Method, device and equipment for predicting icing thickness of power transmission and distribution line and storage medium

Abstract

The invention provides a power transmission and distribution line icing thickness prediction method and device, equipment and a storage medium, and relates to the technical field of power systems. The method comprises the following steps: in response to a prediction instruction for the icing thickness of the power transmission and distribution line, obtaining first meteorological data and line state data of the power transmission and distribution line and second meteorological data corresponding to a to-be-predicted moment; and inputting the first meteorological data, the line state data and the second meteorological data into a power transmission and distribution line icing thickness prediction model for icing thickness prediction to obtain a target icing thickness corresponding to a to-be-predicted moment, the power transmission and distribution line icing thickness prediction model is constructed by adopting an extreme gradient lifting algorithm based on historical meteorological data and historical line state data of the power transmission and distribution line at different historical moments. According to the invention, the prediction result of the icing thickness of the power transmission and distribution line can be obtained more accurately.

Description

Technical Field

[0001] The present application relates to the technical field of power systems, and in particular to a method, device, equipment, and storage medium for predicting ice thickness on transmission and distribution lines. Background Art

[0002] The safe operation of transmission and distribution lines is crucial to the stability of the power system. In cold or volatile climates, heavy ice accumulation can cause transmission and distribution lines to break, toppling towers, and disrupt power supply. Therefore, accurately predicting ice accumulation on transmission and distribution lines is crucial to ensuring the safe and stable operation of the power system.

[0003] Currently, statistical methods are often used to calculate the growth pattern of ice thickness on transmission and distribution lines, and then predict the ice thickness of transmission and distribution lines based on this growth pattern. However, this method of predicting ice thickness on transmission and distribution lines suffers from inaccurate prediction results. Summary of the Invention

[0004] The present application provides a method, device, equipment and storage medium for predicting ice thickness of power transmission and distribution lines to solve the problem that the current method of predicting ice thickness of power transmission and distribution lines results in inaccurate prediction results.

[0005] In a first aspect, the present application provides a method for predicting ice thickness of transmission and distribution lines, comprising:

[0006] In response to a prediction instruction for ice thickness of a power transmission and distribution line, obtaining first meteorological data, line status data, and second meteorological data corresponding to a time to be predicted of the power transmission and distribution line;

[0007] The first meteorological data, line status data and second meteorological data are input into the transmission and distribution line ice thickness prediction model to predict the ice thickness and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical times, and is constructed using the extreme gradient boosting algorithm.

[0008] Optionally, the transmission and distribution line ice thickness prediction model is obtained by training in the following manner: obtaining training samples, the training samples including historical meteorological data and historical line status data of the transmission and distribution lines at different historical moments; based on the training samples, iteratively training the transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation conditions or the number of iterations reaches the preset number of iterations, thereby obtaining the trained transmission and distribution line ice thickness prediction model; wherein, in each iteration, the transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree, and the residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness, and the predicted ice thickness is obtained by accumulating the outputs of the leaf nodes in the decision tree by the transmission and distribution line ice thickness prediction model.

[0009] Optionally, the objective function of the ice thickness prediction model for transmission and distribution lines satisfies the following formula:

[0010]

[0011] Among them, i represents the i-th training sample; y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; represents the loss function of a single training sample; t represents the total number of decision trees in the transmission and distribution line ice thickness prediction model; k represents the kth decision tree in the transmission and distribution line ice thickness prediction model; Ω(f k ) represents the regularization term of the ice thickness prediction model for transmission and distribution lines; represents the objective function; n represents the total number of training samples.

[0012] Optionally, based on the training samples, an ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm is iteratively trained, including: based on the training samples, using a correlation analysis method to determine target training samples related to ice thickness; preprocessing the target training samples to obtain preprocessed samples, the preprocessing including at least one of data formatting processing, data outlier processing, and data missing value filling; based on the preprocessed samples, an ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm is iteratively trained.

[0013] Optionally, based on the training samples, the transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm is iteratively trained, including: in the iterative training, the hyperparameters of the transmission and distribution line ice thickness prediction model are adjusted using a preset adjustment method to obtain the adjusted hyperparameters; the preset adjustment method includes grid search and / or cross validation, and the hyperparameters include at least one of the learning rate, the maximum depth of the decision tree, the minimum weight of the leaf node, and the regularization parameter.

[0014] Optionally, after obtaining the trained transmission and distribution line ice thickness prediction model, the transmission and distribution line ice thickness prediction method also includes: based on the test sample, using preset indicators to evaluate the performance of the transmission and distribution line ice thickness prediction model to obtain an evaluation result, and the evaluation result is used to indicate whether to optimize the transmission and distribution line ice thickness prediction model, and the preset indicators include at least one of the mean square error, mean absolute error and mean relative error.

[0015] Optionally, the first meteorological data includes at least temperature, humidity, wind speed and precipitation; the line status data includes at least ice cover weight, maximum tension and wind deflection angle corresponding to the transmission and distribution lines; the second meteorological data includes at least temperature, humidity, wind speed and precipitation corresponding to the predicted time.

[0016] In a second aspect, the present application provides a device for predicting ice thickness of a power transmission and distribution line, comprising:

[0017] an acquisition module, configured to acquire, in response to a prediction instruction for ice thickness of a power transmission and distribution line, first meteorological data and line status data of the power transmission and distribution line and second meteorological data corresponding to a time to be predicted;

[0018] The prediction module is used to input the first meteorological data, line status data and second meteorological data into the transmission and distribution line ice thickness prediction model to predict the ice thickness and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical times, and is constructed using the extreme gradient boosting algorithm.

[0019] Optionally, the power transmission and distribution line ice thickness prediction device also includes a training module, which is used to train and obtain a power transmission and distribution line ice thickness prediction model in the following manner: obtaining training samples, the training samples including historical meteorological data and historical line status data of the power transmission and distribution lines at different historical moments; based on the training samples, iteratively training the power transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation conditions or the number of iterations reaches the preset number of iterations, thereby obtaining a trained power transmission and distribution line ice thickness prediction model; wherein, in each iteration, the power transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree, and the residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness, and the predicted ice thickness is obtained by accumulating the outputs of the leaf nodes in the decision tree by the power transmission and distribution line ice thickness prediction model.

