triggered, the maintenance personnel take maintenance measures, and the user interface records the maintenance personnel's maintenance measures and feeds them back to the data analysis and fault diagnosis unit, which analyzes the transformer data and maintenance measures after maintenance to learn, adjust and optimize control instructions; The data is transmitted to the data storage unit to store the transformer's historical partial discharge data and fault processing.

Images