Method for predicting electrical strength of oil paper insulation of converter transformer at different service times and temperatures

Abstract

The invention discloses an electrical strength prediction method for converter transformer oil paper insulation at different service times and temperatures, and relates to the technical field of monitoring analysis, and the method comprises the steps: collecting the electrical strength data of the converter transformer oil paper insulation at different service times and temperatures, and building an initial data set; preprocessing the initial data set to obtain a preprocessed data set; based on the preprocessed data set, establishing a relation model among the electrical strength, the service time and the temperature by adopting a machine learning algorithm; training and optimizing the relation model by using a cross validation method to obtain an optimal electrical strength prediction model; and using the optimal electrical strength prediction model to predict the electrical strength of the oil paper insulation of the converter transformer at different service times and temperatures, and outputting a prediction result. The method has the effect of improving the prediction accuracy.

Description

Technical Field

[0001] The present application relates to the field of monitoring and analysis technology, and in particular to a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures. Background Art

[0002] As a key component of high-voltage direct current (HVDC) transmission systems, converter transformers are responsible for converting electrical energy between AC and DC grids. With increasing power demand and longer transmission distances, HVDC technology has become a crucial transmission method in modern power systems. During operation, converter transformers face not only traditional AC stresses but also high-frequency, complex harmonic voltages, and high temperatures. These operating environments place higher demands on the transformer's oil-paper insulation material, whose performance is directly related to the converter transformer's operational safety and reliability.

[0003] In the related art, oil-paper insulation is one of the widely used insulation materials in converter transformers. Its good electrical properties and mechanical strength enable it to provide stable insulation under high voltage conditions. However, long-term high temperatures and frequent voltage fluctuations can cause the oil-paper insulation material to age, resulting in a gradual decrease in its electrical strength and an increased risk of equipment failure. The breakdown strength of oil-paper insulation is affected by many factors, especially temperature and aging time. Rising temperature accelerates the thermal degradation of the insulation material, while long-term electrical stress gradually weakens its insulation capacity. Therefore, understanding the breakdown performance of oil-paper insulation under different aging states and temperatures is of great significance for evaluating the reliability of converter transformers, predicting equipment failures, and optimizing equipment maintenance strategies. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present application provides a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures.

[0005] In a first aspect, the present application provides a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures, using the following technical solutions:

[0006] A method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures comprises the following steps:

[0007] Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures to establish an initial data set;

[0008] Preprocessing the initial data set to obtain a preprocessed data set;

[0009] Based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength, service time and temperature;

[0010] The cross-validation method is used to train and optimize the relationship model to obtain the optimal electrical strength prediction model;

[0011] The optimal electrical strength prediction model is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service time and temperature, and the prediction results are output.

[0012] Preferably, the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures are collected to establish an initial data set, specifically including:

[0013] Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures, including laboratory test data, field test data, and historical record data;

[0014] Laboratory test data: Prepare oil-paper insulation samples to simulate the insulation structure in actual converter transformers; set different temperature conditions in the laboratory and use accelerated aging tests to simulate different service times; test the electrical strength parameters of the oil-paper insulation samples, including breakdown voltage, dielectric loss factor, and partial discharge; record the test results to form a laboratory data set;

[0015] On-site test data: Online monitoring equipment is installed on the operating converter transformer to monitor the status parameters of the oil-paper insulation in real time, including temperature, humidity, partial discharge, and dielectric loss factor. Oil-paper insulation data is regularly collected for laboratory analysis to test the electrical strength of the insulation. The corresponding operating conditions of the converter transformer are also recorded. The monitored oil-paper insulation status parameters are combined with laboratory test data to form an on-site data set.

[0016] Historical data: Collect historical data of converter transformers, extract the electrical strength test results related to oil-paper insulation from the historical data, and combine them with historical ambient temperature data to form a historical data set;

[0017] The laboratory data set, field data set and historical data set are organized into the initial data set according to the preset format.

[0018] Preferably, preprocessing the initial data set to obtain a preprocessed data set specifically includes:

[0019] Preprocessing the initial data set includes data cleaning, normalization, and feature selection. The data cleaning is used to remove outliers, fill missing values, and smooth noisy data. The normalization is used to normalize the data in the initial data set. The feature selection is used to select feature data with a high correlation with electrical strength from the initial data set, including service time, temperature, and dielectric loss factor.

[0020] The preprocessed data is divided into training sets and test sets according to a preset method for model training and verification.

