A method and device for predicting the voltage-second characteristic parameters of an insulator

Abstract

The application discloses a method and device for predicting a voltage-second characteristic parameter of an insulator. The method comprises the following steps: inputting a working condition parameter into a prediction model, and outputting a predicted breakdown time under each preset breakdown voltage; a model establishment process is as follows: performing a voltage-second characteristic test under each historical working condition parameter, determining a training set and a test set, training the prediction model through the training set based on a pre-set random forest algorithm, updating model parameters to obtain an updated prediction model, verifying the prediction model through the test set to update the training set when the number of times of updating the model parameters does not reach a preset number of times, and returning to continue training until the preset number of times is reached, and determining a final prediction model. It can be seen that the prediction model is trained by using the random forest algorithm, so that the prediction model has high tolerance to abnormal values and noise in the training process, and the prediction model can output a breakdown time on the basis of multiple breakdown voltages, and therefore, a more accurate voltage-second characteristic parameter is obtained.

Description

Technical Field

[0001] This application relates to the field of power insulation and lightning protection, and more specifically, to a method and device for predicting the volt-second characteristic parameters of an insulator. Background Technology

[0002] With the continuous development of power systems and the frequent occurrence of power outages caused by lightning strikes, researchers are dedicated to the study of lightning protection for power systems. Insulators, as a widely used component in power system lines, have lightning flashover criteria that form the basis for analyzing the lightning withstand performance of power transmission lines. The lightning flashover criteria for insulators are determined by analyzing and comparing the overvoltages across the insulator terminals to identify whether a flashover has occurred, thereby obtaining the backflashover withstand level of the power system transmission line. Therefore, obtaining accurate flashover characteristics of insulators can effectively pinpoint the backflashover withstand level of insulator-lined power transmission lines.

[0003] Currently, researchers use machine learning algorithms to build models to predict flashover discharge voltage, but they ignore the influence of the duration of voltage applied across the insulator on the lightning flashover criterion. This makes the method of predicting only flashover discharge voltage inaccurate, and therefore the positioning of the backflashover withstand level of the insulator line is not accurate enough. Summary of the Invention

[0004] In view of the above problems, this application is made to provide a method and device for predicting the volt-second characteristic parameters of an insulator, so as to accurately determine the volt-second characteristic breakdown time of the insulator and accurately locate the backflashover withstand level of the insulator line.

[0005] To achieve the above objectives, the following specific solutions are proposed:

[0006] A method for predicting the volt-second characteristic parameters of an insulator includes:

[0007] Obtain the operating parameters of the insulator;

[0008] The operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator under each preset breakdown voltage is output.

[0009] The process of establishing the volt-second characteristic prediction model includes:

[0010] Obtain various historical operating condition parameters of the insulator when it is broken down;

[0011] Volt-second characteristic tests were conducted under each historical operating condition parameter, and the training set and test set were determined based on the test results under various historical operating condition parameters.

[0012] For the latest volt-second characteristic prediction model, based on the pre-set random forest algorithm, the volt-second characteristic prediction model is trained using the training set, the model parameters of the volt-second characteristic prediction model are updated, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters.

[0013] When the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, the volt-second characteristic prediction model is verified through the test set, the training set is updated to obtain the updated training set, and the process of training the volt-second characteristic prediction model based on the pre-set random forest algorithm through the training set for the latest volt-second characteristic prediction model is returned, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1.

[0014] When the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates, the latest volt-second characteristic prediction model is determined as the final volt-second characteristic prediction model.

[0015] Optionally, for the latest volt-second characteristic prediction model, based on a pre-set random forest algorithm, the volt-second characteristic prediction model is trained using the training set to update the model parameters, resulting in an updated volt-second characteristic prediction model, including:

[0016] The latest volt-second characteristic prediction model uses a resampling algorithm to divide the training set into multiple training subsets;

[0017] For each training subset, a decision tree model is performed on the training subset to obtain the decision tree for the training subset;

[0018] Determine the prediction result of each decision tree for its corresponding training subset;

[0019] The prediction results of each decision tree are voted on to determine the final prediction result of the volt-second characteristic prediction model;

[0020] Based on the prediction results of each decision tree and the final prediction result, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model.

[0021] Optionally, the method further includes:

[0022] The average of the prediction results of each decision tree is calculated to obtain the final prediction result of the volt-second characteristic prediction model.

[0023] Optionally, the step of conducting volt-second characteristic tests under each historical operating condition parameter, and determining the training set and test set based on the volt-second characteristic test results under various historical operating condition parameters, includes:

[0024] Under each historical operating condition parameter, the insulator was subjected to several volt-second characteristic tests to obtain multiple sets of volt-second characteristic test data. Each set of volt-second characteristic test data included breakdown time and breakdown voltage.

[0025] The test data of each set of volt-second characteristics for each historical operating condition parameter were preprocessed and normalized to obtain the sample data of that historical operating condition parameter.

