A three-span line state intelligent evaluation system based on numerical weather prediction and machine learning

Abstract

The application provides a three-span line state intelligent evaluation system based on numerical weather prediction and machine learning, relates to the technical field of data processing, and comprises the following steps: arranging the vibration mode, frequency and amplitude of a conductor to form a vibration analysis data set for simulating the dynamic behavior of the conductor in a finite element; establishing a finite element model of the three-span line, running finite element analysis, simulating the response of the three-span line under different meteorological conditions and vibration loads, and identifying the key areas and factors causing line failure; and evaluating the safety state of the line according to the key areas and factors causing line failure, and determining the potential failure type. The application can improve the efficiency and accuracy of line state evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a three-span line state intelligent evaluation system based on numerical weather prediction and machine learning. BACKGROUND

[0002] With the rapid development of the power industry, the safe operation of three-span lines (i.e., power lines crossing railways, highways, and important power transmission channels) has become increasingly prominent. The operating state of these lines is directly related to the stability and reliability of the power system, so it is crucial to accurately and efficiently evaluate their state.

[0003] However, traditional three-span line state evaluation methods rely on manual inspection and periodic detection, and thus mainly have the following defects:

[0004] For example, some traditional methods cannot monitor the operating state of the line in real time, making it difficult to discover and handle potential fault risks in a timely manner, thereby increasing the risk of line failure. When analyzing the state of the line, some lack intelligent data processing and analysis methods, making it difficult to deeply mine the internal laws and characteristics of the line operating state. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a three-span line state intelligent evaluation system based on numerical weather prediction and machine learning, which can improve the efficiency and accuracy of line state evaluation.

[0006] To solve the above technical problems, the technical solution of the present application is as follows:

[0007] In a first aspect, a three-span line state intelligent evaluation system based on numerical weather prediction and machine learning includes:

[0008] An identification module is configured to automatically identify and classify abnormal conditions of the line from preprocessed images by constructing, training, and tuning a convolutional neural network model. The abnormal conditions include device damage data, foreign object intrusion data, and temperature anomaly data.

[0009] A meteorological analysis module is configured to obtain meteorological parameters in real time, including temperature, humidity, wind speed, precipitation, air pressure, and radiation. Based on the meteorological parameters, the influence of the line operating state is analyzed to obtain meteorological analysis results, including the influence of temperature, humidity, and wind speed on the line operating state.

[0010] An evaluation module is configured to monitor and identify the vibration mode and characteristics of the three-span line conductor in real time to evaluate the influence of vibration on the line operating state.

[0011] The finite element analysis module is used to organize the equipment damage data, foreign matter intrusion data and temperature anomaly data to form an anomaly condition data set and serve as boundary conditions for finite element analysis; meteorological analysis results obtained based on meteorological parameter analysis are summarized to form a meteorological analysis data set and serve as environmental loads for finite element analysis; vibration modes, frequencies and amplitudes of the conductor are organized to form a vibration analysis data set for simulating dynamic behavior of the conductor in finite element analysis; a finite element model of the three-span line is established, finite element analysis is run, responses of the three-span line under different meteorological conditions and vibration loads are simulated to identify key areas and factors causing line failure; and the safety state of the line is evaluated according to the key areas and factors causing line failure to determine a potential failure type.

[0012] Further, a convolutional neural network model is constructed, trained and optimized to automatically identify and classify abnormal conditions of the line from the preprocessed images, the abnormal conditions including the equipment damage data, the foreign matter intrusion data and the temperature anomaly data, including:

[0013] Real-time image data of the three-span line is collected by the video monitoring device, and image preprocessing is performed to obtain preprocessed images;

[0014] Samples containing equipment damage, foreign matter intrusion and temperature anomaly conditions are screened out from the preprocessed images, and the samples are labeled to create a labeled data set;

[0015] The labeled data set is divided into a training set, a validation set and a test set;

[0016] A convolutional neural network model is constructed, and parameters of the convolutional neural network model including weights and biases are initialized;

[0017] The training set is used to train the convolutional neural network model, a predicted value is calculated through forward propagation, and parameters of the convolutional neural network model are optimized through a back propagation algorithm to obtain a trained convolutional neural network model;

[0018] The test set is used to evaluate the trained convolutional neural network model, and an accuracy rate is calculated; the trained convolutional neural network model is adjusted according to the accuracy rate to obtain a final convolutional neural network model;

[0019] A new preprocessed image is input into the final convolutional neural network model to obtain abnormal conditions of the line, the abnormal conditions of the line including equipment damage, foreign matter intrusion and temperature anomaly.

