Jumper wind load characteristic rapid calculation method based on GBDT intelligent algorithm

Abstract

The invention relates to the technical field of electric power engineering, in particular to a jumper wind load characteristic rapid calculation method based on a GBDT intelligent algorithm, which comprises the following steps: step 1, data collection and preprocessing; step 2, constructing and training a GBDT model; step 3, wind load characteristic calculation and analysis; and 4, verifying and optimizing a result. According to the method, the calculation efficiency is remarkably improved, the calculation time and cost are reduced, and the method is suitable for design optimization and operation safety evaluation of the jumper wire in the electric power system, especially has important practical significance in engineering application under complex climate conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power engineering, and in particular to a method for quickly calculating wind load characteristics of a jumper based on a GBDT intelligent algorithm. Background Art

[0002] In modern power systems, jumpers are one of the key components of transmission lines, and their wind load characteristics directly affect the safety and stability of the lines. With the development of transmission lines towards high and ultra-high voltage, and the popularization of lines under complex terrain and extreme climate conditions, the calculation of the aerodynamic characteristics of jumpers under different wind speeds, wind attack angles and icing conditions has become particularly important.

[0003] Traditional wind load calculation methods usually rely on wind tunnel experiments and numerical simulations. Although these methods are highly accurate, they have obvious limitations: wind tunnel experiments are costly and time-consuming, and the experimental conditions are difficult to cover all possible actual working conditions; numerical simulations require complex models and a large amount of computing resources, and the results are easily affected by meshing and boundary condition settings. Especially under multivariable and nonlinear conditions, these traditional methods are difficult to efficiently and accurately predict the wind load characteristics of jumpers. Summary of the invention

[0004] The present invention provides a method for quickly calculating the wind load characteristics of a jumper based on the GBDT intelligent algorithm, which can quickly calculate the wind load characteristics of a jumper.

[0005] According to the present invention, a method for quickly calculating wind load characteristics of a jumper based on a GBDT intelligent algorithm comprises the following steps:

[0006] Step 1: Data collection and preprocessing;

[0007] Step 2: GBDT model construction and training;

[0008] Step 3: Calculation and analysis of wind load characteristics;

[0009] Step 4: Result verification and optimization.

[0010] Preferably, in step 1, data collection specifically includes: collecting wind load data of the jumper under different wind speeds, wind attack angles and icing conditions through wind tunnel experiments, field measurements and numerical simulations;

[0011] Wind tunnel experiments provide data on the forces acting on jumpers under different climatic conditions, field measurement data reflect the wind load characteristics of jumpers under actual operating conditions, and numerical simulation data provide supplementary data for different working conditions.

[0012] Preferably, in step 1, preprocessing includes data cleaning, normalization, feature selection and data enhancement.

[0013] Preferably, in step 2, a GBDT wind load prediction model is constructed using the preprocessed data set; during the model construction process, the hyperparameters of GBDT need to be set, including the number of trees, the depth of the tree, and the learning rate; the specific training steps are as follows: first, the initial decision tree is trained using the preprocessed data to calculate the initial prediction value; then, the residual between the initial prediction value and the true value is calculated; then, the residual is used to train the next decision tree, and the above steps are repeated to gradually reduce the prediction error until all trees are trained; the final model is the weighted sum of all decision trees, which can capture the complex patterns of wind load data.

[0014] Preferably, in step 3, the trained GBDT model is used to quickly calculate the wind load characteristics of the jumper under various wind speeds, wind attack angles and icing conditions; the input of the model includes ice type, ice thickness, wind speed, number of sub-conductors and wind attack angle parameters, and the output is the corresponding aerodynamic coefficients, including drag, lift and torque coefficients; the calculation process is as follows: first, according to different input parameters, the model generates corresponding aerodynamic coefficient prediction values; then, by drawing a curve of the aerodynamic coefficient changing with the wind attack angle, the force characteristics and stability of the jumper under different conditions are analyzed.

[0015] Preferably, in step 4, specifically:

[0016] Compare and analyze the model's prediction results with the actual measurement data and calculate the determination coefficient R 2 and mean square error (MSE) indicators to evaluate the accuracy of the model; if the prediction error of the model is large, it is necessary to analyze the source of the error and improve the model performance by adjusting the model parameters or optimizing the algorithm; specific optimization methods include re-tuning the hyperparameters of GBDT, increasing the diversity of the training data set, and using more complex feature engineering to improve the predictive ability of the model.