[0026] Based on sample data of various historical operating conditions, a sample dataset is constructed, and the sample dataset is divided into a training set and a test set.

[0027] Optionally, the operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator at each preset breakdown voltage is output, including:

[0028] The operating parameters are input into a pre-established volt-second characteristic prediction model, which drives the volt-second characteristic prediction model to output intermediate prediction values ​​at each preset breakdown voltage.

[0029] The intermediate predicted value output at each preset breakdown voltage is inversely normalized to obtain the predicted breakdown time at that preset breakdown voltage.

[0030] Optionally, after inputting the operating condition parameters into a pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator at each preset breakdown voltage, the method further includes:

[0031] Based on each preset breakdown voltage and the predicted breakdown time corresponding to that preset breakdown voltage, determine the volt-second characteristic scatter plot of that preset breakdown voltage;

[0032] Curve fitting is performed on the volt-second characteristic scatter points of each preset breakdown voltage to obtain the volt-second characteristic curve.

[0033] A device for predicting the volt-second characteristic parameters of an insulator includes:

[0034] Operating condition parameter acquisition unit, used to acquire the operating condition parameters of the insulator;

[0035] The model output unit is used to input the operating condition parameters into a pre-established volt-second characteristic prediction model and output the predicted breakdown time of the insulator under each preset breakdown voltage.

[0036] The historical parameter acquisition unit is used to acquire various historical operating condition parameters of the insulator when it is broken down.

[0037] The sample set determination unit is used to conduct volt-second characteristic tests under each historical operating condition parameter, and to determine the training set and test set based on the volt-second characteristic test results under various historical operating condition parameters.

[0038] The model parameter update unit is used to train the latest volt-second characteristic prediction model using the training set based on a pre-set random forest algorithm, update the model parameters of the volt-second characteristic prediction model, obtain the updated volt-second characteristic prediction model, and increment the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters.

[0039] The return execution unit is used to verify the volt-second characteristic prediction model through the test set when the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, update the training set to obtain the updated training set, and return to the execution unit for updating the model parameters.

[0040] The final model determination unit is used to determine the latest volt-second characteristic prediction model as the final volt-second characteristic prediction model when the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates.

[0041] Optionally, the model parameter update unit includes:

[0042] The training subset partitioning unit is used to drive the latest volt-second characteristic prediction model to divide the training set into multiple training subsets through a resampling algorithm.

[0043] The decision tree modeling unit is used to perform decision tree modeling on each training subset to obtain the decision tree of the training subset.

[0044] The subset result prediction unit is used to determine the prediction result of each decision tree for its corresponding training subset;

[0045] The voting unit is used to vote on the prediction results of each decision tree to determine the final prediction result of the volt-second characteristic prediction model.

[0046] The update unit is used to update the model parameters of the volt-second characteristic prediction model based on the prediction results of each decision tree and the final prediction result, so as to obtain the updated volt-second characteristic prediction model.

[0047] Optionally, the device may also include:

[0048] The average value calculation unit is used to calculate the average value of the prediction results of each decision tree to obtain the final prediction result of the volt-second characteristic prediction model.

[0049] Optionally, the sample set determination unit includes:

[0050] The volt-second characteristic test unit is used to perform several volt-second characteristic tests on the insulator under each historical operating condition parameter to obtain multiple sets of volt-second characteristic test data. Each set of volt-second characteristic test data includes breakdown time and breakdown voltage.

[0051] The sample data acquisition unit is used to preprocess and normalize the test data of each set of volt-second characteristics for each historical operating condition parameter to obtain the sample data of that historical operating condition parameter.

[0052] The sample set data partitioning unit is used to construct a sample dataset based on sample data of various historical operating condition parameters, and to divide the sample dataset into a training set and a test set.

[0053] Optionally, the model output unit includes:

[0054] The intermediate prediction value output unit is used to input the operating condition parameters into a pre-established volt-second characteristic prediction model and drive the volt-second characteristic prediction model to output intermediate prediction values ​​at each preset breakdown voltage.

[0055] The inverse normalization unit is used to inverse normalize the intermediate predicted value output at each preset breakdown voltage to obtain the predicted breakdown time at that preset breakdown voltage.

[0056] Optionally, the device may also include:

[0057] The volt-second characteristic scatter point determination unit is used to determine the volt-second characteristic scatter point of the preset breakdown voltage based on each preset breakdown voltage and the predicted breakdown time corresponding to the preset breakdown voltage after inputting the operating condition parameters into the pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator under each preset breakdown voltage.

[0058] The curve fitting unit is used to perform curve fitting on the volt-second characteristic scatter points of each preset breakdown voltage to obtain the volt-second characteristic curve.