[0021] Preferably, based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength, service time, and temperature, specifically including:

[0022] Confirming model target data corresponding to the relational model, wherein the model target data includes output feature data, output target data, and model type data;

[0023] Build a relationship model based on model target data and linear regression algorithm.

[0024] Preferably, the relationship model is trained and optimized using a cross-validation method to obtain an optimal electrical strength prediction model, specifically including:

[0025] Divide the processed dataset into feature matrix and target variables;

[0026] Select K-fold cross-validation, and for each fold of data, use the current fold as the validation set and the remaining folds as the training set. Train the relational model on the training set, evaluate the model performance on the validation set, and record the evaluation indicators.

[0027] Calculate the average evaluation index corresponding to all folds, and set the average evaluation index as the performance index corresponding to the relational model;

[0028] The model hyperparameters are adjusted according to the performance indicators corresponding to the relationship model, and the optimal electrical strength prediction model is selected.

[0029] Preferably, the optimal electrical strength prediction model is used to predict the electrical strength of the oil-paper insulation of the converter transformer at different service times and temperatures, and the prediction results are output, specifically including:

[0030] The data to be predicted is obtained, and the data is input into the optimal electrical strength prediction model, and the prediction result is output in a preset manner.

[0031] In a second aspect, the present application provides a system for predicting the electrical strength of oil-paper insulation of a converter transformer at different service times and temperatures, which adopts the following technical solutions:

[0032] A system for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures includes:

[0033] The data collection module is used to collect the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures to establish an initial data set;

[0034] A preprocessing module is used to preprocess the initial data set to obtain a preprocessed data set;

[0035] Model validation module, which uses machine learning algorithms to establish a relationship model between electrical strength, service time, and temperature based on the preprocessed data set;

[0036] Model optimization module, used to train and optimize the relationship model using the cross-validation method to obtain the optimal electrical strength prediction model;

[0037] The prediction module is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service times and temperatures using the optimal electrical strength prediction model and output the prediction results.

[0038] In a third aspect, the present application provides a computer-readable storage medium storing instructions. When the instructions are executed on a computer, the computer executes any one of the above-mentioned methods for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures.

[0039] In summary, this application has the following beneficial technical effects:

[0040] The present application provides a method for predicting the electrical strength of the oil-paper insulation of a converter transformer at different service times and temperatures. The method collects the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures, establishes an initial data set, and preprocesses the initial data set to obtain a preprocessed data set. Based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength and service time and temperature. The relationship model is trained and optimized using a cross-validation method to obtain an optimal electrical strength prediction model. The optimal electrical strength prediction model is then used to predict the electrical strength of the oil-paper insulation of the converter transformer at different service times and temperatures, and the prediction results are output, thereby effectively improving the accuracy of the electrical strength prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 This is a flow chart of a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to an embodiment of the present application.

[0043] Figure 2 This is a schematic diagram of a system for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to an embodiment of the present application. DETAILED DESCRIPTION

[0044] The following is combined with Figure 1-2 This application is described in further detail.

[0045] Example 1

[0046] The embodiments of the present application disclose a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures.

[0047] Reference Figure 1 A method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures comprises the following steps:

[0048] Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures to establish an initial data set;

[0049] Preprocessing the initial data set to obtain a preprocessed data set;

[0050] Based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength, service time and temperature;

[0051] The cross-validation method is used to train and optimize the relationship model to obtain the optimal electrical strength prediction model;

[0052] The optimal electrical strength prediction model is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service time and temperature, and the prediction results are output.

[0053] For example, the electrical strength data of the oil-paper insulation of a converter transformer at different service times (1 year, 3 years, 5 years, 10 years) and temperatures (20°C, 40°C, 60°C, 80°C) are collected to establish an initial data set;

[0054] Preprocess the initial data set, including removing outliers, filling missing values, normalizing and selecting features, to obtain the preprocessed data set;

[0055] Based on the preprocessed data set, the support vector machine algorithm is used to establish the relationship model between electrical strength, service time and temperature;

[0056] The relationship model was trained and optimized using the 5-fold cross-validation method to obtain the optimal model;

[0057] The optimal model is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service time and temperature, and the prediction results are output.