[0017] The present invention provides a method for quickly calculating the wind load characteristics of a jumper based on a GBDT intelligent algorithm. By collecting experimental data and numerical simulation results of jumpers under different wind speeds, wind attack angles and icing conditions, a wind load prediction model is constructed, and the GBDT algorithm is used for training and optimization. Through this method, aerodynamic parameters such as the lift coefficient, drag coefficient and torque coefficient of the jumper can be quickly and accurately calculated. By utilizing the efficiency and accuracy of the GBDT algorithm, combined with actual measurement data and numerical simulation results, the wind load characteristics of the jumper can be quickly and accurately predicted under a variety of complex conditions. This method significantly improves the calculation efficiency, reduces the calculation time and cost, is suitable for the design optimization and operation safety assessment of jumpers in power systems, and has important practical significance, especially in engineering applications under complex climatic conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1It is a flow chart of a method for quickly calculating wind load characteristics of a jumper based on the GBDT intelligent algorithm in an embodiment;

[0019] Figure 2 It is a structural diagram of the GBDT (Gradient Boosting Decision Tree) model in the embodiment;

[0020] FIG3( a ) is a schematic diagram showing the distribution of the flow field around the jumper obtained through a wind tunnel test in an embodiment;

[0021] FIG3( b ) is a schematic diagram of aerodynamic time course curves of the jumper at different time points in the embodiment;

[0022] FIG3( c ) is a schematic diagram of a wind attack angle variation curve in an embodiment;

[0023] Figure 4 This is a diagram of the architecture of a model for rapid calculation of wind load characteristics of jumpers based on the GBDT intelligent algorithm in the embodiment;

[0024] FIG5( a ) is a schematic diagram of the prediction result of the lift coefficient in the embodiment;

[0025] FIG5( b ) is a schematic diagram of the prediction result of the drag coefficient in the embodiment;

[0026] FIG5( c ) is a schematic diagram of the prediction results of the torque coefficient in the embodiment. DETAILED DESCRIPTION

[0027] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are only for explaining the present invention and are not intended to limit it.

[0028] Example

[0029] like Figure 1 As shown, this embodiment provides a method for quickly calculating wind load characteristics of jumpers based on the GBDT intelligent algorithm, which includes the following steps:

[0030] Step 1: Data collection and preprocessing.

[0031] The data collection is specifically as follows: wind load data of jumpers under different wind speeds, wind attack angles and icing conditions are collected through wind tunnel experiments, field measurements and numerical simulations.

[0032] Wind tunnel experiments provide data on the forces acting on jumpers under different climatic conditions, field measurement data reflect the wind load characteristics of jumpers under actual operating conditions, and numerical simulation data provide supplementary data for different working conditions.

[0033] In order to ensure the consistency and adaptability of the data, the collected raw data needs to be preprocessed, including data cleaning (removing outliers), normalization (standardizing data of different dimensions), feature selection (screening out the variables that have the greatest impact on the aerodynamic coefficients) and data enhancement (generating more training samples to improve the generalization ability of the model). These preprocessing steps help improve the efficiency of model training and prediction accuracy.

[0034] Step 2: GBDT model construction and training.

[0035] The GBDT wind load prediction model is constructed using the preprocessed data set; the core of the GBDT model is ensemble learning, which captures the complex nonlinear relationship in the wind load characteristics of the jumper by constructing multiple decision trees. During the model construction process, the hyperparameters of GBDT need to be set, including the number of trees, the depth of the tree, and the learning rate; the specific training steps are as follows: first, the initial decision tree is trained using the preprocessed data to calculate the initial prediction value; then, the residual between the initial prediction value and the true value is calculated; then, the residual is used to train the next decision tree, and the above steps are repeated to gradually reduce the prediction error until all trees are trained; the final model is the weighted sum of all decision trees, which can capture the complex pattern of wind load data.

[0036] Figure 2 The structural diagram of the GBDT (Gradient Boosting Decision Tree) model is shown. The GBDT model gradually improves the prediction accuracy by integrating multiple decision trees. The figure shows the residual calculation process of each tree: starting from the first tree, the residual between the predicted output and the true value is used as the input of the next tree, and so on, until all trees are built. The final prediction result is the accumulation of the outputs of all decision trees. This ensemble learning method can effectively capture complex nonlinear relationships and is particularly suitable for multivariate prediction of jumper wind load characteristics. In this way, the GBDT model can learn from the training data and accurately predict the aerodynamic coefficients under different conditions, significantly improving the accuracy and reliability of the prediction model.

[0037] Step 3: Calculate and analyze wind load characteristics.