[0020] Further, the test set is used to evaluate the trained convolutional neural network model, and an accuracy rate is calculated, including:

[0021] Using the trained convolutional neural network model to predict each image in the test set, record the predicted label of each image; compare the predicted label of each image in the test set with the real label;

[0022] Construct a confusion matrix to record the number of real classification and predicted classification of each category, the confusion matrix includes true positive TP, false positive FP, true negative TN and false negative FN;

[0023] According to the confusion matrix, calculate the accuracy rate, that is, the proportion of all correctly predicted samples to the total samples;

[0024] Initialize a list to save the accuracy rate of each iteration; for each of the K subsets, as a validation set, execute the following steps:

[0025] Combine the remaining K-1 subsets to form a temporary "test set"; use the trained convolutional neural network model to predict each image in the temporary "test set", and record the predicted label; compare the predicted label of each image in the temporary "test set" with the real label, and update the confusion matrix; according to the confusion matrix, calculate the accuracy rate of the current iteration, and add it to the accuracy rate record, reset the confusion matrix to prepare for the next iteration; calculate the average accuracy rate of K iterations as the final accuracy rate of the model.

[0026] Further, real-time monitoring and identifying the vibration mode and characteristics of the three-span line conductor to evaluate the impact of vibration on the operation state of the line, including:

[0027] Define the initial temperature T of simulated annealing algorithm o , termination temperature T f , cooling coefficient α and iteration number L at each temperature; set the initial solution S o , that is, the initial position set of the vibration sensor;

[0028] Define the objective function for evaluating the pros and cons of the position set S;

[0029] For each iteration at the current temperature T, generate a new solution S new ;

[0030] Calculate the objective function value f(S new ) of the new solution, if f(S new ) is greater than the objective function value f(S current ) of the current solution, accept the new solution and update S current = S new , where S current represents the current solution; if f(S new ) ≤ f(S current), the new solution is accepted, and after completing the inner loop, the temperature is reduced. If the current temperature T falls below the termination temperature T f Then, the simulated annealing process is stopped, and the final solution, i.e., the final set of positions of the vibration sensor, is output.

[0031] A plurality of vibration sensors are installed at key positions of the three-span line to monitor the vibration data of the conductor in real time. The vibration data is preprocessed to obtain preprocessed vibration data.

[0032] The vibration characteristics of the preprocessed vibration data are extracted using Fourier transform, including frequency, amplitude, and phase.

[0033] Based on the vibration characteristics, a vibration analysis model is established to identify the vibration mode of the conductor and analyze the impact of vibration on the line operation state to obtain vibration analysis results, including the vibration mode, frequency, and amplitude of the conductor.

[0034] Further, based on the vibration characteristics, a vibration analysis model is established to identify the vibration mode of the conductor and analyze the impact of vibration on the line operation state to obtain vibration analysis results, including the vibration mode, frequency, and amplitude of the conductor, including:

[0035] According to the characteristics of the vibration data, the vibration data is labeled to determine and label the corresponding vibration mode label to form a labeled data set.

[0036] The labeled data set is divided into a training set, a validation set, and a test set. The training set is used to train the recurrent neural network model. During the training process, the weights and biases of the recurrent neural network model are continuously adjusted through the backpropagation algorithm and the optimizer to obtain the trained recurrent neural network model.

[0037] The trained recurrent neural network model is used to classify new vibration data to obtain the corresponding predicted vibration mode label.

[0038] By comparing the predicted vibration mode label with the predefined vibration mode label, different vibration modes, including wind vibration and galloping, are identified.

[0039] The correlation between the identified different vibration modes and the line operation state is analyzed.

[0040] Based on the correlation, the risk level of the line under different vibration modes is evaluated.

[0041] Further, according to the characteristics of the vibration data, the vibration data is labeled to determine and label the corresponding vibration mode label to form a labeled data set, including:

[0042] The categories of vibration modes are determined, including wind vibration and galloping.

