Power transmission line icing thickness driving factor mining method based on interpretable machine learning

Abstract

The invention relates to a power transmission line icing thickness driving factor mining method based on interpretable machine learning, and relates to the technical field of power system maintenance. The method comprises the following steps: acquiring historical meteorological data and historical icing data corresponding to a power transmission line, wherein the historical meteorological data comprises meteorological data corresponding to each driving factor; the driving factors are meteorological factors which influence the icing thickness of the power transmission line; determining a correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data, and determining feature data based on the correlation coefficient; according to meteorological data corresponding to the feature data, training to obtain at least one icing thickness prediction model; determining contribution degrees corresponding to the driving factors based on an icing thickness prediction model; the contribution degree is determined based on an interaction effect between the driving factors; and determining a driving factor combination based on the contribution degree. By adopting the method, the accuracy of determining the icing thickness driving factor can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of power system maintenance, and in particular, to a method, device, computer device, storage medium, and computer program product for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning. Background Art

[0002] The power system is an indispensable infrastructure in modern society. Among them, overhead transmission lines are an important part of the power system. The phenomenon of ice coating on transmission lines is a natural disaster faced by the power system in high-cold and high-humidity regions in winter. This natural disaster can lead to major accidents such as wire breakage, tower collapse, and power grid paralysis, seriously threatening the safe operation of the power system.

[0003] Currently, due to the complexity of the ice formation mechanism and the diversity of influencing factors, the source of features of the current ice coating thickness prediction method is unclear, seriously affecting the accuracy and timeliness of ice coating disaster emergency decision-making. Related technologies are difficult to capture complex interaction effects and are prone to ignoring the collinearity problem among high-dimensional features, resulting in the omission or misjudgment of key driving factors. And related technologies rely on linear assumptions, unable to effectively identify non-linear driving factors, and difficult to balance the contradiction between feature redundancy and information loss, resulting in insufficient accuracy of the extracted driving factors of ice coating thickness. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning, which can improve the accuracy of driving factors.

[0005] In a first aspect, the present application provides a method for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning. The method includes:

[0006] Obtain historical meteorological data and historical ice coating data corresponding to the transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factor is a meteorological element that affects the ice coating thickness of the transmission line;

[0007] Based on the historical meteorological data and the historical ice coating data, determine the correlation coefficient corresponding to the driving factor, and determine the feature data based on the correlation coefficient;

[0008] Train at least one ice coating thickness prediction model according to the meteorological data corresponding to the feature data;

[0009] Based on the ice coating thickness prediction model, determine the contribution degree corresponding to each driving factor; the contribution degree is determined based on the interaction effect between the driving factors;

[0010] Determine a combination of driving factors based on the contribution degree, and the combination of driving factors is used to determine a disaster prevention decision for icing disasters of the transmission line.

[0011] In one embodiment, the determining the correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data includes:

[0012] For the meteorological data corresponding to each driving factor, determine the correlation coefficient between the meteorological data corresponding to each driving factor and the historical icing data based on the Pearson correlation coefficient, and determine it as the correlation coefficient corresponding to each driving factor.

[0013] In one embodiment, the determining the characteristic data based on the correlation coefficient includes:

[0014] For each driving factor, when the correlation coefficient corresponding to the driving factor is greater than a preset coefficient value, determine the meteorological data corresponding to the driving factor;

[0015] Based on the meteorological data corresponding to the driving factor and the correlation coefficient corresponding to the driving factor, determine the characteristic data corresponding to the driving factor.

[0016] In one embodiment, the training at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data includes:

[0017] Divide the historical meteorological data and the historical icing data into a training set and a test set;

[0018] Based on the training set and the test set, train at least one pre-configured initial prediction model respectively to obtain at least one trained initial prediction model;

[0019] Determine the model accuracy index of the at least one trained initial prediction model, and based on the model accuracy index, determine at least one icing thickness prediction model from the at least one trained initial prediction model.

[0020] In one embodiment, the determining the contribution degree corresponding to each driving factor based on the icing thickness prediction model includes:

[0021] Based on the icing thickness prediction model and the test set, determine the contribution value of the driving factor and the total contribution value of each driving factor;

[0022] Based on the contribution values of each driving factor, determine the interaction effect between different driving factors and the main effect of a single driving factor;

[0023] For each of the said driving factors, based on the contribution value, total contribution value, main effect and multiple interaction effects of the driving factor, determine the contribution degree corresponding to the driving factor.

[0024] In one embodiment, the determining the combination of driving factors based on the contribution degree includes:

[0025] Sort the driving factors corresponding to each of the contribution degrees in descending order of the contribution degree to obtain a driving factor sequence;

[0026] In the driving factor sequence, determine a preset number of driving factors as the driving factor combination.

[0027] In one embodiment, the obtaining the historical meteorological data and historical icing data corresponding to the transmission line includes:

[0028] Obtain the meteorological data corresponding to each driving factor, and obtain the historical icing data, where the driving factors are various elements affecting weather phenomena;

[0029] For the meteorological data corresponding to each of the driving factors, and the obtained historical icing data, delete local abnormal data points to obtain the cleaned historical meteorological data and historical icing data;

[0030] Align the time of the cleaned historical meteorological data and historical icing data to obtain the aligned historical meteorological data and historical icing data.

[0031] In a second aspect, the present application also provides a device for mining driving factors of transmission line icing thickness. The device includes:

[0032] A data acquisition module, configured to acquire historical meteorological data and historical icing data corresponding to a transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factors are meteorological elements that affect the icing thickness of the transmission line;

[0033] A feature determination module, configured to determine the correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data, and determine the feature data based on the correlation coefficient;

[0034] A model training module, configured to train at least one icing thickness prediction model according to the meteorological data corresponding to the feature data;

[0035] A contribution degree determination module, configured to determine the contribution degree corresponding to each of the driving factors based on the icing thickness prediction model; the contribution degree is determined based on the interaction effect between the driving factors;

[0036] A driving factor determination module determines a driving factor combination based on the contribution degree, and the driving factor combination is used to determine a disaster prevention decision for the icing disaster of the transmission line.