[0020] Optionally, the objective function of the ice thickness prediction model for transmission and distribution lines satisfies the following formula:

[0021]

[0022] Among them, i represents the i-th training sample; y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; represents the loss function of a single training sample; t represents the total number of decision trees in the transmission and distribution line ice thickness prediction model; k represents the kth decision tree in the transmission and distribution line ice thickness prediction model; Ω(f k ) represents the regularization term of the ice thickness prediction model for transmission and distribution lines; represents the objective function; n represents the total number of training samples.

[0023] Optionally, when the training module is used to iteratively train the ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm based on training samples, it is specifically used to: based on the training samples, use a correlation analysis method to determine the target training samples related to the ice thickness; preprocess the target training samples to obtain preprocessed samples, and the preprocessing includes at least one of data formatting processing, data outlier processing and data missing value filling; based on the preprocessed samples, iteratively train the ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm.

[0024] Optionally, when the training module is used to iteratively train a power transmission and distribution line ice thickness prediction model based on an extreme gradient boosting algorithm based on training samples, it is specifically used to: during the iterative training, adjust the hyperparameters of the power transmission and distribution line ice thickness prediction model using a preset adjustment method to obtain adjusted hyperparameters; the preset adjustment method includes grid search and / or cross-validation, and the hyperparameters include at least one of the learning rate, the maximum depth of the decision tree, the minimum weight of the leaf node, and the regularization parameter.

[0025] Optionally, the training module is also used to: after obtaining the trained transmission and distribution line ice thickness prediction model, based on the test sample, use preset indicators to evaluate the performance of the transmission and distribution line ice thickness prediction model to obtain an evaluation result, and the evaluation result is used to indicate whether to optimize the transmission and distribution line ice thickness prediction model. The preset indicators include at least one of the mean square error, mean absolute error and mean relative error.

[0026] Optionally, the first meteorological data includes at least temperature, humidity, wind speed and precipitation; the line status data includes at least ice cover weight, maximum tension and wind deflection angle corresponding to the transmission and distribution lines; the second meteorological data includes at least temperature, humidity, wind speed and precipitation corresponding to the predicted time.

[0027] In a third aspect, the present application provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor;

[0028] Memory stores computer-executable instructions;

[0029] The processor executes the computer-executable instructions stored in the memory to implement the method for predicting ice thickness of transmission and distribution lines as described in the first aspect of the present application.

[0030] In a fourth aspect, the present application provides a computer-readable storage medium, in which computer program instructions are stored. When the computer program instructions are executed, the method for predicting ice thickness of transmission and distribution lines as described in the first aspect of the present application is implemented.

[0031] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed, implements the method for predicting ice thickness of transmission and distribution lines as described in the first aspect of the present application.

[0032] The present application provides a method, apparatus, device, and storage medium for predicting ice thickness on power transmission and distribution lines. In response to a prediction instruction for ice thickness on a power transmission and distribution line, the method obtains first meteorological data, line status data, and second meteorological data corresponding to the time to be predicted for the power transmission and distribution line; the first meteorological data, line status data, and second meteorological data are input into a power transmission and distribution line ice thickness prediction model to predict ice thickness, thereby obtaining a target ice thickness corresponding to the time to be predicted. The power transmission and distribution line ice thickness prediction model is constructed using an extreme gradient boosting algorithm based on historical meteorological data and historical line status data of the power transmission and distribution line at different historical times. It can address the deficiencies of existing technologies in nonlinear processing, feature selection, and computational efficiency, and has good generalization capabilities. Furthermore, when predicting ice thickness using the power transmission and distribution line ice thickness prediction model, more accurate prediction results can be obtained, prediction accuracy can be improved, and accurate prediction support can be provided for the operation of the power system, thereby helping to ensure the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0034] Figure 1 This is a flow chart of a method for predicting ice thickness on power transmission and distribution lines provided in one embodiment of the present application;

[0035] Figure 2 A flowchart of a method for training a transmission and distribution line ice thickness prediction model provided in one embodiment of the present application;

[0036] Figure 3 A flowchart of a method for training a transmission and distribution line ice thickness prediction model provided in another embodiment of the present application;

[0037] Figure 4 A schematic diagram of a correlation analysis provided in one embodiment of the present application;

[0038] Figure 5 A schematic diagram of the structure of a device for predicting ice thickness on power transmission and distribution lines provided in one embodiment of the present application;

[0039] Figure 6 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application.

[0040] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0041] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0042] 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 used 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 must comply with relevant laws, regulations and standards, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0043] First, some technical terms involved in this application are explained:

[0044] The Extreme Gradient Boosting (XGBoost) algorithm is an efficient gradient boosting algorithm with the advantage of handling nonlinear relationships. Through optimized objective functions, feature selection methods, and parallel computing capabilities, the XGBoost algorithm can provide efficient and accurate prediction solutions in complex data environments.

[0045] The safe operation of transmission and distribution lines is crucial to the stability of the power system. In cold or volatile climates, heavy ice accumulation can cause transmission and distribution lines to break, toppling towers, and disrupt power supply. Therefore, accurately predicting ice accumulation on transmission and distribution lines is crucial to ensuring the safe and stable operation of the power system.

[0046] At present, statistical methods are usually used to calculate the growth law of ice thickness on transmission and distribution lines, and then the ice thickness of transmission and distribution lines is predicted based on the growth law. In addition, the ice thickness of transmission and distribution lines can also be predicted based on physical models. Although the above method can explain the formation mechanism of ice, it has the problem of inaccurate prediction results. Specifically, there are the following problems: (1) Nonlinear problem: The relationship between climate conditions (such as temperature, humidity and wind speed) and ice formation is complex, and the above method is difficult to process highly nonlinear data; (2) Difficulty in feature selection: There are many factors that affect ice thickness, and how to screen out key features to improve prediction accuracy is a challenge; (3) Low computational efficiency: When facing large-scale data or multi-site real-time prediction, the above method often has low computational efficiency and is difficult to meet the needs of rapid response.

[0047] Taking into account the advantage of the XGBoost algorithm in processing nonlinear relationships, based on the above problems, the present application provides a method, device, equipment and storage medium for predicting the ice thickness of transmission and distribution lines. Based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical moments, the XGBoost algorithm is used to construct a transmission and distribution line ice thickness prediction model. The first meteorological data, line status data and second meteorological data corresponding to the time to be predicted of the transmission and distribution lines obtained can be input into the transmission and distribution line ice thickness prediction model to predict the ice thickness, and the target ice thickness corresponding to the time to be predicted can be obtained. This can solve the shortcomings of the existing technology in nonlinear processing, feature selection and computational efficiency, and more accurately predict the ice thickness of the transmission and distribution lines, thereby providing accurate prediction support for the operation of the power system, and helping to ensure the safe and stable operation of the power system.