[0059] By means of the above technical solution, this application obtains the operating condition parameters of the insulator, inputs the operating condition parameters into a pre-established volt-second characteristic prediction model, and outputs the predicted breakdown time of the insulator under each preset breakdown voltage. The process of establishing the volt-second characteristic prediction model is as follows: obtaining multiple historical operating condition parameters of the insulator under which it has been broken down, conducting volt-second characteristic tests under each historical operating condition parameter, and determining the training set and test set based on the test results under various historical operating condition parameters. For the latest volt-second characteristic prediction model, training the volt-second characteristic prediction model using the training set based on a pre-set random forest algorithm, updating the model parameters of the volt-second characteristic prediction model, obtaining the updated volt-second characteristic prediction model, and incrementing the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1. The initial version is a volt-second characteristic prediction model with pre-set default model parameters. Further, when the number of times the model parameters of the volt-second characteristic prediction model are updated does not reach the preset update count, the volt-second characteristic prediction model is validated using the test set, the training set is updated to obtain an updated training set, and the process of training the volt-second characteristic prediction model using the training set based on a pre-set random forest algorithm, updating the model parameters of the volt-second characteristic prediction model, and incrementing the number of times the model parameters of the volt-second characteristic prediction model are updated is performed. When the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset update count, the latest volt-second characteristic prediction model is determined as the final volt-second characteristic prediction model. It can be seen that the volt-second characteristic prediction model is trained using the random forest algorithm, which makes the volt-second characteristic prediction model have a good tolerance for outliers and noise during the training process. Moreover, the volt-second characteristic prediction model can output the predicted breakdown time under various breakdown voltages, thus obtaining more accurate volt-second characteristic parameters, making the positioning of the backflashover lightning withstand level of the insulator line more accurate. Attached Figure Description

[0060] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0061] Figure 1 A flowchart illustrating the prediction of the volt-second characteristic parameters of an insulator provided in an embodiment of this application;

[0062] Figure 2 A schematic diagram of a process for constructing a volt-second characteristic prediction model provided in an embodiment of this application;

[0063] Figure 3 A schematic diagram of a device for predicting the volt-second characteristic parameters of an insulator provided in an embodiment of this application;

[0064] Figure 4 This is a schematic diagram of the structure of a device for predicting the volt-second characteristic parameters of an insulator, provided in an embodiment of this application. Detailed Implementation

[0065] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0066] The proposed solution can be implemented based on a terminal with data processing capabilities, such as a computer, server, or cloud platform.

[0067] Next, combined Figure 1 The method for predicting the volt-second characteristic parameters of insulators in this application may include the following steps:

[0068] Step S110: Obtain the operating parameters of the insulator.

[0069] Specifically, the operating parameters of an insulator can represent the parameters required for the volt-second characteristic test, such as the test name, insulator model, insulator dry arc distance, material dielectric constant, temperature, humidity, atmospheric pressure, applied voltage polarity and voltage amplitude.

[0070] Step S120: Input the operating parameters into the pre-established volt-second characteristic prediction model and output the predicted breakdown time of the insulator under each preset breakdown voltage.

[0071] The preset breakdown voltage can be represented as not less than the flashover discharge voltage U. 50％ The voltage, and the preset breakdown voltage can be customized.

[0072] For example, when multiple preset breakdown voltages are 1.0U... 50％ 1.1U 50％ 1.2U 50％ 1.4U 50％ 1.6U 50％ 1.8U 50％ and 2.0U 50％ When the volt-second characteristic prediction model receives the operating condition parameters as input, it can output 1.0U. 50％ Predicted breakdown time under 1.1U 50％Predicted breakdown time under 1.2U 50％ Predicted breakdown time under 1.4U 50％ Predicted breakdown time under 1.6U 50％ Predicted breakdown time under 1.8U 50％ Predicted breakdown time and 2.0U 50％ The predicted breakdown time.

[0073] Specifically, such as Figure 2 As shown, the process of establishing a volt-second characteristic prediction model may include the following steps:

[0074] Step S121: Obtain various historical operating condition parameters of the insulator being broken down.

[0075] Specifically, each historical operating condition parameter of an insulator breakdown can include multiple operating condition parameters from each historical volt-second characteristic test, such as test name, insulator model, insulator dry arc distance, material dielectric constant, temperature, humidity, atmospheric pressure, applied voltage polarity, applied voltage amplitude, and scatter plot data of the volt-second characteristic of the breakdown, which can include breakdown time and breakdown voltage.

[0076] Among them, the operating condition parameter values ​​can be different between various historical operating condition parameters.

[0077] Step S122: Conduct volt-second characteristic tests under each historical operating condition parameter, and determine the training set and test set based on the volt-second characteristic test results under various historical operating condition parameters.

[0078] Understandably, conducting volt-second characteristic tests under various historical operating parameters can yield multiple volt-second characteristic test results, which can then be used as training and testing sets for model training.

[0079] The training set serves as the training material for the model to enhance its parameters, while the test set verifies whether the model has been trained correctly during the training process.

[0080] Step S123: For the latest volt-second characteristic prediction model, based on the pre-set random forest algorithm, train the volt-second characteristic prediction model through the training set, update the model parameters of the volt-second characteristic prediction model, obtain the updated volt-second characteristic prediction model, and increment the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1.