[0037] In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0038] Obtain historical meteorological data and historical icing data corresponding to the transmission line. The historical meteorological data includes meteorological data corresponding to each driving factor. The driving factor is a meteorological element that affects the icing thickness of the transmission line;

[0039] Based on the historical meteorological data and the historical icing data, determine the correlation coefficient corresponding to the driving factor, and determine the characteristic data based on the correlation coefficient;

[0040] Train at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data;

[0041] Based on the icing thickness prediction model, determine the contribution degree corresponding to each driving factor. The contribution degree is determined based on the interaction effect between the driving factors;

[0042] Determine a driving factor combination based on the contribution degree, and the driving factor combination is used to determine a disaster prevention decision for the icing disaster of the transmission line.

[0043] In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0044] Obtain historical meteorological data and historical icing data corresponding to the transmission line. The historical meteorological data includes meteorological data corresponding to each driving factor. The driving factor is a meteorological element that affects the icing thickness of the transmission line;

[0045] Based on the historical meteorological data and the historical icing data, determine the correlation coefficient corresponding to the driving factor, and determine the characteristic data based on the correlation coefficient;

[0046] Train at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data;

[0047] Based on the icing thickness prediction model, determine the contribution degree corresponding to each driving factor. The contribution degree is determined based on the interaction effect between the driving factors;

[0048] Determine a combination of driving factors based on the contribution degree, and the combination of driving factors is used to determine a disaster prevention decision for icing disasters of the transmission line.

[0049] In a fifth aspect, the present application also provides a computer program product, including a computer program, which when executed by a processor implements the following steps:

[0050] Obtain historical meteorological data and historical icing data corresponding to the transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factor is a meteorological element that affects the icing thickness of the transmission line;

[0051] Based on the historical meteorological data and the historical icing data, determine the correlation coefficient corresponding to the driving factor, and determine the characteristic data based on the correlation coefficient;

[0052] Train at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data;

[0053] Based on the icing thickness prediction model, determine the contribution degree corresponding to each driving factor; the contribution degree is determined based on the interaction effect between the driving factors;

[0054] Determine a combination of driving factors based on the contribution degree, and the combination of driving factors is used to determine a disaster prevention decision for icing disasters of the transmission line.

[0055] The above method, device, computer device, storage medium, and computer program product for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning obtain historical meteorological data and historical ice coating data corresponding to the transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor. Based on the historical meteorological data and the historical ice coating data, the correlation coefficient corresponding to each driving factor is determined, and the feature data for training the ice coating thickness prediction model is determined based on the correlation coefficient. Then, at least one ice coating thickness prediction model is trained according to the meteorological data corresponding to the feature data. Finally, based on the ice coating thickness prediction model, the contribution degree corresponding to each driving factor is determined, and the driving factor combination is determined based on the contribution degree, so as to obtain various driving factors that mainly affect the ice coating thickness, and the disaster prevention decision for ice coating disasters on the transmission line can be determined according to the driving factor combination. Based on this, the driving factors in the historical meteorological data can be mined for the first time to obtain the correlation coefficient between each driving factor and the historical ice coating data, and the feature data for training is determined according to the correlation coefficient, and at least one ice coating thickness prediction model is trained. Through the prediction process of the ice coating thickness prediction model, the second mining is carried out to mine the contribution degree determined by the interaction effect between the driving factors, and the driving factor combination is determined based on the contribution degree, which can effectively analyze the influence of the driving factors on the ice coating thickness and the influence of the interaction between the driving factors on the ice coating thickness, thereby improving the accuracy of extracting the driving factors of the ice coating thickness. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0057] Figure 1 It is an application environment diagram of a method for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning in an embodiment;

[0058] Figure 2 It is a flowchart of a method for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning in an embodiment;

[0059] Figure 3 It is a flowchart of the steps for determining the contribution degree of driving factors in an embodiment;

[0060] Figure 4 It is a flowchart of a method for mining driving factors of ice coating thickness on transmission lines based on interpretable machine learning in another embodiment;

[0061] Figure 5 The structural block diagram of the driving factor mining device for the icing thickness of a transmission line in an embodiment;

[0062] Figure 6 The internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0063] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0064] The driving factor mining method for the icing thickness of a transmission line based on interpretable machine learning provided by the embodiments of the present application can be applied to an application environment as Figure 1 shown. Among them, the terminal 102 communicates with the server 104 through a network. The terminal 102 can be a sensing device set on the transmission line and used to record the icing data in winter, and the terminal 102 can be a meteorological monitoring device used to monitor the meteorological data of the area where the transmission line is located. The meteorological data and icing data in the same area can be sent to the server 104 as a set of data through the communication network. The meteorological data in the past period (such as several years) can be determined as historical meteorological data, and the icing data in the past period can be determined as historical icing data. The server 104 can package the meteorological data and icing data corresponding to the transmission line in the same area as the data for determining the driving factors.

[0065] Based on this, the server 104 can determine the correlation coefficient between various driving factors corresponding to the historical meteorological data and the historical icing data according to the historical meteorological data and the historical icing data, and determine the feature data from the historical meteorological data by the correlation coefficient. The server 104 can use the meteorological data corresponding to the feature data as training data to train the classification model to obtain at least one icing thickness prediction model. The server 104 uses the trained icing thickness model for prediction, and analyzes the contribution degree corresponding to each driving factor based on the trained icing thickness model and the test set, so as to determine the driving factor combination based on the contribution degree.