[0048] It should be noted that the method for predicting ice thickness of power transmission and distribution lines provided in the embodiment of the present application can be applied in a server, which can be an independent server or a service cluster.

[0049] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0050] Figure 1 This is a flow chart of a method for predicting ice thickness on power transmission and distribution lines provided in one embodiment of the present application. Figure 1 As shown, the method for predicting ice thickness of transmission and distribution lines in an embodiment of the present application includes:

[0051] S101 : In response to a prediction instruction for ice thickness of a power transmission and distribution line, obtain first meteorological data, line status data, and second meteorological data corresponding to a time to be predicted of the power transmission and distribution line.

[0052] In an embodiment of the present application, a prediction instruction for the ice thickness of the power transmission and distribution line may be input by a user to an electronic device executing an embodiment of the present method, or may be sent by another device to an electronic device executing an embodiment of the present method. For example, the first meteorological data of the power transmission and distribution line may be collected in real time by a meteorological sensor corresponding to the power transmission and distribution line, and the line status data of the power transmission and distribution line may be collected by a line monitoring device corresponding to the power transmission and distribution line. The time to be predicted is a certain time in the future, and the embodiment of the present application does not limit the time to be predicted. The second meteorological data corresponding to the time to be predicted is, for example, obtained based on a weather forecast issued by a meteorological station.

[0053] Optionally, the first meteorological data includes at least temperature, humidity, wind speed and precipitation; the line status data includes at least ice cover weight, maximum tension and wind deflection angle corresponding to the transmission and distribution lines; the second meteorological data includes at least temperature, humidity, wind speed and precipitation corresponding to the predicted time.

[0054] For example, first meteorological data such as temperature, humidity, wind speed and precipitation can be collected in real time by meteorological sensors, and line status data such as ice weight, maximum tension and wind deflection angle corresponding to the power transmission and distribution lines can be collected by line monitoring equipment. The first meteorological data can be uploaded to the data collection platform in the electronic device that executes the embodiment of the present method in real time by wireless communication. The data collection frequency of the first meteorological data can be set to once every 10 minutes, for example, to ensure that it can reflect real-time meteorological changes. The line status data can be collected regularly according to the patrol cycle of the power transmission and distribution lines, and matched with the real-time first meteorological data. The second meteorological data, for example, includes data such as temperature, humidity, wind speed and precipitation corresponding to the time to be predicted.

[0055] S102. Input the first meteorological data, the line status data, and the second meteorological data into the transmission and distribution line ice thickness prediction model to predict the ice thickness and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical times, and is constructed using the extreme gradient boosting algorithm.

[0056] In this step, the transmission and distribution line ice thickness prediction model is constructed based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical moments using the extreme gradient boosting (XGBoost) algorithm. The transmission and distribution line ice thickness prediction model can be understood as a regression model based on the XGBoost algorithm. The XGBoost algorithm iteratively optimizes the prediction results of the transmission and distribution line ice thickness prediction model by constructing a series of decision trees, gradually reducing the prediction error. For specific information on how to train and obtain the transmission and distribution line ice thickness prediction model, please refer to the subsequent embodiments, which will not be repeated here.

[0057] For example, after obtaining first meteorological data and line status data for a transmission and distribution line, as well as second meteorological data corresponding to a predicted time, the first meteorological data, line status data, and second meteorological data can be input into a transmission and distribution line ice thickness prediction model to predict ice thickness and obtain a target ice thickness corresponding to the predicted time. After obtaining the target ice thickness corresponding to the predicted time, the target ice thickness corresponding to the predicted time can be output and used to support the stable operation of the power system.

[0058] The method for predicting ice thickness of power transmission and distribution lines provided in an embodiment of the present application responds to a prediction instruction for ice thickness of power transmission and distribution lines, obtains first meteorological data, line status data, and second meteorological data corresponding to the time to be predicted of the power transmission and distribution lines; inputs the first meteorological data, line status data, and second meteorological data into a power transmission and distribution line ice thickness prediction model to predict ice thickness, and obtains a target ice thickness corresponding to the time to be predicted. The power transmission and distribution line ice thickness prediction model is constructed using an extreme gradient boosting algorithm based on historical meteorological data and historical line status data of the power transmission and distribution lines at different historical times. It can address the deficiencies of existing technologies in nonlinear processing, feature selection, and computational efficiency, and has good generalization capabilities. Therefore, when predicting ice thickness using the power transmission and distribution line ice thickness prediction model, it can obtain more accurate prediction results, improve prediction accuracy, and provide accurate prediction support for the operation of the power system, thereby helping to ensure the safe and stable operation of the power system.

[0059] Figure 2 This is a flow chart of a method for training a transmission and distribution line ice thickness prediction model according to an embodiment of the present application. Figure 2 As shown, the training method of the embodiment of the present application includes:

[0060] S201. Obtain training samples, where the training samples include historical meteorological data and historical line status data of power transmission and distribution lines at different historical moments.

[0061] For example, historical meteorological data may include data such as temperature, humidity, wind speed, and precipitation, while historical line status data may include data such as ice weight, maximum tension, and wind yaw angle corresponding to the transmission and distribution lines. Historical meteorological data can be understood as historical meteorological characteristics, including temperature, humidity, wind speed, and precipitation, while historical line status data can also be understood as historical line status characteristics, including ice weight, maximum tension, and wind yaw angle corresponding to the transmission and distribution lines. Historical meteorological data of the transmission and distribution lines at different historical moments can be collected using meteorological sensors corresponding to the transmission and distribution lines, and line status data of the transmission and distribution lines at different historical moments can be collected using line monitoring equipment corresponding to the transmission and distribution lines. Data collection should be high-frequency and real-time to ensure that the collected data reflects dynamic changes in meteorological conditions and line status. For example, the collection frequency of historical meteorological data can be set to every 10 minutes, and historical line status data can be collected regularly based on the inspection cycle of the transmission and distribution lines and matched with historical meteorological data. The collected historical meteorological data and historical line status data can be stored as original dynamic data files as training samples to facilitate subsequent data processing and training of the transmission and distribution line ice thickness prediction model.

[0062] S202. Based on the training samples, iteratively train the transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation condition or the number of iterations reaches the preset number of iterations, thereby obtaining the trained transmission and distribution line ice thickness prediction model.

[0063] Among them, in each iteration, the transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree. The residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness. The predicted ice thickness is obtained by accumulating the output of the leaf nodes in the decision tree by the transmission and distribution line ice thickness prediction model.