[0058] It should be noted that the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures were collected to establish the initial data set, which specifically includes:

[0059] Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures, including laboratory test data, field test data, and historical record data;

[0060] Laboratory test data: Prepare oil-paper insulation samples to simulate the insulation structure in actual converter transformers; set different temperature conditions in the laboratory and use accelerated aging tests to simulate different service times; test the electrical strength parameters of the oil-paper insulation samples, including breakdown voltage, dielectric loss factor, and partial discharge; record the test results to form a laboratory data set;

[0061] Specifically, different temperature gradients are set (such as 20°C, 40°C, 60°C, and 80°C); different service times (such as 1 year, 3 years, 5 years, and 10 years) are simulated through accelerated aging tests; and the electrical strength parameters of the samples are tested at each time point and temperature point.

[0062] On-site test data: Online monitoring equipment is installed on the operating converter transformer to monitor the status parameters of the oil-paper insulation in real time, including temperature, humidity, partial discharge, and dielectric loss factor. Oil-paper insulation data is regularly collected for laboratory analysis to test the electrical strength of the insulation. The corresponding operating conditions of the converter transformer are also recorded. The monitored oil-paper insulation status parameters are combined with laboratory test data to form an on-site data set.

[0063] Historical data: Collect historical data of converter transformers, extract the electrical strength test results related to oil-paper insulation from the historical data, and combine them with historical ambient temperature data to form a historical data set;

[0064] The laboratory data set, field data set and historical data set are organized into the initial data set according to the preset format.

[0065] Specifically, organize laboratory data sets, field data sets, and historical data sets into structured data sets; ensure that the data format is unified, including fields such as time, temperature, humidity, and electrical strength parameters; use a database or spreadsheet to store data for subsequent analysis and modeling.

[0066] Furthermore, the initial data set is preprocessed to obtain a preprocessed data set, specifically including:

[0067] Preprocessing the initial data set includes data cleaning, normalization, and feature selection. The data cleaning is used to remove outliers, fill missing values, and smooth noisy data. The normalization is used to normalize the data in the initial data set. The feature selection is used to select feature data with a high correlation with electrical strength from the initial data set, including service time, temperature, and dielectric loss factor.

[0068] The preprocessed data is divided into training sets and test sets according to a preset method for model training and verification.

[0069] Furthermore, based on the preprocessed data set, a machine learning algorithm was used to establish a relationship model between electrical strength, service time, and temperature, including:

[0070] Confirming model target data corresponding to the relational model, wherein the model target data includes output feature data, output target data, and model type data;

[0071] Build a relationship model based on model target data and linear regression algorithm.

[0072] Furthermore, the cross-validation method is used to train and optimize the relationship model to obtain the optimal electrical strength prediction model, which specifically includes:

[0073] Divide the processed dataset into feature matrix and target variables;

[0074] Select K-fold cross-validation, and for each fold of data, use the current fold as the validation set and the remaining folds as the training set. Train the relational model on the training set, evaluate the model performance on the validation set, and record the evaluation indicators.

[0075] Calculate the average evaluation index corresponding to all folds, and set the average evaluation index as the performance index corresponding to the relational model;

[0076] The model hyperparameters are adjusted according to the performance indicators corresponding to the relationship model, and the optimal electrical strength prediction model is selected.

[0077] Specifically, in the embodiment of the present application, the optimal electrical strength prediction model is set as:

[0078] z=z0+a·x+b·y+c·x2+d·y2+f·x·y

[0079] Among them, z represents the electrical strength, x and y represent the aging days and temperature respectively, z0 represents the preset basic value, and a, b, c, d, and f represent the preset coefficients respectively.

[0080] Furthermore, the optimal electrical strength prediction model is used to predict the electrical strength of the converter transformer oil-paper insulation at different service times and temperatures, and the prediction results are output, including:

[0081] The data to be predicted is obtained, and the data is input into the optimal electrical strength prediction model, and the prediction result is output in a preset manner.

[0082] Example 2

[0083] The embodiments of the present application also disclose a system for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures.

[0084] Reference Figure 2 , a system for predicting the electrical strength of oil-paper insulation of converter transformers under different service times and temperatures, including:

[0085] The data collection module is used to collect the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures to establish an initial data set;

[0086] A preprocessing module is used to preprocess the initial data set to obtain a preprocessed data set;

[0087] Model validation module, which uses machine learning algorithms to establish a relationship model between electrical strength, service time, and temperature based on the preprocessed data set;

[0088] Model optimization module, used to train and optimize the relationship model using the cross-validation method to obtain the optimal electrical strength prediction model;

[0089] The prediction module is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service times and temperatures using the optimal electrical strength prediction model and output the prediction results.