[0081] The initial version of the volt-second characteristic prediction model was a volt-second characteristic prediction model with default model parameters pre-set.

[0082] Specifically, the random forest algorithm can be represented by randomly selecting a subset of samples from the original training set to form a subset, so that the decision trees are trained on different subsets, thereby enhancing the diversity of the model. Then, when selecting the best splitting feature for each node of each decision tree, only a subset of randomly selected features need to be considered, thereby preventing certain features from having too much influence on the entire model and improving the robustness of the trained model. Finally, the final regression result of the model can be obtained by averaging or weighted averaging the prediction results of multiple decision trees.

[0083] It is understandable that, since the volt-second characteristic prediction model becomes more accurate after multiple rounds of training and the model parameters are updated multiple times, the model parameter update round marker can be used to record the current training round number after each round of training.

[0084] Step S124: Determine whether the number of times the model parameters of the volt-second characteristic prediction model have been updated has reached the preset number of updates. If not, proceed to step S125; if yes, proceed to step S126.

[0085] Specifically, the preset update count can represent the total number of training rounds required for the volt-second characteristic prediction model.

[0086] Step S125: Validate the volt-second characteristic prediction model using the test set, update the training set to obtain the updated training set, and return to execute step S123.

[0087] Understandably, training the volt-second characteristic prediction model with different training sets can improve the accuracy of the model parameters.

[0088] Specifically, when the number of times the model parameters of the volt-second characteristic prediction model are updated does not reach the preset number of updates, it means that the volt-second characteristic prediction model still needs to undergo another round of training. Therefore, after verifying the volt-second characteristic prediction model through the test set and updating the training set to obtain the updated training set, the process can return to step S123.

[0089] Step S126: Determine the latest volt-second characteristic prediction model as the final volt-second characteristic prediction model.

[0090] Specifically, when the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates, it can be said that the volt-second characteristic prediction model has been trained for the preset total number of training rounds, and then the latest volt-second characteristic prediction model can be determined as the final volt-second characteristic prediction model.

[0091] The insulator volt-second characteristic parameter prediction method provided in this embodiment obtains the insulator's operating condition parameters, inputs these parameters into a pre-established volt-second characteristic prediction model, and outputs the predicted breakdown time of the insulator at each preset breakdown voltage. The process of establishing the volt-second characteristic prediction model is as follows: obtaining multiple historical operating condition parameters of the insulator under which it has been broken down, conducting volt-second characteristic tests under each historical operating condition parameter, and determining training and testing sets based on the test results under various historical operating condition parameters. For the latest volt-second characteristic prediction model, training the volt-second characteristic prediction model using the training set based on a pre-set random forest algorithm, updating the model parameters of the volt-second characteristic prediction model, obtaining the updated volt-second characteristic prediction model, and incrementing the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1. The initial version of the voltage-second characteristic prediction model is a voltage-second characteristic prediction model with pre-set default model parameters. Further, when the number of times the model parameters of the voltage-second characteristic prediction model are updated does not reach the preset number of updates, the voltage-second characteristic prediction model is verified through the test set, the training set is updated to obtain an updated training set, and the process of training the voltage-second characteristic prediction model on the latest voltage-second characteristic prediction model using the training set based on the pre-set random forest algorithm is returned, the model parameters of the voltage-second characteristic prediction model are updated to obtain an updated voltage-second characteristic prediction model, and the number of times the model parameters of the voltage-second characteristic prediction model are updated is incremented by 1. When the number of times the model parameters of the voltage-second characteristic prediction model are updated reaches the preset number of updates, the latest voltage-second characteristic prediction model is determined as the final voltage-second characteristic prediction model. It can be seen that the volt-second characteristic prediction model is trained using the random forest algorithm, which makes the volt-second characteristic prediction model have a good tolerance for outliers and noise during the training process. Moreover, the volt-second characteristic prediction model can output the predicted breakdown time under various breakdown voltages, thus obtaining more accurate volt-second characteristic parameters, making the positioning of the backflashover lightning withstand level of the insulator line more accurate.

[0092] In some embodiments of this application, the process of training the latest volt-second characteristic prediction model based on a pre-set random forest algorithm using the training set to train and update the model parameters of the volt-second characteristic prediction model, as mentioned in the above embodiments, to obtain an updated volt-second characteristic prediction model, is described. This process may include:

[0093] S1 drives the latest volt-second characteristic prediction model to divide the training set into multiple training subsets through a resampling algorithm.

[0094] Specifically, the volt-second characteristic prediction model can use the bootsrap resampling algorithm to randomly select a number of sample data from the training set to form a training subset. Repeating this operation multiple times can yield multiple training subsets.

[0095] S2. For each training subset, perform decision tree modeling on the training subset to obtain the decision tree of the training subset.

[0096] S3. Determine the prediction result of each decision tree for its corresponding training subset.