[0064] In this step, based on the training samples, the extreme gradient boosting (XGBoost) algorithm is used to construct a transmission and distribution line ice thickness prediction model. Among them, the XGBoost algorithm iteratively optimizes the prediction results of the transmission and distribution line ice thickness prediction model by constructing a series of decision trees, gradually reducing the prediction error, and gradually constructing decision trees through iterative training and obtaining the residual of each decision tree, thereby optimizing the performance of the transmission and distribution line ice thickness prediction model. The transmission and distribution line ice thickness prediction model can be understood as a regression model based on the XGBoost algorithm. For example, first, the transmission and distribution line ice thickness prediction model is initialized by an initial ice thickness prediction value. For example, the average ice thickness obtained based on historical ice data can be used as the initial ice thickness prediction value. Due to the rapid changes in meteorological conditions and the presence of nonlinear effects, the initial ice thickness prediction value usually has a large error with the actual ice situation. For each collected training sample, the residual is the difference between the reference ice thickness corresponding to the historical moment to be predicted (i.e., the actual ice thickness) and the predicted ice thickness output by the transmission and distribution line ice thickness prediction model. In icing scenarios, the residuals reflect changes in icing caused by sudden temperature drops, sudden changes in wind speed, or humidity fluctuations that are not captured by the current transmission and distribution line ice thickness prediction model. Using the residuals obtained from the previous transmission and distribution line ice thickness prediction model, a new decision tree is constructed using the XGBoost algorithm. The goal is to learn the icing patterns from the residuals. For example, when the temperature drops suddenly or the wind speed is abnormal, the growth of ice may be more obvious. The new decision tree can be used to capture errors caused by extreme meteorological conditions. Each new decision tree acts as a "weak learner," continuously correcting the deficiencies in the previous round of transmission and distribution line ice thickness prediction models, gradually improving the overall prediction accuracy.

[0065] After the new decision tree training is completed, its prediction results will be accumulated with the predicted ice thickness output by the original transmission and distribution line ice thickness prediction model. This process is usually multiplied by a learning rate (reduction factor) to control the contribution of each decision tree to the overall transmission and distribution line ice thickness prediction model, thereby preventing overfitting. Specifically, the current predicted ice thickness = the previous predicted ice thickness + learning rate × the fitting value of the new decision tree for the ice residual. This accumulation process enables the transmission and distribution line ice thickness prediction model to gradually correct the deviation of the initial prediction and become closer and closer to the actual ice situation. Among them, each iteration will re-obtain the residual based on the updated transmission and distribution line ice thickness prediction model, construct a new decision tree fitting residual, and update the prediction of the transmission and distribution line ice thickness prediction model, until the calculated loss function value meets the preset evaluation conditions or the number of iterations reaches the preset number of iterations, and the trained transmission and distribution line ice thickness prediction model is obtained. During each iteration, the XGBoost algorithm can also use regularization terms to limit the complexity of the decision tree, preventing the ice thickness prediction model for power transmission and distribution lines from overfitting when faced with noisy training samples.

[0066] It can be understood that through iterative training, each newly constructed decision tree takes the current residual as the target, fits a part of the error, and gradually corrects the prediction results of the ice thickness prediction model of the transmission and distribution lines. The final predicted ice thickness is obtained by adding the output (adjustment amount) of the leaf nodes in all decision trees to the initial ice thickness prediction value. The root node in the decision tree is responsible for the preliminary data segmentation and selecting the optimal splitting condition (such as whether the temperature is lower than the temperature threshold) in order to minimize the overall prediction error. For data that falls into a leaf node, the value of its leaf node will be considered as the ice thickness adjustment amount under this sub-group. After the outputs of the leaf nodes in all decision trees are accumulated and the initial ice thickness prediction value is added, the final predicted ice thickness is obtained. Among them, the loss function is, for example, the mean squared error (MSE), which is the following formula 1:

[0067]

[0068] Among them, y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; L represents the loss function value.

[0069] The regularization term is used to penalize the complexity of the ice thickness prediction model for transmission and distribution lines, satisfying the following formula 2:

[0070]

[0071] Where T represents the total number of leaf nodes in the decision tree; w jThe weight of the jth leaf node; γ is used to control the number of leaf nodes; λ is used to control the sum of squares of weights.

[0072] Optionally, the objective function of the ice thickness prediction model for transmission and distribution lines satisfies the following formula 3:

[0073]

[0074] Among them, i represents the i-th training sample; y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; represents the loss function of a single training sample, which is used to measure the error between the reference ice thickness and the predicted ice thickness; t represents the total number of decision trees in the transmission and distribution line ice thickness prediction model; k is an index, which represents the kth decision tree in the transmission and distribution line ice thickness prediction model; Ω(f k ) represents the regularization term of the ice thickness prediction model for transmission and distribution lines; represents the objective function; n represents the total number of training samples.

[0075] It can be understood that the objective function is used to obtain the loss function value of n training samples and limit the complexity of the decision tree through the regularization term.

[0076] Optionally, during iterative training, a preset adjustment method is used to adjust the hyperparameters of the transmission and distribution line ice thickness prediction model to obtain adjusted hyperparameters; the preset adjustment method includes grid search and / or cross validation, and the hyperparameters include at least one of the learning rate, the maximum depth of the decision tree, the minimum weight of the leaf node, and the regularization parameter.

[0077] For example, grid search, cross-validation, and other methods can be used to adjust and optimize the hyperparameters of the transmission and distribution line ice thickness prediction model. These hyperparameters may include, for example, the learning rate, the maximum depth of the decision tree, the minimum weight of leaf nodes, and the regularization parameter. They may also include the subsample ratio, feature sampling ratio, and number of iterations. For specific hyperparameters, please refer to the subsequent embodiments and will not be detailed here.

[0078] The training method for the ice thickness prediction model for power transmission and distribution lines provided in an embodiment of the present application obtains training samples, the training samples including historical meteorological data and historical line status data of the power transmission and distribution lines at different historical moments; based on the training samples, iteratively trains the ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm until the calculated loss function value satisfies a preset evaluation condition or the number of iterations reaches a preset number of iterations, thereby obtaining a trained ice thickness prediction model for power transmission and distribution lines. In each iteration, the ice thickness prediction model for power transmission and distribution lines gradually constructs a new decision tree to fit the residual of the previous decision tree. The residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness. The predicted ice thickness is obtained by accumulating the outputs of the leaf nodes in the decision tree by the ice thickness prediction model for power transmission and distribution lines. This method can fully learn the relationship between historical meteorological data and historical line status data, addressing the deficiencies of the prior art in nonlinear processing, feature selection, and computational efficiency. It has good generalization ability and can be used to more accurately predict the ice thickness of power transmission and distribution lines.