[0066] The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed on the cloud or other network servers. Among them, the terminal 102 can be, but is not limited to, various meteorological monitoring devices, such as temperature, humidity, pressure, wind speed, wind direction monitoring devices, and icing thickness monitoring devices. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0067] In an exemplary embodiment, asFigure 2 As shown, a method for mining the driving factors of ice coating thickness on transmission lines based on interpretable machine learning is provided. Taking the server 104 in Figure 1 as an example, the method includes the following steps S202 to S208. Among them:

[0068] Step S202: Obtain the historical meteorological data and historical ice coating data corresponding to the transmission line. The historical meteorological data includes the meteorological data corresponding to each driving factor.

[0069] Among them, the meteorological data can be time series data of variables such as temperature, humidity, wind speed, and air pressure. The time span of the data should be as long as possible to cover different seasons and climate conditions. For example, daily data for the past 5 to 10 years can be collected. The ice coating data can be the ice coating thickness data of the transmission line within the corresponding time period obtained through monitoring devices installed on the transmission line. The driving factor can also be called a driving factor or influencing factor, which refers to various elements or conditions that can have a promoting, influencing, or determining effect on a certain system, phenomenon, process, or result. The driving factors in the embodiments of this application are various meteorological elements that affect the ice coating thickness of the transmission line. The meteorological elements can include temperature, humidity, wind speed, air pressure, etc.

[0070] Specifically, the server can receive the historical meteorological data composed of time series data of variables such as temperature, humidity, wind speed, and air pressure sent by each terminal. The server can receive the historical ice coating data of the transmission line within the corresponding time sent by each terminal. For example, the time series data of temperature can be the temperature range for each day in the past year, and the historical ice coating data can be the maximum ice coating thickness for each day in the past year.

[0071] Step S204: Based on the historical meteorological data and the historical ice coating data, determine the correlation coefficient corresponding to the driving factor, and determine the feature data based on the correlation coefficient.

[0072] Among them, the correlation coefficient represents the linear correlation degree between each driving factor and the ice coating thickness respectively. The value range of the correlation coefficient can be [-1, 1]. The feature data is obtained by screening the historical meteorological data according to the absolute value size of each driving factor and generating features for the screened meteorological data.

[0073] Specifically, for the meteorological data corresponding to each driving factor, the server can determine the correlation coefficient between the driving factor and the historical icing data based on the observed value of the meteorological data and the observed value of the corresponding historical icing data. Based on this, the server can obtain the correlation coefficients between each driving factor and the historical icing data respectively. The server can screen out part of the historical meteorological data from the historical meteorological data according to the correlation coefficients between each driving factor and the historical icing data respectively, and determine the characteristic data corresponding to the historical meteorological data.

[0074] Step S206: Train at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data.

[0075] Among them, the icing thickness prediction model can be an artificial intelligence model with different model structures. The icing thickness prediction model can be a common regression model or classification model such as a recurrent neural network, a convolutional neural network, a decision tree, a support vector machine, etc.

[0076] Specifically, the server can pre-construct multiple icing thickness prediction models of different types. For each icing thickness prediction model, the server can obtain the same meteorological data, and generate the characteristic data corresponding to the meteorological data according to the data structure of the input data that each model can accept, so as to obtain the characteristic data corresponding to each icing thickness prediction model. Based on this, the server can train the icing thickness prediction model through the characteristic data corresponding to each icing thickness prediction model to obtain at least one trained icing thickness prediction model.

[0077] In addition, the server can pre-construct an icing thickness prediction model of one type. For this icing thickness prediction model, the server can determine the meteorological data for training based on the correlation coefficient, and process the meteorological data according to at least one feature generation method to obtain at least one type of characteristic data. The server can train the icing thickness prediction model with each type of characteristic data to obtain the trained icing thickness prediction model corresponding to each type of characteristic data. That is, use the same type of icing thickness prediction model, but train it with different characteristic data respectively to obtain multiple trained icing thickness prediction models.

[0078] For example, the server can determine the meteorological data Data and determine the corresponding characteristic data X of Data. The server can train different types of icing thickness prediction models A, B, and C through the characteristic data X to obtain the trained icing thickness prediction models A, B, and C. Additionally, the server can determine the meteorological data Data and determine the corresponding characteristic data X1, X2, and X3 of Data. The server uses the characteristic data to train the same type of icing thickness prediction model A respectively to obtain the trained icing thickness prediction models A1, A2, and A3.

[0079] Step S208: Based on the ice thickness prediction model, determine the contribution degree corresponding to each driving factor.

[0080] Among them, the contribution degree is determined based on the interaction effect between driving factors. The interaction effect refers to the combined influence effect generated when two or more independent variables act on a dependent variable at the same time. In the embodiments of the present application, specifically, it refers to the combined influence effect generated when two or more driving factors act on the ice thickness.

[0081] Specifically, the server can perform prediction according to the trained ice thickness prediction model to obtain the prediction result, and analyze the ice thickness prediction model, the prediction result and the actual result to determine the contribution value of the characteristic data corresponding to each driving factor, and based on the contribution value of each characteristic data, determine the interaction effect between driving factors. For a single driving factor, the server can determine the contribution degree corresponding to the driving factor based on the interaction effect corresponding to the driving factor.

[0082] Step S210: Determine the driving factor combination based on the contribution degree.

[0083] Specifically, the driving factor combination is used to determine the disaster prevention decision for the ice disaster of the transmission line. The server can determine multiple driving factors with contribution degrees greater than the preset value as the driving factor combination. The server can send the driving factor combination and the contribution degrees corresponding to each driving factor in the driving factor combination to the user terminal, so that the user terminal can determine the disaster prevention decision for the ice disaster of the transmission line according to the driving factor combination.