[0097] Specifically, during the prediction process, each node of each decision tree only considers the features of the selected training subset when choosing the best splitting feature. Therefore, for each decision tree, the prediction result is the prediction result for its corresponding training subset.

[0098] S4. Vote on the prediction results of each decision tree to determine the final prediction result of the volt-second characteristic prediction model.

[0099] Understandably, since the prediction results of each decision tree may be different, a voting method can be used to select the prediction result with the highest number of votes from each decision tree as the final prediction result of the volt-second characteristic prediction model.

[0100] In addition, the final prediction result of the volt-second characteristic prediction model can be obtained by calculating the average of the prediction results of each decision tree.

[0101] Understandably, since the prediction results of each decision tree may differ, to take into account the differences between the decision trees, the average of the prediction results of each decision tree can be chosen as the final prediction result of the volt-second characteristic prediction model.

[0102] S5. Based on the prediction results of each decision tree and the final prediction result, update the model parameters of the volt-second characteristic prediction model to obtain the updated volt-second characteristic prediction model.

[0103] It is understandable that since the prediction results of each decision tree may be different, there may be differences between the final prediction result and the prediction results of each decision tree. Therefore, the model parameters can be adjusted according to the differences between the final prediction result and the prediction results of each decision tree to update the model parameters of the volt-second characteristic prediction model and obtain the updated volt-second characteristic prediction model.

[0104] The insulator volt-second characteristic parameter prediction method provided in this embodiment uses the latest volt-second characteristic prediction model and a resampling algorithm to divide the training set into multiple training subsets. Decision tree modeling is performed on each training subset to obtain decision trees. The prediction result of each decision tree for its corresponding training subset is determined. Then, the prediction results of each decision tree are voted on or averaged to determine the final prediction result of the model. Finally, based on the prediction results of each decision tree and the final prediction result, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model. It can be seen that training the volt-second characteristic prediction model using nonlinear modeling has a higher tolerance for outliers and noise, and the prediction accuracy is high.

[0105] In some embodiments of this application, the process of step S122, which involves conducting volt-second characteristic tests under each historical operating condition parameter and determining the training set and test set based on the test results under various historical operating condition parameters, is described. This process may include:

[0106] S1. Under each historical operating condition parameter, conduct several volt-second characteristic tests on the insulator to obtain multiple sets of volt-second characteristic test data.

[0107] Each set of volt-second characteristic test data may include breakdown time and breakdown voltage.

[0108] S2. Preprocess and normalize the test data of each group of volt-second characteristics for each historical operating condition parameter to obtain sample data for that historical operating condition parameter.

[0109] Specifically, the required feature information data can be selected based on the feature correlation, and the test data of each set of volt-second characteristics for each historical operating condition parameter can be preprocessed accordingly.

[0110] The preprocessing methods can specifically include deleting duplicate records, deleting abnormal data with obviously abnormal feature parameter values, filling in missing data, and processing data format consistency.

[0111] The following formula can be used to standardize the test data of each set of volt-second characteristics for each historical operating condition parameter:

[0112]

[0113] Where x' is the sample data obtained after standardization, and x is the feature information data selected from each group of volt-second characteristic test data. min x is the minimum value among the volt-second characteristic test data of each group. max This represents the maximum value among the volt-second characteristic test data for each group.

[0114] S3. Based on sample data of various historical operating conditions, construct a sample dataset and divide the sample dataset into a training set and a test set.

[0115] Specifically, the sample dataset can be divided into training and test sets according to a preset custom partitioning ratio.

[0116] For example, 80% of the sample dataset can be divided into the training set and 20% into the test set.

[0117] In some embodiments of this application, the process of inputting the operating condition parameters into a pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator at each preset breakdown voltage is described. This process may include:

[0118] S1. Input the operating condition parameters into the pre-established volt-second characteristic prediction model, and drive the volt-second characteristic prediction model to output the intermediate prediction value at each preset breakdown voltage.

[0119] It is understandable that, since the training samples of the volt-second characteristic prediction model are normalized training sets during the training process, the intermediate prediction values ​​output by the volt-second characteristic prediction model at each preset breakdown voltage are normalized data, and need to be denormalized to obtain the predicted breakdown time information.

[0120] S2. Normalize the intermediate predicted value output under each preset breakdown voltage to obtain the predicted breakdown time under that preset breakdown voltage.

[0121] Specifically, the intermediate predicted values ​​of the output at each preset breakdown voltage can be inversely normalized using the following formula:

[0122] U=(U max -U min )*U′+U min

[0123] Where U′ is the intermediate predicted value output by the volt-second characteristic prediction model, U max U is the predicted maximum breakdown time at a preset breakdown voltage, determined by the volt-second characteristic prediction model after training. min U is the predicted minimum breakdown time at the preset breakdown voltage, determined by the volt-second characteristic prediction model after training.