[0079] Based on the above embodiment, the hyper parameters of the ice thickness prediction model for transmission and distribution lines are as follows:

[0080] (1) Learning rate, which is used to control the contribution of each decision tree to the ice thickness prediction model for transmission and distribution lines, and is equivalent to the step size of gradient boosting. A lower learning rate can make the learning of the ice thickness prediction model for transmission and distribution lines more robust, reduce the amplitude of each update, and help improve generalization performance, but more decision trees are needed to achieve the same effect. In ice thickness prediction, a learning rate of, for example, 0.05 to 0.1 can be used to balance the training speed and accuracy of the ice thickness prediction model for transmission and distribution lines, which is conducive to capturing the nonlinear relationship between weather and line status, while avoiding overfitting caused by too fast convergence.

[0081] (2) The maximum depth of the tree is used to control the maximum depth of a single decision tree, that is, the complexity of the feature interaction that the transmission and distribution line ice thickness prediction model can learn. The greater the depth, the stronger the fitting ability of the transmission and distribution line ice thickness prediction model. However, a decision tree that is too deep is prone to overfitting the training samples. For example, a medium depth (6-8) can be used to fully learn the relationship between historical meteorological data and historical line status data. At the same time, the risk of overfitting can be reduced by limiting the depth. For example, a decision tree with a depth of 6 is often sufficient to capture the effects of temperature, humidity, etc. on ice coverage without introducing too many noise features.

[0082] (3) Minimum weight of leaf nodes, used to determine the minimum threshold for the sum of sample weights contained in each leaf node; a larger value requires each leaf to contain at least enough sample weights to avoid splitting nodes due to very few training samples; in ice thickness prediction, this hyperparameter can prevent the ice thickness prediction model of transmission and distribution lines from building too deep branches for abnormal weather or individual line conditions; for example, setting the minimum weight of leaf nodes to 5 means that splitting is allowed only when the sum of sample weights contained in each leaf node after splitting is ≥5, which helps to reduce the impact of noise and prevent overfitting;

[0083] (4) Subsample ratio, which is used to determine the proportion of randomly sampled training samples when constructing each decision tree. For example, a subsample ratio value of <1 can introduce randomness and reduce the correlation between decision trees, thereby improving the generalization ability of the transmission and distribution line ice thickness prediction model. A subsample ratio value of about 0.8, for example, can ensure that each decision tree sees most of the data to learn the ice pattern. At the same time, by introducing moderate randomness by sampling slightly less than 100% of the samples, overfitting caused by extreme weather events can be alleviated. The subsample ratio value of 0.8 is an empirical value commonly used in practice.

[0084] (5) Feature sampling ratio, which is used to control the ratio of features randomly sampled during the training of each decision tree, similar to column sampling in random forests. The value of the feature sampling ratio is, for example, 0.8, which is used in combination with the subsample ratio to introduce random sampling of features. In ice thickness prediction, historical meteorological data and historical line status data may contain many features. By using only 80% of the features in each decision tree, it helps to reduce the interference between related features and improve the generalization performance of the ice thickness prediction model for transmission and distribution lines for unseen data. A feature sampling ratio that is too low may lead to underfitting, and a feature sampling ratio that is too high may result in insufficient randomness. The feature sampling ratio value of 0.8 is an empirical choice to balance bias and variance.

[0085] (6) Regularization parameter (regularization term), which is used to control the complexity of the transmission and distribution line ice thickness prediction model to prevent overfitting. The values of the regularization parameter are, for example: lambda = 1 (L2 regularization coefficient), alpha = 0 or 0.1 (L1 regularization coefficient); the intensity can be adjusted according to the complexity of the transmission and distribution line ice thickness prediction model, generally starting from lambda = 1 and alpha = 0. If there is a tendency to overfitting, these two values can be increased; it can be understood that the L2 regularization coefficient is used to penalize large weights, which can make the transmission and distribution line ice thickness prediction model smoother and more conservative. Setting lambda = 0 is a good choice. bda=1 means that a certain penalty is imposed on the sum of the squares of the leaf weights, which helps prevent the transmission and distribution line ice thickness prediction model from assigning too high weights to individual features and can reduce the impact of noisy weather data in ice thickness prediction. The L1 regularization coefficient is used to make the weights sparse and weaken irrelevant or weakly correlated features. For cases with many features (such as a large amount of sensor data), setting alpha≈0.1 can make the transmission and distribution line ice thickness prediction model automatically ignore some unimportant features. If the complexity of the transmission and distribution line ice thickness prediction model is still too high, the alpha value can be further increased to strengthen the regularization. The regularization parameter reduces the complexity of the transmission and distribution line ice thickness prediction model by adding a penalty term to the objective function, thereby improving generalization performance. A reasonable regularization parameter can prevent the transmission and distribution line ice thickness prediction model from overfitting historical meteorological anomalies and enhance the robustness of future predictions. The regularization parameter directly acts on the output (prediction adjustment value) of each leaf node, and by penalizing the square (L2 regularization coefficient) or absolute value (L1 regularization coefficient) of the leaf node weight, limiting its value to a small value, thereby preventing overfitting of the transmission and distribution line ice thickness prediction model.

[0086] (7) The number of iterations is used to determine the number of decision trees to be boosted, which is equivalent to the number of iterations or the number of rounds of gradient boosting. More decision trees can usually improve training accuracy, but too many decision trees will increase the risk of overfitting. Practical experience shows that 300-500 decision trees are often sufficient to capture the icing pattern. More decision trees may have little gain and are prone to overfitting. Therefore, a larger number of iterations is usually set first, such as 500. When the error no longer decreases after several consecutive rounds (such as 10 rounds), the training can be stopped in advance to obtain a number of decision trees close to the optimal number. This ensures that the transmission and distribution line ice thickness prediction model does not continue to learn noise after reaching the optimal complexity, thereby improving the generalization ability.

[0087] Figure 3This is a flowchart of a training method for a power transmission and distribution line ice thickness prediction model provided by another embodiment of the present application. Based on the above embodiment, this embodiment of the present application further illustrates the training method for a power transmission and distribution line ice thickness prediction model. Figure 3 As shown, the training method of the transmission and distribution line ice thickness prediction model in the embodiment of the present application may include:

[0088] S301. Obtain training samples, where the training samples include historical meteorological data and historical line status data of power transmission and distribution lines at different historical moments.