[0084] In the above method for mining the driving factors of ice coating thickness on transmission lines based on interpretable machine learning, historical meteorological data and historical ice coating data corresponding to the transmission lines are obtained, and the historical meteorological data includes meteorological data corresponding to each driving factor. Based on the historical meteorological data and the historical ice coating data, the correlation coefficients corresponding to each driving factor are determined, and the characteristic data for training the ice coating thickness prediction model is determined based on the correlation coefficients. Then, at least one ice coating thickness prediction model is trained according to the meteorological data corresponding to the characteristic data. Finally, based on the ice coating thickness prediction model, the contribution degrees corresponding to each driving factor are determined, and the driving factor combination is determined based on the contribution degrees, so as to obtain various driving factors that mainly affect the ice coating thickness, and the disaster prevention decision for ice coating disasters on the transmission line can be determined according to the driving factor combination. Based on this, the driving factors in the historical meteorological data can be mined for the first time to obtain the correlation coefficients between each driving factor and the historical ice coating data, and the characteristic data for training is determined according to the correlation coefficients, and at least one ice coating thickness prediction model is trained. Through the prediction process of the ice coating thickness prediction model, the second mining is carried out to mine the contribution degrees determined by the interaction effects between the driving factors, and the driving factor combination is determined based on the contribution degrees, which can effectively analyze the influence of the driving factors on the ice coating thickness and the influence of the interaction between the driving factors on the ice coating thickness, so as to improve the accuracy of extracting the driving factors of the ice coating thickness.

[0085] In an exemplary embodiment, the specific implementation process of the step of "determining the correlation coefficients corresponding to the driving factors based on the historical meteorological data and the historical ice coating data" includes:

[0086] For the meteorological data corresponding to each driving factor, the correlation coefficients between the meteorological data corresponding to each driving factor and the historical ice coating data are determined based on the Pearson correlation coefficient, and are determined as the correlation coefficients corresponding to each driving factor.

[0087] Among them, the Pearson correlation coefficient is a statistical index used to measure the linear correlation degree between two variables.

[0088] Specifically, the server can traverse each driving factor to determine multiple groups of observed values composed of the meteorological data corresponding to the driving factor and the historical ice coating data. The server can determine the linear correlation relationship between the meteorological data and the historical ice coating data through the calculation strategy of the Pearson correlation coefficient. In one example, the formula of the Pearson correlation coefficient is as follows:

[0089]

[0090] Among them, let X be the meteorological data and Y be the historical ice coating data. X and Y can form n groups of observed values (x1, y1), (x2, y2)......(x n , y n ), and are the means of meteorological data and historical icing data respectively. The value range of the Pearson correlation coefficient r is [-1, 1]. When 0 < r < 1, it indicates that there is a positive linear correlation between the two variables, and the closer r is to 1, the stronger the linear correlation; when -1 < r < 0, it indicates that there is a negative linear correlation between the two variables, and the closer r is to -1, the stronger the negative linear correlation.

[0091] In one example, the server can determine the correlation coefficient between temperature and icing thickness, the correlation coefficient between humidity and icing thickness, the correlation coefficient between pressure and icing thickness, etc. through the Pearson correlation coefficient.

[0092] In this embodiment, the correlation coefficient between the meteorological data corresponding to each driving factor and the historical icing data through the Pearson correlation coefficient can improve the accuracy and efficiency of determining the correlation coefficient.

[0093] In an exemplary embodiment, the specific implementation process of the step "determine feature data based on the correlation coefficient" includes:

[0094] For each driving factor, when the correlation coefficient corresponding to the driving factor is greater than the preset coefficient value, determine the meteorological data corresponding to the driving factor. Based on the meteorological data corresponding to the driving factor and the correlation coefficient corresponding to the driving factor, determine the feature data corresponding to the driving factor.

[0095] Specifically, the server can traverse the correlation coefficients of each driving factor. The server can take the absolute value of each correlation coefficient and determine whether the absolute value is greater than the preset coefficient value. If the absolute value is less than the preset coefficient value, discard the driving factor corresponding to the absolute value. If the absolute value is greater than the preset coefficient value, determine the driving factor corresponding to the absolute value. When the server determines multiple driving factors with absolute values greater than the preset coefficient value, it can obtain the meteorological data of the above driving factors.

[0096] The server can determine the weight of each driving factor according to the correlation coefficient of the above driving factor, and determine the initial feature data of the meteorological data of each driving factor according to the preset feature determination strategy, and configure the above initial feature data according to the weight to obtain the target feature data corresponding to the driving factor.

[0097] In this embodiment, by using the preset coefficient value to screen out some driving factors from multiple driving factors and processing the meteorological data corresponding to some driving factors based on the correlation coefficients corresponding to some driving factors to obtain the feature numbers corresponding to the driving factors, it can screen out driving factors with extremely low importance in advance and improve the accuracy of the feature data.

[0098] In an exemplary embodiment, the specific implementation process of the step "training at least one ice accretion thickness prediction model according to the meteorological data corresponding to the feature data" includes:

[0099] Dividing the historical meteorological data and historical ice accretion data into a training set and a test set; training at least one pre-configured initial prediction model based on the training set and the test set respectively to obtain at least one trained initial prediction model; determining the model accuracy metrics of at least one trained initial prediction model, and based on the model accuracy metrics, determining at least one ice accretion thickness prediction model from at least one trained initial prediction model.

[0100] Specifically, the server may merge the historical meteorological data and historical ice accretion data into a data set, and divide the data set into a training set and a test set according to a preset ratio. The server may train at least one pre-configured initial prediction model according to the training set and the test set to obtain at least one trained initial prediction model. For example, the server may generate the feature data corresponding to different initial prediction models according to the data structure of the feature data required by the initial prediction model, and train the initial prediction model based on the respective corresponding feature data to obtain at least one trained initial prediction model. For example, the server may generate multiple different types of feature data, each type of feature data corresponding to an initial prediction model, and the server may train the initial prediction model based on the feature data, and one type of feature data can train one initial prediction model.