[0038] Using the trained GBDT model, the wind load characteristics of the jumper under various wind speeds, wind attack angles and icing conditions are quickly calculated; the model input includes ice type, ice thickness, wind speed, number of sub-conductors and wind attack angle parameters, and the output is the corresponding aerodynamic coefficients, including drag, lift and torque coefficients; the calculation process is as follows: first, according to different input parameters, the model generates the corresponding aerodynamic coefficient prediction values; then, by plotting the curve of the aerodynamic coefficient changing with the wind attack angle, the force characteristics and stability of the jumper under different conditions are analyzed. This method can not only quickly calculate the wind load characteristics of the jumper under various working conditions, but also provide intuitive results for design optimization and safety assessment.

[0039] Figures 3(a), 3(b), and 3(c) show the overall process of the calculation method for the aerodynamic characteristics of the jumper wind load. First, Figure 3(a) shows the distribution of the flow field around the jumper obtained through the wind tunnel test. These experimental data provide a basis for subsequent analysis. Based on the wind tunnel test data and numerical simulation, sub-Figure 3(b) plots the aerodynamic time history curve of the jumper at different time points, reflecting the dynamic changes of the force on the jumper. Then, using the data in the time history curve, the aerodynamic coefficients such as lift, drag and torque are calculated and plotted as a wind angle of attack change curve, as shown in Figure 3(c), to analyze the influence of different wind angles of attack on the aerodynamic characteristics. This method combines experimental and computational means to form a complete aerodynamic coefficient calculation process.

[0040] Table 1 shows the input and output parameters of the GBDT model for the calculation of the wind load characteristics of the jumper. The input parameters include ice type (indicating different ice shapes, such as crescent, D-shaped, etc.), wind speed, ice thickness, number of sub-conductors and wind attack angle. These variables are important factors affecting the aerodynamic characteristics of the jumper. The output parameters include the drag coefficient, lift coefficient and torque coefficient, which are used to describe the aerodynamic performance of the jumper under different conditions. The table clearly lists the symbol, unit and value range of each parameter as the basic data set for model training and prediction. These input data are trained and analyzed by the GBDT algorithm to output the corresponding aerodynamic coefficients, which are used to evaluate the mechanical characteristics of the jumper under various environmental conditions.

[0041] Table 1 Input and output parameters of GBDT model

[0042]

[0043]

[0044] Note: *1 represents crescent-shaped icing, 2 represents D-shaped icing, and 3 represents no icing; # represents the wind attack angle range of 0 to 180°, with an increment of 5°.

[0045] Figure 4The fast calculation model architecture of jumper wind load characteristics based on GBDT intelligent algorithm is shown. The figure includes input layer, hidden layer and output layer, simulating the calculation process of data from input to result output. The input layer receives five main parameters: ice type, ice thickness, wind speed, number of sub-conductors and wind attack angle. These variables are processed by multi-layer neural network and passed to the output layer. The hidden layer contains multiple computing nodes, each of which uses the weights and parameters trained by GBDT algorithm to process the input data, so as to obtain the lift, drag and torque coefficients of wind load. This model architecture shows how to achieve fast and accurate prediction of jumper wind load characteristics through multi-layer nonlinear transformation and GBDT integrated learning method.

[0046] Step 4: Result verification and optimization.

[0047] In order to ensure the prediction accuracy and reliability of the GBDT model, the prediction results of the model need to be verified and optimized. Specifically:

[0048] Compare and analyze the model's prediction results with the actual measured data and calculate the determination coefficient R 2 and mean square error (MSE) indicators to evaluate the accuracy of the model; if the prediction error of the model is large, it is necessary to analyze the source of the error and improve the model performance by adjusting the model parameters or optimizing the algorithm; specific optimization methods include re-tuning the hyperparameters of GBDT (such as the number of trees, depth, and learning rate), increasing the diversity of the training data set (introducing more wind load data under different conditions), and using more complex feature engineering (such as nonlinear feature combinations, interactive features, etc.) to improve the prediction ability of the model.

[0049] Figure 5(a), Figure 5(b), and Figure 5(c) show the accuracy of the prediction model based on the GBDT intelligent algorithm, specifically comparing the relationship between the true value and the predicted value of the aerodynamic coefficient. In each figure, the horizontal axis represents the true value, the vertical axis represents the predicted value, and the fitting line shows the ideal consistency between the predicted value and the true value. 2 0.840, 0.907 and 0.932 respectively. These high R 2 The values ​​show that the GBDT model has high accuracy and reliability in predicting the aerodynamic coefficients and can be effectively applied to the design optimization and operation safety assessment of jumpers in power systems.