[0124] To more intuitively demonstrate the volt-second characteristics of the insulator and more clearly locate the backflashover withstand level of the line where the insulator is located, a volt-second characteristic curve can be plotted after obtaining the predicted breakdown time corresponding to each predicted breakdown voltage of the insulator. Based on this, in some embodiments of this application, after step S120 above, which involves inputting the operating parameters into the pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator at each preset breakdown voltage, a process of plotting the volt-second characteristic curve can also be included. This process may include:

[0125] S1. Based on each preset breakdown voltage and the predicted breakdown time corresponding to that preset breakdown voltage, determine the volt-second characteristic scatter plot of that preset breakdown voltage.

[0126] Specifically, the x-axis of each volt-second characteristic scatter point represents the predicted breakdown time, and the y-axis represents the preset breakdown voltage corresponding to that predicted breakdown time.

[0127] S2. Perform curve fitting on the scatter plot of the volt-second characteristic of each preset breakdown voltage to obtain the volt-second characteristic curve.

[0128] Understandably, since there is a relationship between the breakdown time and the breakdown voltage of an insulator, the volt-second characteristic scatter plots of each preset breakdown voltage can be used to obtain the volt-second characteristic curve of the breakdown voltage changing with the breakdown time, thereby more clearly locating the backflashover withstand level of the line where the insulator is located.

[0129] The apparatus for predicting the volt-second characteristic parameters of an insulator provided in the embodiments of this application will be described below. The apparatus for predicting the volt-second characteristic parameters of an insulator described below can be referred to in correspondence with the method for predicting the volt-second characteristic parameters of an insulator described above.

[0130] See Figure 3 , Figure 3 This is a schematic diagram of a device for predicting the volt-second characteristic parameters of an insulator, as disclosed in an embodiment of this application.

[0131] like Figure 3 As shown, the device may include:

[0132] Operating condition parameter acquisition unit 11 is used to acquire the operating condition parameters of the insulator;

[0133] Model output unit 12 is used to input the operating condition parameters into a pre-established volt-second characteristic prediction model and output the predicted breakdown time of the insulator under each preset breakdown voltage.

[0134] The historical parameter acquisition unit 13 is used to acquire various historical operating condition parameters of the insulator when it is broken down.

[0135] The sample set determination unit 14 is used to conduct volt-second characteristic tests under each historical operating condition parameter, and to determine the training set and test set based on the volt-second characteristic test results under various historical operating condition parameters.

[0136] The model parameter update unit 15 is used to train the latest volt-second characteristic prediction model based on a pre-set random forest algorithm using the training set, update the model parameters of the volt-second characteristic prediction model, obtain the updated volt-second characteristic prediction model, and increment the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters.

[0137] The execution unit 16 is used to verify the volt-second characteristic prediction model through the test set when the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, update the training set to obtain the updated training set, and return to the execution unit for updating the model parameters.

[0138] The final model determination unit 17 is used to determine the latest volt-second characteristic prediction model as the final volt-second characteristic prediction model when the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates.

[0139] Optionally, the model parameter update unit includes:

[0140] The training subset partitioning unit is used to drive the latest volt-second characteristic prediction model to divide the training set into multiple training subsets through a resampling algorithm.

[0141] The decision tree modeling unit is used to perform decision tree modeling on each training subset to obtain the decision tree of the training subset.

[0142] The subset result prediction unit is used to determine the prediction result of each decision tree for its corresponding training subset;

[0143] The voting unit is used to vote on the prediction results of each decision tree to determine the final prediction result of the volt-second characteristic prediction model.

[0144] The update unit is used to update the model parameters of the volt-second characteristic prediction model based on the prediction results of each decision tree and the final prediction result, so as to obtain the updated volt-second characteristic prediction model.

[0145] Optionally, the device may also include:

[0146] The average value calculation unit is used to calculate the average value of the prediction results of each decision tree to obtain the final prediction result of the volt-second characteristic prediction model.

[0147] Optionally, the sample set determination unit includes:

[0148] The volt-second characteristic test unit is used to perform several volt-second characteristic tests on the insulator under each historical operating condition parameter to obtain multiple sets of volt-second characteristic test data. Each set of volt-second characteristic test data includes breakdown time and breakdown voltage.

[0149] The sample data acquisition unit is used to preprocess and normalize the test data of each set of volt-second characteristics for each historical operating condition parameter to obtain the sample data of that historical operating condition parameter.

[0150] The sample set data partitioning unit is used to construct a sample dataset based on sample data of various historical operating condition parameters, and to divide the sample dataset into a training set and a test set.

[0151] Optionally, the model output unit includes:

[0152] The intermediate prediction value output unit is used to input the operating condition parameters into a pre-established volt-second characteristic prediction model and drive the volt-second characteristic prediction model to output intermediate prediction values ​​at each preset breakdown voltage.

[0153] The inverse normalization unit is used to inverse normalize the intermediate predicted value output at each preset breakdown voltage to obtain the predicted breakdown time at that preset breakdown voltage.