[0101] The server may determine the model accuracy metrics of each initial prediction model according to calculation strategies such as the sum of mean squared errors, coefficient of determination, and mean absolute error. The server may determine at least one initial prediction model as the ice accretion thickness prediction model from multiple trained initial prediction models. For example, the server may determine that the model accuracy metrics of model A, model B, and model C are 90, 50, and 80 through the calculation strategy, and determine the two models with the highest model accuracy metrics among them as the ice accretion thickness prediction models, that is, determine model A and model C as the ice accretion thickness prediction models.

[0102] In an example, the model accuracy metric may be the average of the sum of mean squared errors, coefficient of determination, and mean absolute error. The model accuracy metric may be determined by a weighted sum of the sum of mean squared errors, coefficient of determination, and mean absolute error.

[0103] In this embodiment, by training at least one initial prediction model with a training set and a test set to obtain at least one trained initial prediction model, and determining the ice accretion thickness prediction model according to the model accuracy metrics of each initial prediction model, the optimal model can be selected from multiple models, improving the prediction accuracy of the ice accretion thickness prediction model.

[0104] In an exemplary embodiment, as Figure 3 shown, the step of "determining the contribution degree corresponding to each driving factor based on the ice thickness prediction model" includes steps S302 to S306. Among them:

[0105] Step S302: Based on the ice thickness prediction model and the test set, determine the contribution value of the driving factor and the total contribution value of each driving factor.

[0106] Among them, the contribution value can be the SHAP value determined based on SHAP (SHapley Additive exPlanations). SHAP is not a single algorithm in the traditional sense, but an interpretability framework for explaining the prediction results of machine learning models. Through SHAP, the contribution of the feature data corresponding to various driving factors to the ice thickness can be determined.

[0107] Specifically, the server can determine the contribution value of the driving factor corresponding to each feature data in the ice thickness prediction model based on the attribution analysis method of the SHAP framework. For example, the server can input the test set into the ice thickness prediction model, and through the SHAP framework, the prediction process of the ice thickness prediction model can be analyzed to obtain the contribution value corresponding to the feature data of each driving factor. Based on this, the server can determine the contribution value corresponding to the feature data of each driving factor, and the server adds up the contribution values corresponding to each driving factor to obtain the total contribution value of each driving factor.

[0108] Step S304: Based on the contribution value of each driving factor, determine the interaction effect between different driving factors and the main effect of a single driving factor.

[0109] Specifically, the server can process the contribution value of the driving factor through a preset effect determination function to obtain the interaction effect between different driving factors, and the server can only determine the main effect of a single driving factor.

[0110] In one example, the effect determination function is mainly used to show the interaction effects between feature data and the influence of main effects on the model output. In the scenario of ice accretion thickness, for two feature data x and y, assuming the effect determination function is shap.dependence_plot(x, shap_values, data), this function will plot a scatter plot with the value of feature x as the abscissa and the SHAP value of feature x as the ordinate. If there is an interaction effect, the scatter points will be color-coded or differentiated in other ways according to the value of feature y, so that the change in the influence of x on ice accretion thickness under different y values can be observed, thereby seeing the interaction between features. The main effect is the trend of the influence of each feature x on ice accretion thickness alone. Without considering other features, how the change of x causes the change of ice accretion thickness, and this change can be reflected by the SHAP value.

[0111] Step S306, for each driving factor, based on the contribution value, total contribution value, main effect, and multiple interaction effects of the driving factor, determine the contribution degree corresponding to the driving factor.

[0112] Specifically, the server can analyze the feature data corresponding to each type of driving factor. For example, the server can determine the contribution ratio of the contribution value corresponding to the feature data in the total contribution value. The server can perform statistical analysis based on the contribution ratio and the main effect to obtain the first contribution degree. The server can perform unified analysis based on the contribution ratio and multiple different interaction effects to obtain multiple second contribution degrees. The first contribution degree is used to represent the contribution of a single driving factor to ice accretion thickness. The second contribution degree is used to represent the contribution of a single driving factor combined with other driving factors to ice accretion thickness. After the server combines the first contribution degree and the second contribution degree, it obtains the contribution degree corresponding to the driving factor. The server can determine the contribution degree corresponding to each type of driving factor according to the above process.

[0113] In this embodiment, the contribution value of the driving factor and the total contribution value are determined through the ice accretion thickness prediction model and the test set. The interaction effect and the main effect can also be determined according to the contribution value of the driving factor, so as to further determine the first contribution degree corresponding to the main effect and the second contribution degree corresponding to the interaction effect, and obtain the contribution degree of the driving factor, which can improve the accuracy of determining the contribution degree.

[0114] In an exemplary embodiment, the specific implementation process of the step "determine the driving factor combination based on the contribution degree" includes:

[0115] Sort the driving factors corresponding to each contribution degree in the order of the size of the contribution degree to obtain a driving factor sequence; in the driving factor sequence, determine a preset number of driving factors as the driving factor combination.

[0116] Specifically, the server can sort the driving factors in descending order of contribution degree to obtain a driving factor sequence. For example, the driving factor sequence can be (temperature: 0.4, humidity: 0.3, wind speed: 0.2, air pressure: 0.15). The server can determine a preset number of driving factors from the driving factor sequence and combine the driving factors to obtain a driving factor combination that affects the main icing thickness.

[0117] In this embodiment, by sorting the driving factors according to the contribution degree by the server to obtain a driving factor sequence and determining a driving factor combination in the driving factor sequence, the accuracy and rationality of the driving factor combination can be improved.