[0043] defining a unique label for each vibration mode, converting the characteristics of the vibration data, i.e. frequency or amplitude, into genes in the genetic algorithm, which are used to form the chromosomes of the individuals;

[0044] randomly generating an initial population, where each individual represents a vibration mode label sequence;

[0045] extracting key features for each vibration data sample, including time domain and frequency domain features;

[0046] defining a fitness function for evaluating the degree of matching between each individual and the vibration data characteristics;

[0047] According to the fitness function, select the corresponding individual into the next generation, randomly select two individuals as parents, perform cross operation to produce new offspring, randomly mutate the genes of the offspring, i.e. the labels of the vibration modes, repeat the selection, cross and mutation operations until the termination condition is met, when the termination condition is met, stop iteration, and output the final individual as the labeling result of the vibration data;

[0048] decode the final individual into the corresponding vibration mode label sequence, and correspond the vibration mode label sequence with the original vibration data sample to form a labeled data set.

[0049] Further, a finite element model of the three-span line is established, and the finite element analysis is run to simulate the response of the three-span line under different weather conditions and vibration loads, to identify the key areas and factors that cause line failure, including:

[0050] determining the specific structure of the three-span line, including the geometric shape, material and connection method of the conductor, insulator and tower;

[0051] According to the specific structure of the three-span line, create a three-span line geometric model, specify the corresponding material properties for each component in the three-span line geometric model, including elastic modulus, density and Poisson's ratio; define the connection relationship between components, including hinged or fixed connection;

[0052] meshing the three-span line geometric model, setting the boundary conditions of the three-span line geometric model, including fixed constraints and load application positions;

[0053] According to the weather analysis data set, define different weather conditions, including wind speed, wind direction and temperature, and apply them as environmental loads to the three-span line geometric model;

[0054] Using the vibration analysis data set, determine the vibration mode, frequency and amplitude of the conductor, and apply the vibration load to the conductor part;

[0055] Setting the solution parameters, including time step and iteration number, starting the solver, starting the simulation analysis, obtaining the simulation results, including stress distribution contour, deformation graph and vibration response curve;

[0056] According to the simulation results, identify the key areas of stress concentration, large deformation or abnormal vibration response;

[0057] Analysis of the response characteristics of the key areas under different weather conditions and vibration load to determine the factors leading to line failure.

[0058] Further, the three-span line geometry model is meshed, including:

[0059] Set the number of particles, velocity range and position range for the particle swarm optimization algorithm; Set the learning rate and iteration number for the gradient descent algorithm;

[0060] Generate a set of initial particles, each particle representing a meshing scheme, the position of the particle representing the meshing parameters, including mesh size and distribution;

[0061] Evaluate each particle's meshing scheme represented by the evaluation function, update the velocity and position of each particle, repeat the evaluation of fitness and update the particle until the convergence condition is met, to obtain the global final particle;

[0062] Extract the meshing parameters from the global final particle; as the initial scheme for gradient descent;

[0063] Disturb the initial meshing scheme, calculate the gradient of the optimization function with respect to the meshing parameters by finite difference method, adjust the meshing parameters in the direction of gradient descent, repeat the calculation of gradient and adjustment of meshing parameters until the preset iteration number is reached, to obtain the optimized meshing parameters;

[0064] Mesh the three-span line geometry model by the optimized meshing parameters.

[0065] Secondly, a three-span line state intelligent evaluation method based on numerical weather prediction and machine learning, the method comprising the following steps:

[0066] By constructing, training and tuning the convolutional neural network model, automatically identifying and classifying the abnormal conditions of the line from the preprocessed images, including equipment damage data, foreign object intrusion data and temperature anomaly data;

[0067] Real-time acquisition of meteorological parameters, including temperature, humidity, wind speed, precipitation, air pressure and radiation; Based on the meteorological parameters, analyze their influence on the line operation state to obtain the meteorological analysis results, including the influence of temperature, humidity and wind speed on the line operation state;

[0068] Real-time monitoring and identifying the vibration mode and its characteristics of the three-span line conductor to evaluate the influence of vibration on the line operation state;

[0069] The device damage data, foreign matter intrusion data and temperature anomaly data are sorted to form an abnormal condition data set, and are used as boundary conditions for finite element analysis; the meteorological analysis results obtained based on meteorological parameter analysis are summarized to form a meteorological analysis data set, and are used as environmental loads for finite element analysis; the vibration mode, frequency and amplitude of the conductor are sorted to form a vibration analysis data set, which is used to simulate the dynamic behavior of the conductor in finite element; a finite element model of the three-span line is established, and finite element analysis is performed to simulate the response of the three-span line under different meteorological conditions and vibration loads, so as to identify the key areas and factors leading to line failure; the safety state of the line is evaluated according to the key areas and factors leading to line failure, and the potential failure type is determined.