[0154] Optionally, the device may also include:

[0155] The volt-second characteristic scatter point determination unit is used to determine the volt-second characteristic scatter point of the preset breakdown voltage based on each preset breakdown voltage and the predicted breakdown time corresponding to the preset breakdown voltage after inputting the operating condition parameters into the pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator under each preset breakdown voltage.

[0156] The curve fitting unit is used to perform curve fitting on the volt-second characteristic scatter points of each preset breakdown voltage to obtain the volt-second characteristic curve.

[0157] The device for predicting the volt-second characteristic parameters of insulators provided in this application embodiment can be applied to devices for predicting the volt-second characteristic parameters of insulators, such as terminals: mobile phones, computers, etc. Optionally, Figure 4 The hardware block diagram of the device for predicting the volt-second characteristic parameters of an insulator is shown. (Refer to...) Figure 4 The hardware structure of the device for predicting the volt-second characteristic parameters of insulators may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4.

[0158] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0159] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0160] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0161] The memory stores a program, which the processor can call. The program is used for:

[0162] Obtain the operating parameters of the insulator;

[0163] The operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator under each preset breakdown voltage is output.

[0164] The process of establishing the volt-second characteristic prediction model includes:

[0165] Obtain various historical operating condition parameters of the insulator when it is broken down;

[0166] Volt-second characteristic tests were conducted under each historical operating condition parameter, and the training set and test set were determined based on the test results under various historical operating condition parameters.

[0167] For the latest volt-second characteristic prediction model, based on the pre-set random forest algorithm, the volt-second characteristic prediction model is trained using the training set, the model parameters of the volt-second characteristic prediction model are updated, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters.

[0168] When the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, the volt-second characteristic prediction model is verified through the test set, the training set is updated to obtain the updated training set, and the process of training the volt-second characteristic prediction model based on the pre-set random forest algorithm through the training set for the latest volt-second characteristic prediction model is returned, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1.

[0169] When the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates, the latest volt-second characteristic prediction model is determined as the final volt-second characteristic prediction model.

[0170] Optionally, the refined and extended functions of the program can be found in the description above.

[0171] This application embodiment also provides a storage medium that can store a program suitable for execution by a processor, the program being used for:

[0172] Obtain the operating parameters of the insulator;

[0173] The operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator under each preset breakdown voltage is output.

[0174] The process of establishing the volt-second characteristic prediction model includes:

[0175] Obtain various historical operating condition parameters of the insulator when it is broken down;

[0176] Volt-second characteristic tests were conducted under each historical operating condition parameter, and the training set and test set were determined based on the test results under various historical operating condition parameters.

[0177] For the latest volt-second characteristic prediction model, based on the pre-set random forest algorithm, the volt-second characteristic prediction model is trained using the training set, the model parameters of the volt-second characteristic prediction model are updated, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters.

[0178] When the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, the volt-second characteristic prediction model is verified through the test set, the training set is updated to obtain the updated training set, and the process of training the volt-second characteristic prediction model based on the pre-set random forest algorithm through the training set for the latest volt-second characteristic prediction model is returned, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1.

[0179] When the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates, the latest volt-second characteristic prediction model is determined as the final volt-second characteristic prediction model.

[0180] Optionally, the refined and extended functions of the program can be found in the description above.

[0181] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0182] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0183] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting the volt-second characteristic parameters of an insulator, characterized in that, include: Obtain the operating parameters of the insulator; The operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator under each preset breakdown voltage is output. The process of establishing the volt-second characteristic prediction model includes: Obtain various historical operating condition parameters of the insulator when it is broken down; Volt-second characteristic tests were conducted under each historical operating condition parameter, and the training set and test set were determined based on the test results under various historical operating condition parameters. For the latest volt-second characteristic prediction model, based on the pre-set random forest algorithm, the volt-second characteristic prediction model is trained using the training set, the model parameters of the volt-second characteristic prediction model are updated, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters. When the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, the volt-second characteristic prediction model is verified through the test set, the training set is updated to obtain the updated training set, and the process of training the volt-second characteristic prediction model based on the pre-set random forest algorithm through the training set for the latest volt-second characteristic prediction model is returned, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model, and the number of times the model parameters of the volt-second characteristic prediction model are updated is incremented by 1. When the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates, the latest volt-second characteristic prediction model is determined as the final volt-second characteristic prediction model.

2. The method according to claim 1, characterized in that, The method for predicting the latest volt-second characteristics, based on a pre-set random forest algorithm, trains the volt-second characteristic prediction model using the training set, updates the model parameters, and obtains the updated volt-second characteristic prediction model, including: The latest volt-second characteristic prediction model uses a resampling algorithm to divide the training set into multiple training subsets; For each training subset, a decision tree model is performed on the training subset to obtain the decision tree for the training subset; Determine the prediction result of each decision tree for its corresponding training subset; The prediction results of each decision tree are voted on to determine the final prediction result of the volt-second characteristic prediction model; Based on the prediction results of each decision tree and the final prediction result, the model parameters of the volt-second characteristic prediction model are updated to obtain the updated volt-second characteristic prediction model.