[0089] The detailed description of this step can be found in Figure 2 The relevant description of S201 in the illustrated embodiment will not be repeated here.

[0090] In the embodiment of this application, Figure 2 Step S202 may further include the following three steps S302 to S304:

[0091] S302: Based on the training samples, a correlation analysis method is used to determine target training samples related to ice thickness.

[0092] For example, historical meteorological data can be understood as historical meteorological characteristics, including characteristics such as temperature, humidity, wind speed and precipitation, and historical line status data can also be understood as historical line status characteristics, including characteristics such as ice weight, maximum tension and wind deflection angle corresponding to the transmission and distribution lines. The above-mentioned features can be standardized to remove redundant features to reduce the risk of overfitting. For each of the above-mentioned features, a correlation analysis method can be used to determine the important features that affect the ice thickness, thereby determining the target training samples related to the ice thickness. Among them, in order to avoid the influence of features of different dimensions on the model effect, all features can be standardized, such as using the standard score standardization (Z-score standardization) method to standardize all features to ensure that each feature has the same scale. Figure 4 A schematic diagram of a correlation analysis provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the correlation between the ice thickness on the horizontal axis and the maximum tension inclination angle, maximum tension wind angle, maximum tension, ice thickness, ice weight, temperature and data time on the vertical axis is shown. For example, the correlation between the ice thickness on the horizontal axis and the maximum tension inclination angle on the vertical axis is -0.046, indicating that the two are negatively correlated.

[0093] S303: Preprocess the target training sample to obtain a preprocessed sample, where the preprocessing includes at least one of data formatting, data outlier processing, and data missing value filling.

[0094] For example, in order to ensure the quality of the target training samples, the target training samples can be formatted and stored in a database for subsequent training of the ice thickness prediction model for the transmission and distribution lines. Among them, the data is stored in the form of a structured database (such as a relational database), specifically including converting the collected timestamps, location coordinates, meteorological data (such as temperature, wind speed, humidity, etc.), ice weight and other information into a standardized format to ensure the uniformity and operability between the data. For different types of meteorological data (such as temperature and humidity), a unified unit and format can be used for conversion to ensure that the data will not cause errors due to inconsistent formats in subsequent processing. The processed data is stored in a database for data query, analysis and subsequent modeling training. Table 1 shows the data corresponding to the stored target training samples.

[0095] Table 1

[0096]

[0097] The target training samples can be processed for data outliers. Common outliers such as temperature that is too low, temperature that is too high, or abnormal wind speed data will affect the accuracy of the transmission and distribution line ice thickness prediction model. A reasonable threshold range can be set to eliminate data that exceeds the threshold. For example, a reasonable range for each type of meteorological data can be set based on practical experience and meteorological principles. For example, the temperature cannot be lower than -50°C and cannot be higher than 50°C, and the wind speed cannot exceed 100m / s. Abnormal data that exceeds the threshold range can be automatically detected and marked by a preset algorithm. It is necessary to ensure that the processing of data outliers does not introduce new deviations to ensure the effectiveness and accuracy of the subsequent training of the transmission and distribution line ice thickness prediction model.

[0098] Missing data values can be filled in the target training samples. For example, the mean interpolation method, the K-nearest neighbor algorithm (a supervised learning algorithm), or an interpolation method based on a regression model can be used to reasonably fill in missing data (such as missing values caused by equipment failure or transmission delay), fill in the gaps in meteorological data and line status information, and ensure data integrity. It is necessary to ensure that the processing of filling missing data values does not introduce new biases to ensure the effectiveness and accuracy of the subsequent training of the transmission and distribution line ice thickness prediction model. In addition, for data that is zero for a long time, the data can be deleted and not included in the calculation of the transmission and distribution line ice thickness prediction model.

[0099] S304. Based on the preprocessed samples, iteratively train the transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation condition or the number of iterations reaches the preset number of iterations, thereby obtaining the trained transmission and distribution line ice thickness prediction model.

[0100] Among them, in each iteration, the transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree. The residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness. The predicted ice thickness is obtained by accumulating the output of the leaf nodes in the decision tree by the transmission and distribution line ice thickness prediction model.

[0101] The detailed description of this step can be found in Figure 2 The relevant description of S202 in the illustrated embodiment will not be repeated here.

[0102] S305. Based on the test samples, the performance of the ice thickness prediction model for the transmission and distribution lines is evaluated using preset indicators to obtain an evaluation result. The evaluation result is used to indicate whether the ice thickness prediction model for the transmission and distribution lines is optimized; wherein the preset indicators include at least one of the mean square error, the mean absolute error, and the mean relative error.

[0103] For example, the performance of the ice thickness prediction model for power transmission and distribution lines can be verified using preset indicators such as mean square error (MSE), mean absolute error (MAE), and mean relative error (MRE). Specifically, after the performance training of the ice thickness prediction model for power transmission and distribution lines is completed, the prediction effect of the ice thickness prediction model for power transmission and distribution lines can be comprehensively evaluated based on the test samples included in the test set using preset indicators. Among them, the mean square error (MSE) is used to obtain the mean of the squared differences between all predicted ice thicknesses and the actual ice thickness (reference ice thickness), to evaluate the overall deviation of the prediction results of the ice thickness prediction model for power transmission and distribution lines, and can reflect the overall deviation of the prediction results; a smaller MSE indicates that the prediction error of the ice thickness prediction model for power transmission and distribution lines is smaller. The mean absolute error (MAE) is used to obtain the mean of the absolute errors between all predicted ice thicknesses and the actual ice thickness, to evaluate the accuracy of the prediction of the ice thickness prediction model for power transmission and distribution lines; a smaller MAE indicates that the deviation between the prediction result and the actual ice thickness is smaller. The mean relative error (MRE) is used to obtain the average relative error between all predicted ice thicknesses and the actual ice thickness, evaluating the prediction accuracy of the transmission and distribution line ice thickness prediction model under different scales of data. A smaller MRE indicates that the transmission and distribution line ice thickness prediction model has good prediction capabilities. Based on the above preset indicators, the prediction capabilities of the transmission and distribution line ice thickness prediction model are comprehensively evaluated, and the evaluation results are used to indicate whether the transmission and distribution line ice thickness prediction model should be optimized to ensure that the transmission and distribution line ice thickness prediction model can adapt to ice thickness prediction under different meteorological conditions and provide accurate prediction results for subsequent real-time monitoring and early warning.