[0118] In an exemplary embodiment, the specific implementation process of the step "obtain historical meteorological data and historical icing data corresponding to the transmission line" includes:

[0119] Obtain the meteorological data corresponding to each driving factor and obtain historical icing data, where the driving factors are various elements affecting weather phenomena; for the meteorological data corresponding to each driving factor and the obtained historical icing data, delete local abnormal data points to obtain the cleaned historical meteorological data and historical icing data; perform alignment processing on the time of the cleaned historical meteorological data and historical icing data to obtain the aligned historical meteorological data and historical icing data.

[0120] Specifically, the server can collect data: winter monitoring data such as meteorological data (including: temperature, humidity, air pressure, wind direction, wind speed, etc.) and icing thickness data (including: icing thickness, icing time, icing duration, etc.). The server can process the meteorological data corresponding to each driving factor, determine missing values and abnormal values, and process the missing values and abnormal values. For example, the server can fill the missing values and abnormal values with the average value, and the server can determine local abnormal data points containing abnormal values and delete the abnormal data points to obtain the cleaned historical meteorological data and historical icing data.

[0121] In one example, the server can use an improved local outlier factor (LOF) algorithm to identify abnormal data points and process the abnormal values of the abnormal data points. For data j, the determination criterion for abnormal values is:

[0122]

[0123] where and respectively represent the mean and standard deviation of the LOF values of the features. This method can effectively identify local outliers and improve the reliability of the data.

[0124] In addition, the server can align the time of the cleaned historical meteorological data and historical icing data to obtain the aligned historical meteorological data and historical icing data.

[0125] In one example, the server performs time alignment processing on the data to align the time of the icing thickness and meteorological data. For example, the server can perform data normalization processing to eliminate the influence of dimension on the subsequent mining of driving factors. An improved Min-Max normalization method can be selected to introduce a non-linear transformation to enhance the discrimination of features:

[0126]

[0127] where α is an adjustable non-linear factor (usually set between 0.1 and 10), and μ j and σj are the mean and standard deviation of feature j respectively. This non-linear normalization method can better handle features with different scales and distributions, and is particularly effective for features with long-tailed distributions.

[0128] In this embodiment, through data cleaning and data alignment, the historical meteorological data and historical icing data can be preprocessed to obtain the processed historical meteorological data and historical icing data, which can improve the reliability of the data.

[0129] As Figure 4 shown below, a specific embodiment is combined to describe in detail the specific execution process of the above-mentioned method for mining the driving factors of the icing thickness of transmission lines based on interpretable machine learning, including the following steps:

[0130] 1. Data collection and preprocessing module.

[0131] 1.1. Collect data: including meteorological data (temperature, humidity, air pressure, wind direction, wind speed, etc.), icing thickness data (obtained by the front-end tension sensor, including icing thickness, icing time, icing duration, etc.) and other winter monitoring data (time span from November of each year to February of the next year).

[0132] 1.2. Data cleaning: handling missing values and outliers. An improved Local Outlier Factor (LOF) algorithm is used to identify and remove abnormal data points. For data j, the determination criterion for outliers is:

[0133]

[0134] where and represent the mean and standard deviation of the LOF values of feature j respectively. The k-d tree algorithm can be used to accelerate the calculation process of LOF to handle large-scale data sets. It can be implemented through LocalOutlierFactor in sklearn.

[0135] 1.3. Data Formatting: Perform time consistency processing on the data to align the icing thickness and meteorological data in terms of time. Optional data normalization processing can be carried out to eliminate the influence of dimensions on subsequent mining of driving factors. An optional improved Min-Max normalization method can be introduced with a non-linear transformation to enhance the discrimination of features:

[0136]

[0137] where α is an adjustable non-linear factor (usually set between 0.1 and 10), and μ j and σj are the mean and standard deviation of feature j respectively. The non-linear normalization method can better handle features with different scales and distributions, especially for features with long-tailed distributions.

[0138] 2. Correlation Analysis Module: Apply the correlation analysis module to traverse the correlation coefficients between driving factors and icing thickness, and search for factors affecting the icing thickness to optimize the data combination. Specifically, in Python, call the corrcoef() function of numpy for calculation.

[0139] 3. Random Forest Module:

[0140] 3.1. Sample Setting: Divide the samples into a training set and a test set with a ratio of 8:2. Specifically, use the split_dataset method for sample division.

[0141] 3.2. Model Training: The Random Forest (RF) algorithm is an extended variant of Bagging in ensemble learning and is one of the currently widely used machine learning models, often used to solve classification and regression problems. The main principle of RF is to draw M samples from the original training set with replacement, where each sample has the same size as the original training set. Use Classification and Regression Tree (CART) to model each sample to obtain M modeling results, and combine the prediction results of each tree together to generate a representative value as the final prediction result. RF is easy to use and only requires adjusting one or two hyperparameters (i.e., the number of trees and the maximum tree depth). Model training includes hyperparameter tuning and training. Specifically, call the BayesianOptimization method of bayes_opt for hyperparameter tuning and call the fit method for model training.

[0142] 3.3. Model accuracy test: Three metrics are selected to evaluate the regression model. They are: Mean Squared Error (MSE), Coefficient of Determination (R 2 ), and Mean Absolute Error (MAE). The definitions of the above metrics are as follows:

[0143]

[0144] where: y represents the measured value of ice coating thickness, represents the predicted value of ice coating thickness, and n represents the total number of samples. represents the average value of the measured values of ice coating thickness, and i represents the i-th sample. Specifically, the predict method is used to predict the test set, and the r2_score, mean_squared_error, and mean_absolute_error methods in sklearn are combined to calculate the mean squared error, coefficient of determination, and mean absolute error. The optimal regression model is obtained.

[0145] 4. Interpretability module: Based on the attribution analysis method of SHAP theory in cooperative games, calculate the marginal contribution value of each driving factor to the ice coating thickness in the random forest model, and combine the SHAP additive interpretability model to quantify the contribution size of each driving factor. Specifically, call TreeExplainer in the shap library to calculate the shap value of the sample, use shap.simmary_plot to obtain the total contribution of each driving factor to the ice coating thickness, use shap.dependence_plot to obtain the interaction effect and main effect of the driving factors on the ice coating thickness, and use shap.plots.heatmap to obtain the influence of the driving factors in each sample on the ice coating thickness. Through the interpretability module, the main driving factor combination of the ice coating thickness is obtained.