3. The method according to claim 2, characterized in that, Also includes: The average of the prediction results of each decision tree is calculated to obtain the final prediction result of the volt-second characteristic prediction model.

4. The method according to claim 1, characterized in that, The process involves conducting volt-second characteristic tests under each historical operating condition parameter, and determining the training and test sets based on the test results under various historical operating condition parameters, including: Under each historical operating condition parameter, the insulator was subjected to several volt-second characteristic tests to obtain multiple sets of volt-second characteristic test data. Each set of volt-second characteristic test data included breakdown time and breakdown voltage. The test data of each set of volt-second characteristics for each historical operating condition parameter were preprocessed and normalized to obtain the sample data of that historical operating condition parameter. Based on sample data of various historical operating conditions, a sample dataset is constructed, and the sample dataset is divided into a training set and a test set.

5. The method according to claim 1, characterized in that, The operating parameters are input into a pre-established volt-second characteristic prediction model, and the predicted breakdown time of the insulator at each preset breakdown voltage is output, including: The operating parameters are input into a pre-established volt-second characteristic prediction model, which drives the volt-second characteristic prediction model to output intermediate prediction values ​​at each preset breakdown voltage. The intermediate predicted value output at each preset breakdown voltage is inversely normalized to obtain the predicted breakdown time at that preset breakdown voltage.

6. The method according to claim 1, characterized in that, After inputting the operating parameters into a pre-established volt-second characteristic prediction model and outputting the predicted breakdown time of the insulator at each preset breakdown voltage, the method further includes: Based on each preset breakdown voltage and the predicted breakdown time corresponding to that preset breakdown voltage, the volt-second characteristic scatter plot of that preset breakdown voltage is determined; Curve fitting is performed on the volt-second characteristic scatter points of each preset breakdown voltage to obtain the volt-second characteristic curve.

7. A device for predicting the volt-second characteristic parameters of an insulator, characterized in that, include: Operating condition parameter acquisition unit, used to acquire the operating condition parameters of the insulator; The model output unit is used to input the operating condition parameters into a pre-established volt-second characteristic prediction model and output the predicted breakdown time of the insulator under each preset breakdown voltage. The historical parameter acquisition unit is used to acquire various historical operating condition parameters of the insulator when it is broken down. The sample set determination unit is used to conduct volt-second characteristic tests under each historical operating condition parameter, and to determine the training set and test set based on the volt-second characteristic test results under various historical operating condition parameters. The model parameter update unit is used to train the latest volt-second characteristic prediction model using the training set based on a pre-set random forest algorithm, update the model parameters of the volt-second characteristic prediction model, obtain the updated volt-second characteristic prediction model, and increment the number of times the model parameters of the volt-second characteristic prediction model have been updated by 1. The initial version of the volt-second characteristic prediction model is a volt-second characteristic prediction model with pre-set default model parameters. The return execution unit is used to verify the volt-second characteristic prediction model through the test set when the number of times the model parameters of the volt-second characteristic prediction model are updated has not reached the preset number of updates, update the training set to obtain the updated training set, and return to the execution unit for updating the model parameters. The final model determination unit is used to determine the latest volt-second characteristic prediction model as the final volt-second characteristic prediction model when the number of times the model parameters of the volt-second characteristic prediction model are updated reaches the preset number of updates.

8. The apparatus according to claim 7, characterized in that, The model parameter update unit includes: The training subset partitioning unit is used to drive the latest volt-second characteristic prediction model to divide the training set into multiple training subsets through a resampling algorithm. The decision tree modeling unit is used to perform decision tree modeling on each training subset to obtain the decision tree of the training subset. The subset result prediction unit is used to determine the prediction result of each decision tree for its corresponding training subset; The voting unit is used to vote on the prediction results of each decision tree to determine the final prediction result of the volt-second characteristic prediction model. The update unit is used to update the model parameters of the volt-second characteristic prediction model based on the prediction results of each decision tree and the final prediction result, so as to obtain the updated volt-second characteristic prediction model.

9. The apparatus according to claim 8, characterized in that, Also includes: The average value calculation unit is used to calculate the average value of the prediction results of each decision tree to obtain the final prediction result of the volt-second characteristic prediction model.

10. The apparatus according to claim 7, characterized in that, The sample set determination unit includes: The volt-second characteristic test unit is used to perform several volt-second characteristic tests on the insulator under each historical operating condition parameter to obtain multiple sets of volt-second characteristic test data. Each set of volt-second characteristic test data includes breakdown time and breakdown voltage. The sample data acquisition unit is used to preprocess and normalize the test data of each set of volt-second characteristics for each historical operating condition parameter to obtain the sample data of that historical operating condition parameter. The sample set data partitioning unit is used to construct a sample dataset based on sample data of various historical operating condition parameters, and to divide the sample dataset into a training set and a test set.

Images