[0104] During the evaluation process, if the performance of the transmission and distribution line ice thickness prediction model does not meet expectations, the transmission and distribution line ice thickness prediction model can be further optimized by adjusting hyperparameters or increasing training samples.

[0105] The training method of the transmission and distribution line ice thickness prediction model provided in the embodiment of the present application obtains training samples, and the training samples include historical meteorological data and historical line status data of the transmission and distribution lines at different historical moments; based on the training samples, a correlation analysis method is used to determine target training samples related to ice thickness, and the target training samples are preprocessed to obtain preprocessed samples, and the preprocessing includes at least one of data formatting processing, data outlier processing, and data missing value filling, which can effectively ensure the quality of the training samples; based on the preprocessed samples, iteratively train the transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation condition or the number of iterations reaches the preset number of iterations, thereby obtaining the trained transmission and distribution line ice thickness prediction model; based on the test samples, the performance of the transmission and distribution line ice thickness prediction model is evaluated using preset indicators to obtain an evaluation result, which is used to indicate whether the transmission and distribution line ice thickness prediction model is optimized, and the preset indicators include at least one of mean square error, mean absolute error, and mean relative error, which can ensure that the transmission and distribution line ice thickness prediction model can adapt to ice thickness prediction under different meteorological conditions. Among them, in each iteration, the transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree. The residual is obtained based on the reference ice thickness and predicted ice thickness corresponding to the historical moment to be predicted. The predicted ice thickness is obtained by accumulating the output of the leaf nodes in the decision tree by the transmission and distribution line ice thickness prediction model. It can fully learn the relationship between historical meteorological data and historical line status data, solve the shortcomings of existing technologies in nonlinear processing, feature selection and computational efficiency, and has good generalization ability, so that it can be used to more accurately predict the ice thickness of transmission and distribution lines.

[0106] The following are device embodiments of the present application, which can be used to implement the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

[0107] Figure 5 This is a schematic diagram of the structure of the device for predicting ice thickness on power transmission and distribution lines provided in one embodiment of the present application. Figure 5 As shown, the device 500 for predicting ice thickness of power transmission and distribution lines according to the embodiment of the present application includes: an acquisition module 501 and a prediction module 502.

[0108] The acquisition module 501 is configured to acquire first meteorological data, line status data, and second meteorological data corresponding to a time to be predicted of the transmission and distribution line in response to a prediction instruction for ice thickness of the transmission and distribution line.

[0109] The prediction module 502 is used to input the first meteorological data, the line status data and the second meteorological data into the transmission and distribution line ice thickness prediction model to predict the ice thickness and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution lines at different historical times, and is constructed using the extreme gradient boosting algorithm.

[0110] Optionally, the power transmission and distribution line ice thickness prediction device 500 may further include a training module 503, which is used to train and obtain a power transmission and distribution line ice thickness prediction model in the following manner: obtaining training samples, the training samples including historical meteorological data and historical line status data of the power transmission and distribution lines at different historical moments; based on the training samples, iteratively training the power transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm until the calculated loss function value meets the preset evaluation condition or the number of iterations reaches the preset number of iterations, thereby obtaining a trained power transmission and distribution line ice thickness prediction model; wherein, in each iteration, the power transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree, and the residual is obtained based on the reference ice thickness corresponding to the historical moment to be predicted and the predicted ice thickness, and the predicted ice thickness is obtained by accumulating the outputs of the leaf nodes in the decision tree by the power transmission and distribution line ice thickness prediction model.

[0111] Optionally, the objective function of the ice thickness prediction model for transmission and distribution lines satisfies the following formula:

[0112]

[0113] Among them, i represents the i-th training sample; y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; represents the loss function of a single training sample; t represents the total number of decision trees in the transmission and distribution line ice thickness prediction model; k represents the kth decision tree in the transmission and distribution line ice thickness prediction model; Ω(f k ) represents the regularization term of the ice thickness prediction model for transmission and distribution lines; represents the objective function; n represents the total number of training samples.

[0114] Optionally, when the training module 503 is used to iteratively train the ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm based on the training samples, it can be specifically used to: based on the training samples, use the correlation analysis method to determine the target training samples related to the ice thickness; preprocess the target training samples to obtain preprocessed samples, and the preprocessing includes at least one of data formatting processing, data outlier processing and data missing value filling; based on the preprocessed samples, iteratively train the ice thickness prediction model for power transmission and distribution lines based on the extreme gradient boosting algorithm.

[0115] Optionally, when the training module 503 is used to iteratively train a power transmission and distribution line ice thickness prediction model based on an extreme gradient boosting algorithm based on training samples, it can be specifically used to: in the iterative training, adjust the hyperparameters of the power transmission and distribution line ice thickness prediction model using a preset adjustment method to obtain adjusted hyperparameters; the preset adjustment method includes grid search and / or cross-validation, and the hyperparameters include at least one of the learning rate, the maximum depth of the decision tree, the minimum weight of the leaf node, and the regularization parameter.

[0116] Optionally, the training module 503 can also be used to: after obtaining the trained transmission and distribution line ice thickness prediction model, based on the test sample, use preset indicators to evaluate the performance of the transmission and distribution line ice thickness prediction model to obtain an evaluation result, and the evaluation result is used to indicate whether to optimize the transmission and distribution line ice thickness prediction model. The preset indicators include at least one of the mean square error, mean absolute error and mean relative error.

[0117] Optionally, the first meteorological data includes at least temperature, humidity, wind speed and precipitation; the line status data includes at least ice cover weight, maximum tension and wind deflection angle corresponding to the transmission and distribution lines; the second meteorological data includes at least temperature, humidity, wind speed and precipitation corresponding to the predicted time.

[0118] The device of the embodiment of the present application can be used to execute the technical solution of any of the above-mentioned method embodiments. Its implementation principles and technical effects are similar and will not be repeated here.

[0119] Figure 6 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present application. Figure 6 As shown, the electronic device 600 may include: at least one processor 601 and a memory 602 .

[0120] The memory 602 is used to store programs. Specifically, the programs may include program codes, and the program codes include computer-executable instructions.

[0121] The memory 602 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0122] Processor 601 is configured to execute computer-executable instructions stored in memory 602 to implement the method for predicting ice thickness on power transmission and distribution lines described in the aforementioned method embodiment. Processor 601 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. Specifically, when implementing the method for predicting ice thickness on power transmission and distribution lines described in the aforementioned method embodiment, the electronic device may be, for example, a server or other electronic device with processing capabilities.