[0146] In this embodiment, by combining data collection and processing, Pearson correlation analysis, random forest model, and SHAP value interpretation, it is possible to automatically calculate the interaction and contribution of driving factors to the ice coating thickness of transmission lines, and solve the problems that traditional methods fail to deeply explore the interaction effects, linear effects, etc. between driving factors. In addition, the embodiment of this application introduces an automatic parameter tuning method, enabling the model to automatically adjust various prediction model hyperparameters according to different data characteristics, so that the machine learning model can maintain high accuracy in different meteorological environments and expand the application scope of the method.

[0147] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indications of the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear statement in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0148] Based on the same inventive concept, an embodiment of the present application also provides a device for mining driving factors of transmission line icing thickness for implementing the above-mentioned method for mining driving factors of transmission line icing thickness based on interpretable machine learning. The implementation solution provided by this device to solve the problem is similar to the implementation solution described in the above method. Therefore, the specific limitations in one or more embodiments of the device for mining driving factors of transmission line icing thickness provided below can refer to the limitations on the method for mining driving factors of transmission line icing thickness based on interpretable machine learning in the above text, and will not be repeated here.

[0149] In an exemplary embodiment, as Figure 5 shown, a device 500 for mining driving factors of transmission line icing thickness is provided, including: a data acquisition module 501, a feature determination module 502, a model training module 503, a contribution degree determination module 504, and a driving factor determination module 505, where:

[0150] The data acquisition module 501 is used to acquire historical meteorological data and historical icing data corresponding to the transmission line, and the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factor is a meteorological element that affects the icing thickness of the transmission line;

[0151] The feature determination module 502 is used to determine the correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data, and determine the feature data based on the correlation coefficient;

[0152] The model training module 503 is used to train at least one icing thickness prediction model according to the meteorological data corresponding to the feature data;

[0153] The contribution degree determination module 504 is used to determine the contribution degree corresponding to each driving factor based on the icing thickness prediction model; the contribution degree is determined based on the interaction effect between the driving factors;

[0154] The driving factor determination module 505 determines a combination of driving factors based on the contribution degree, and the combination of driving factors is used to determine the disaster prevention decision for icing disasters of the transmission line.

[0155] Further, the feature determination module 502 is specifically configured to: for the meteorological data corresponding to each driving factor, determine the correlation coefficient between the meteorological data corresponding to each driving factor and the historical icing data based on the Pearson correlation coefficient, and determine it as the correlation coefficient corresponding to each driving factor.

[0156] Further, the feature determination module 502 is specifically further configured to: for each driving factor, when the correlation coefficient corresponding to the driving factor is greater than the preset coefficient value, determine the meteorological data corresponding to the driving factor; based on the meteorological data corresponding to the driving factor and the correlation coefficient corresponding to the driving factor, determine the feature data corresponding to the driving factor.

[0157] Further, the model training module 503 is specifically further configured to: divide the historical meteorological data and historical icing data into a training set and a test set; train at least one pre-configured initial prediction model based on the training set and the test set respectively to obtain at least one trained initial prediction model; determine the model accuracy index of at least one trained initial prediction model, and based on the model accuracy index, determine at least one icing thickness prediction model from at least one trained initial prediction model.

[0158] Further, the driving factor module 504 is specifically configured to: based on the icing thickness prediction model and the test set, determine the contribution value of the driving factor and the total contribution value of each driving factor; based on the contribution value of each driving factor, determine the interaction effect between different driving factors and the main effect of a single driving factor; for each driving factor, based on the contribution value, total contribution value, main effect and multiple interaction effects of the driving factor, determine the contribution degree corresponding to the driving factor.

[0159] Further, the driving factor module 504 is specifically further configured to: sort the driving factors corresponding to each contribution degree in descending order of the contribution degree to obtain a driving factor sequence; in the driving factor sequence, determine a preset number of driving factors as the combination of driving factors.

[0160] Further, the data acquisition module 502 is specifically configured to: acquire the meteorological data corresponding to each driving factor and acquire the historical icing data, where the driving factor is various elements affecting weather phenomena; for the meteorological data corresponding to each driving factor and the acquired historical icing data, delete local abnormal data points to obtain the cleaned historical meteorological data and historical icing data; perform alignment processing on the time of the cleaned historical meteorological data and historical icing data to obtain the aligned historical meteorological data and historical icing data.

[0161] Each module in the above-mentioned device for mining the driving factors of the ice thickness on the transmission line can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above modules can be embedded in the processor of the computer device in the form of hardware or be independent of it, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0162] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 6 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store historical meteorological data and historical ice-covering data. The input / output interface of the computer device is used for the processor to exchange information with external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it realizes a method for mining the driving factors of the ice thickness on the transmission line based on interpretable machine learning.

[0163] Those skilled in the art can understand that Figure 6 the structure shown in

[0164] is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0165] Obtain the historical meteorological data and historical ice-covering data corresponding to the transmission line. The historical meteorological data includes the meteorological data corresponding to each driving factor. The driving factor is a meteorological element that affects the ice thickness of the transmission line.

[0166] Based on the historical meteorological data and the historical ice-covering data, determine the correlation coefficient corresponding to the driving factor, and determine the feature data based on the correlation coefficient.

[0167] Training at least one ice thickness prediction model based on the meteorological data corresponding to the feature data;

[0168] Based on the ice thickness prediction model, determining the contribution degree corresponding to each of the driving factors; the contribution degree is determined based on the interaction effect between the driving factors;

[0169] Determining a combination of driving factors based on the contribution degree, the combination of driving factors being used to determine a disaster prevention decision for ice disasters of the transmission line.