[0090] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in similar ways. As long as they do not deviate from the concept of the invention, they should all fall within the scope of protection of the present invention.

[0091] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0092] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention.

Claims

1. A method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures, characterized in that: The following steps are involved: Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures to establish an initial data set; Preprocessing the initial data set to obtain a preprocessed data set; Based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength, service time and temperature; The cross-validation method is used to train and optimize the relationship model to obtain the optimal electrical strength prediction model; The optimal electrical strength prediction model is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service time and temperature, and the prediction results are output.

2. The method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to claim 1, characterized in that: Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures to establish an initial data set, including: Collect electrical strength data of converter transformer oil-paper insulation at different service times and temperatures, including laboratory test data, field test data, and historical record data; Laboratory test data: Prepare oil-paper insulation samples to simulate the insulation structure in actual converter transformers; set different temperature conditions in the laboratory and use accelerated aging tests to simulate different service times; test the electrical strength parameters of the oil-paper insulation samples, including breakdown voltage, dielectric loss factor, and partial discharge; record the test results to form a laboratory data set; On-site test data: Online monitoring equipment is installed on the operating converter transformer to monitor the status parameters of the oil-paper insulation in real time, including temperature, humidity, partial discharge, and dielectric loss factor. Oil-paper insulation data is regularly collected for laboratory analysis to test the electrical strength of the insulation. The corresponding operating conditions of the converter transformer are also recorded. The monitored oil-paper insulation status parameters are combined with laboratory test data to form an on-site data set. Historical data: Collect historical data of converter transformers, extract the electrical strength test results related to oil-paper insulation from the historical data, and combine them with historical ambient temperature data to form a historical data set; The laboratory data set, field data set and historical data set are organized into the initial data set according to the preset format.

3. The method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to claim 2, characterized in that: Preprocess the initial data set to obtain a preprocessed data set, specifically including: Preprocessing the initial data set includes data cleaning, normalization, and feature selection. The data cleaning is used to remove outliers, fill missing values, and smooth noisy data. The normalization is used to normalize the data in the initial data set. The feature selection is used to select feature data with a high correlation with electrical strength from the initial data set, including service time, temperature, and dielectric loss factor. The preprocessed data is divided into training sets and test sets according to a preset method for model training and verification.

4. The method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to claim 3, characterized in that: Based on the preprocessed data set, a machine learning algorithm is used to establish a relationship model between electrical strength, service time, and temperature. Specifically, the following are the steps: Confirming model target data corresponding to the relational model, wherein the model target data includes output feature data, output target data, and model type data; Build a relationship model based on model target data and linear regression algorithm.

5. The method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to claim 4, characterized in that: The cross-validation method is used to train and optimize the relationship model to obtain the optimal electrical strength prediction model, which includes: Divide the processed dataset into feature matrix and target variables; Select K-fold cross-validation, and for each fold of data, use the current fold as the validation set and the remaining folds as the training set. Train the relational model on the training set, evaluate the model performance on the validation set, and record the evaluation indicators. Calculate the average evaluation index corresponding to all folds, and set the average evaluation index as the performance index corresponding to the relational model; The model hyperparameters are adjusted according to the performance indicators corresponding to the relationship model, and the optimal electrical strength prediction model is selected.

6. The method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures according to claim 5, characterized in that: The optimal electrical strength prediction model is used to predict the electrical strength of the oil-paper insulation of the converter transformer at different service times and temperatures, and the prediction results are output, including: The data to be predicted is obtained, and the data is input into the optimal electrical strength prediction model, and the prediction result is output in a preset manner.

7. A system for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures, applied to a method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures as claimed in any one of claims 1 to 6, characterized in that: include: The data collection module is used to collect the electrical strength data of the oil-paper insulation of the converter transformer at different service times and temperatures to establish an initial data set; A preprocessing module is used to preprocess the initial data set to obtain a preprocessed data set; Model validation module, which uses machine learning algorithms to establish a relationship model between electrical strength, service time, and temperature based on the preprocessed data set; Model optimization module, used to train and optimize the relationship model using the cross-validation method to obtain the optimal electrical strength prediction model; The prediction module is used to predict the electrical strength of the oil-paper insulation of the converter transformer under different service times and temperatures using the optimal electrical strength prediction model and output the prediction results.

8. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are run on a computer, the computer is caused to execute the method for predicting the electrical strength of oil-paper insulation of a converter transformer under different service times and temperatures as claimed in any one of claims 1 to 6.

Images