[0123] Optionally, the electronic device 600 may further include a communication interface 603. In a specific implementation, if the communication interface 603, the memory 602, and the processor 601 are implemented independently, the communication interface 603, the memory 602, and the processor 601 may be interconnected via a bus and communicate with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be divided into address buses, data buses, control buses, etc., but this does not mean that there is only one bus or only one type of bus.

[0124] Optionally, in a specific implementation, if the communication interface 603, the memory 602 and the processor 601 are integrated on a chip, the communication interface 603, the memory 602 and the processor 601 can complete communication through an internal interface.

[0125] The present application also provides a computer-readable storage medium, in which computer program instructions are stored. When a processor executes the computer program instructions, the above-mentioned method for predicting ice thickness of power transmission and distribution lines is implemented.

[0126] The present application also provides a computer program product, including a computer program, which, when executed, implements the above-mentioned method for predicting ice thickness of power transmission and distribution lines.

[0127] The computer-readable storage medium may be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The computer-readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0128] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium may also be an integral part of the processor. The processor and the readable storage medium may be located in an application-specific integrated circuit. Of course, the processor and the readable storage medium may also exist as discrete components in the power transmission and distribution line ice thickness prediction device.

[0129] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for predicting ice thickness of power transmission and distribution lines, characterized in that: include: In response to a prediction instruction for ice thickness of a power transmission and distribution line, obtaining first meteorological data, line status data, and second meteorological data corresponding to a time to be predicted of the power transmission and distribution line; The first meteorological data, the line status data and the second meteorological data are input into the transmission and distribution line ice thickness prediction model to predict the ice thickness, and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution line at different historical times, and is constructed using the extreme gradient boosting algorithm.

2. The method for predicting ice thickness of power transmission and distribution lines according to claim 1, characterized in that: The ice thickness prediction model for transmission and distribution lines is obtained by training in the following way: Acquire training samples, wherein the training samples include historical meteorological data and historical line status data of the transmission and distribution line at different historical moments; Based on the training samples, the power transmission and distribution line ice thickness prediction model based on the extreme gradient boosting algorithm is iteratively trained until the calculated loss function value meets the preset evaluation conditions or the number of iterations reaches the preset number of iterations, thereby obtaining the trained power transmission and distribution line ice thickness prediction model; wherein, in each iteration, the power transmission and distribution line ice thickness prediction model gradually constructs a new decision tree to fit the residual of the previous decision tree, and the residual is obtained based on the reference ice thickness and the predicted ice thickness corresponding to the historical moment to be predicted, and the predicted ice thickness is obtained by accumulating the outputs of the leaf nodes in the decision tree by the power transmission and distribution line ice thickness prediction model.

3. The method for predicting ice thickness of power transmission and distribution lines according to claim 2, characterized in that: The objective function of the ice thickness prediction model for transmission and distribution lines satisfies the following formula: Among them, i represents the i-th training sample; y i represents the reference ice thickness of the i-th training sample; represents the predicted ice thickness of the i-th training sample; represents the loss function of a single training sample; t represents the total number of decision trees in the transmission and distribution line ice thickness prediction model; k represents the kth decision tree in the transmission and distribution line ice thickness prediction model; Ω(f k ) represents the regularization term of the ice thickness prediction model for the transmission and distribution line; represents the objective function; n represents the total number of training samples.

4. The method for predicting ice thickness of power transmission and distribution lines according to claim 2, characterized in that: The iterative training of the ice thickness prediction model for the power transmission and distribution lines based on the extreme gradient boosting algorithm based on the training samples includes: Based on the training samples, a correlation analysis method is used to determine target training samples related to ice thickness; Preprocessing the target training sample to obtain a preprocessed sample, wherein the preprocessing includes at least one of data formatting, data outlier processing, and data missing value filling; Based on the preprocessed samples, an ice thickness prediction model for power transmission and distribution lines based on an extreme gradient boosting algorithm is iteratively trained.

5. The method for predicting ice thickness of power transmission and distribution lines according to claim 2, characterized in that: The iterative training of the ice thickness prediction model for the power transmission and distribution lines based on the extreme gradient boosting algorithm based on the training samples includes: In the iterative training, a preset adjustment method is used to adjust the hyperparameters of the transmission and distribution line ice thickness prediction model to obtain adjusted hyperparameters; the preset adjustment method includes grid search and / or cross-validation, and the hyperparameters include at least one of the learning rate, the maximum depth of the decision tree, the minimum weight of the leaf node, and the regularization parameter.

6. The method for predicting ice thickness of power transmission and distribution lines according to claim 2, characterized in that: After obtaining the trained transmission and distribution line ice thickness prediction model, the method further includes: Based on the test samples, the performance of the transmission and distribution line ice thickness prediction model is evaluated using preset indicators to obtain an evaluation result. The evaluation result is used to indicate whether the transmission and distribution line ice thickness prediction model is optimized. The preset indicators include at least one of the mean square error, the mean absolute error and the mean relative error.

7. The method for predicting ice thickness of power transmission and distribution lines according to any one of claims 1 to 6, characterized in that: The first meteorological data includes at least temperature, humidity, wind speed and precipitation; the line status data includes at least ice weight, maximum tension and wind deflection angle corresponding to the transmission and distribution line; the second meteorological data includes at least temperature, humidity, wind speed and precipitation corresponding to the predicted time.

8. A device for predicting ice thickness of power transmission and distribution lines, characterized in that: include: an acquisition module, configured to acquire, in response to a prediction instruction for ice thickness of a power transmission and distribution line, first meteorological data and line status data of the power transmission and distribution line and second meteorological data corresponding to a time to be predicted; The prediction module is used to input the first meteorological data, the line status data and the second meteorological data into the transmission and distribution line ice thickness prediction model to predict the ice thickness and obtain the target ice thickness corresponding to the predicted time. The transmission and distribution line ice thickness prediction model is based on the historical meteorological data and historical line status data of the transmission and distribution line at different historical times, and is constructed using the extreme gradient boosting algorithm.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method for predicting ice thickness of power transmission and distribution lines according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed, the method for predicting ice thickness of a power transmission and distribution line according to any one of claims 1 to 7 is implemented.

11. A computer program product comprising a computer program, characterized in that When the computer program is executed, the method for predicting ice thickness of a power transmission and distribution line according to any one of claims 1 to 7 is implemented.

Images