[0170] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0171] Obtaining historical meteorological data and historical ice thickness data corresponding to a transmission line, the historical meteorological data including meteorological data corresponding to each driving factor; the driving factor being a meteorological element that affects the ice thickness of the transmission line;

[0172] Based on the historical meteorological data and the historical ice thickness data, determining the correlation coefficient corresponding to the driving factor, and determining feature data based on the correlation coefficient;

[0173] Training at least one ice thickness prediction model based on the meteorological data corresponding to the feature data;

[0174] Based on the ice thickness prediction model, determining the contribution degree corresponding to each of the driving factors; the contribution degree is determined based on the interaction effect between the driving factors;

[0175] Determining a combination of driving factors based on the contribution degree, the combination of driving factors being used to determine a disaster prevention decision for ice disasters of the transmission line.

[0176] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0177] Obtaining historical meteorological data and historical ice thickness data corresponding to a transmission line, the historical meteorological data including meteorological data corresponding to each driving factor; the driving factor being a meteorological element that affects the ice thickness of the transmission line;

[0178] Based on the historical meteorological data and the historical ice thickness data, determining the correlation coefficient corresponding to the driving factor, and determining feature data based on the correlation coefficient;

[0179] Training at least one ice thickness prediction model based on the meteorological data corresponding to the feature data;

[0180] Based on the ice thickness prediction model, determine the contribution degree corresponding to each of the driving factors; the contribution degree is determined based on the interaction effect between the driving factors;

[0181] Determine a combination of driving factors based on the contribution degree, and the combination of driving factors is used to determine a disaster prevention decision for the icing disaster of the transmission line.

[0182] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0183] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memories can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0184] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0185] The above-described embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for mining the driving factors of ice coating thickness on transmission lines based on interpretable machine learning, characterized in that The method includes: Obtaining historical meteorological data and historical icing data corresponding to a transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factor is a meteorological element that affects the icing thickness of the transmission line; Based on the historical meteorological data and the historical icing data, determining the correlation coefficient corresponding to the driving factor, and determining characteristic data based on the correlation coefficient; Training at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data; Based on the icing thickness prediction model, determining the contribution degree corresponding to each driving factor; the contribution degree is determined based on the interaction effect between the driving factors; Determining a driving factor combination based on the contribution degree, where the driving factor combination is used to determine the disaster prevention decision for the icing disaster of the transmission line.

2. The method according to claim 1, wherein The determining the correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data includes: For the meteorological data corresponding to each driving factor, determining the correlation coefficient between the meteorological data corresponding to each driving factor and the historical icing data based on the Pearson correlation coefficient, and determining it as the correlation coefficient corresponding to each driving factor.

3. The method according to claim 2, wherein The determining the characteristic data based on the correlation coefficient includes: For each driving factor, when the correlation coefficient corresponding to the driving factor is greater than a preset coefficient value, determining the meteorological data corresponding to the driving factor; Based on the meteorological data corresponding to the driving factor and the correlation coefficient corresponding to the driving factor, determining the characteristic data corresponding to the driving factor.

4. The method according to claim 1, characterized in that, The training at least one icing thickness prediction model according to the meteorological data corresponding to the characteristic data includes: Dividing the historical meteorological data and the historical icing data into a training set and a test set; Based on the training set and the test set, respectively training at least one pre-configured initial prediction model to obtain at least one trained initial prediction model; Determining the model accuracy index of the at least one trained initial prediction model, and based on the model accuracy index, determining at least one icing thickness prediction model from the at least one trained initial prediction model.

5. The method according to claim 4, wherein The determining the contribution degree corresponding to each driving factor based on the icing thickness prediction model includes: Based on the icing thickness prediction model and the test set, determining the contribution value of the driving factor and the total contribution value of each driving factor; Based on the contribution values of each driving factor, determining the interaction effect between different driving factors and the main effect of a single driving factor; For each driving factor, based on the contribution value, total contribution value, main effect and multiple interaction effects of the driving factor, determining the contribution degree corresponding to the driving factor.

6. The method according to claim 5, wherein The determining the driving factor combination based on the contribution degree includes: Sorting the driving factors corresponding to each contribution degree in descending order of the contribution degree to obtain a driving factor sequence; In the driving factor sequence, determining a preset number of driving factors as the driving factor combination.

7. The method according to claim 1, characterized in that, The obtaining the historical meteorological data and the historical icing data corresponding to the transmission line includes: Obtain the meteorological data corresponding to each driving factor and obtain historical icing data, where the driving factors are various elements affecting weather phenomena; For the meteorological data corresponding to each driving factor and the obtained historical icing data, delete local abnormal data points to obtain the cleaned historical meteorological data and historical icing data; Align the time of the cleaned historical meteorological data and historical icing data to obtain the aligned historical meteorological data and historical icing data.

8. An apparatus for mining driving factors of ice coating thickness on a transmission line, characterized in that The device includes: A data acquisition module for acquiring historical meteorological data and historical icing data corresponding to a transmission line, where the historical meteorological data includes meteorological data corresponding to each driving factor; the driving factors are meteorological elements that affect the icing thickness of the transmission line; A feature determination module for determining the correlation coefficient corresponding to the driving factor based on the historical meteorological data and the historical icing data, and determining feature data based on the correlation coefficient; A model training module for training at least one icing thickness prediction model according to the meteorological data corresponding to the feature data; A contribution degree determination module for determining the contribution degree corresponding to each driving factor based on the icing thickness prediction model; the contribution degree is determined based on the interaction effect between the driving factors; A driving factor determination module that determines a driving factor combination based on the contribution degree, and the driving factor combination is used to determine the disaster prevention decision for the icing disaster of the transmission line.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images