[0050] Programming Implementation and Engineering Application

[0051] Develop automated calculation programs based on Python or other programming languages ​​to achieve full-process automation from data input to result output. The key to programmatic implementation is: first, design a user-friendly interface that allows engineers to input different operating parameters (such as wind speed, wind angle of attack, ice type, etc.); then, the program automatically calls the pre-trained GBDT model to calculate the wind load characteristics; finally, generate a detailed calculation report, including the predicted value, change curve and analysis results of the aerodynamic coefficient. This automated calculation program supports batch processing of multiple operating conditions, significantly improves the efficiency and accuracy of wind load characteristic calculations, and provides strong technical support for power system design optimization and operation safety assessment.

[0052] This embodiment provides a method for quickly calculating the wind load characteristics of jumpers based on the GBDT intelligent algorithm. By utilizing the efficiency and accuracy of the GBDT algorithm, combined with actual measurement data and numerical simulation results, the wind load characteristics of jumpers can be quickly and accurately predicted under a variety of complex conditions. This method significantly improves the calculation efficiency, reduces the calculation time and cost, and is suitable for the design optimization and operation safety assessment of jumpers in power systems, especially in engineering applications under complex climatic conditions, and has important practical significance.

[0053] The present invention and its embodiments are described schematically above, and the description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. Therefore, if a person skilled in the art is inspired by it and designs a structural method and an embodiment similar to the technical solution without creativity without departing from the purpose of the invention, they shall all fall within the protection scope of the present invention.

Claims

1. A fast calculation method for wind load characteristics of jumper based on GBDT intelligent algorithm, characterized by: The following steps are involved: Step 1: Data collection and preprocessing; Step 2: GBDT model construction and training; Step 3: Calculation and analysis of wind load characteristics; Step 4: Result verification and optimization.

2. The method for rapidly calculating wind load characteristics of a wire jumper based on the GBDT intelligent algorithm according to claim 1, characterized in that: In step 1, data collection specifically includes: collecting wind load data of the jumper under different wind speeds, wind attack angles and icing conditions through wind tunnel experiments, field measurements and numerical simulations; Wind tunnel experiments provide data on the forces acting on jumpers under different climatic conditions, field measurement data reflect the wind load characteristics of jumpers under actual operating conditions, and numerical simulation data provide supplementary data for different working conditions.

3. The method for rapidly calculating wind load characteristics of a jumper based on the GBDT intelligent algorithm according to claim 2 is characterized in that: In step 1, preprocessing includes data cleaning, normalization, feature selection and data enhancement.

4. The method for rapidly calculating wind load characteristics of a jumper based on the GBDT intelligent algorithm according to claim 3 is characterized in that: In step 2, the preprocessed data set is used to build the GBDT wind load prediction model; during the model construction process, the hyperparameters of GBDT need to be set, including the number of trees, the depth of the tree, and the learning rate; the specific training steps are as follows: first, the initial decision tree is trained using the preprocessed data to calculate the initial prediction value; then, the residual between the initial prediction value and the true value is calculated; then, the residual is used to train the next decision tree, and the above steps are repeated to gradually reduce the prediction error until all trees are trained; the final model is the weighted sum of all decision trees, which can capture the complex patterns of wind load data.

5. The method for rapidly calculating wind load characteristics of a wire jumper based on the GBDT intelligent algorithm according to claim 4 is characterized in that: In step 3, the trained GBDT model is used to quickly calculate the wind load characteristics of the jumper under various wind speeds, wind attack angles and icing conditions; the input of the model includes ice type, ice thickness, wind speed, number of sub-conductors and wind attack angle parameters, and the output is the corresponding aerodynamic coefficients, including drag, lift and torque coefficients; the calculation process is as follows: first, according to different input parameters, the model generates the corresponding aerodynamic coefficient prediction values; then, by drawing the curve of the aerodynamic coefficient changing with the wind attack angle, the force characteristics and stability of the jumper under different conditions are analyzed.

6. The method for rapidly calculating wind load characteristics of a jumper based on the GBDT intelligent algorithm according to claim 5 is characterized in that: In step 4, specifically: Compare and analyze the model's prediction results with the actual measured data and calculate the determination coefficient R 2 and mean square error (MSE) indicators to evaluate the accuracy of the model; if the prediction error of the model is large, it is necessary to analyze the source of the error and improve the model performance by adjusting the model parameters or optimizing the algorithm; specific optimization methods include re-tuning the hyperparameters of GBDT, increasing the diversity of the training data set, and using more complex feature engineering to improve the predictive ability of the model.

Images