[0070] In a third aspect, a computing device includes:

[0071] One or more processors;

[0072] A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, so that the one or more processors implement the method.

[0073] In a fourth aspect, a computer readable storage medium stores a program, which is executed by a processor to implement the method.

[0074] The above-mentioned scheme of the present application at least includes the following beneficial effects:

[0075] By constructing, training and optimizing the convolutional neural network model, the system can automatically identify and classify the abnormal conditions of the line, such as device damage, foreign matter intrusion and temperature anomaly, etc. This way not only improves the accuracy of identification, but also greatly improves the identification efficiency, which helps to discover and handle line problems in time.

[0076] It can monitor meteorological parameters in real time and analyze their influence on the line operation state, which enables the system to respond quickly in adverse weather conditions, providing strong support for preventing line failure.

[0077] By real-time monitoring and identifying the vibration mode and its characteristics of the three-span line conductor, the influence of vibration on the line operation state can be evaluated, which helps to discover potential mechanical failure or structural problems, so as to repair and replace in time.

[0078] Combined with equipment damage data, foreign matter intrusion data, temperature anomaly data, weather analysis results and conductor vibration analysis, the system provides comprehensive input conditions for finite element analysis of the line, which makes the evaluation results more accurate and helps to identify the key areas and factors that cause line failure.

[0079] Through finite element analysis, the system can simulate the response of the three-span line under different weather conditions and vibration loads, so as to timely find and handle potential safety hazards, which not only prolongs the service life of the line, but also significantly reduces the probability of failure, and improves the stability and safety of the entire power grid system. BRIEF DESCRIPTION OF DRAWINGS

[0080] Figure 1 is a schematic diagram of a three-span line state intelligent evaluation system based on numerical weather prediction and machine learning provided by an embodiment of the present application.

[0081] Figure 2 is a flowchart of a three-span line state intelligent evaluation method based on numerical weather prediction and machine learning provided by an embodiment of the present application. DETAILED DESCRIPTION

[0082] The exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0083] As shown in Figure 2 , an embodiment of the present application proposes a three-span line state intelligent evaluation method based on numerical weather prediction and machine learning, comprising:

[0084] Step 1, by constructing, training and optimizing a convolutional neural network model, to automatically identify and classify the abnormal conditions of the line from the preprocessed images, the abnormal conditions including equipment damage data, foreign matter intrusion data and temperature anomaly data;

[0085] Step 2, real-time acquisition of weather parameters, including temperature, humidity, wind speed, precipitation, air pressure and radiation; based on the weather parameters, analyze their impact on the line operation state to obtain weather analysis results, including the impact of temperature, humidity and wind speed on the line operation state;

[0086] Step 3, real-time monitoring and identification of the vibration mode and characteristics of the three-span line conductor to evaluate the impact of vibration on the line operation state;

[0087] Step 4: Organize the device damage data, foreign object intrusion data, and temperature anomaly data to form an abnormal condition dataset, which serves as the boundary conditions for finite element analysis; summarize the meteorological analysis results based on meteorological parameter analysis to form a meteorological analysis dataset, which serves as the environmental load for finite element analysis; organize the vibration mode, frequency, and amplitude of the conductor to form a vibration analysis dataset, which is used to simulate the dynamic behavior of the conductor in finite element analysis; establish a finite element model of the three-span line, run the finite element analysis, and simulate the response of the three-span line under different meteorological conditions and vibration loads to identify the key areas and factors that cause line failure; evaluate the safety status of the line based on the key areas and factors that cause line failure, and determine the potential failure type.

[0088] In the embodiments of the present application, through the convolutional neural network model, the system can automatically identify and classify the abnormal conditions of the line from the preprocessed images, greatly improving the efficiency and accuracy of anomaly detection. Compared with traditional manual inspection, the introduction of machine learning model reduces the misjudgment and missed detection caused by human factors, enhances the reliability of anomaly detection, and accumulates abnormal condition data for further analysis and prediction, providing data support for line maintenance and management. Real-time acquisition of meteorological parameters and analysis of their impact on line operation status can help respond to sudden weather events in a timely manner. Through analysis of meteorological parameters, potential problems that may occur under certain weather conditions can be predicted in advance, thereby enabling preventive maintenance. In extreme weather conditions, the system can issue warnings in advance to help operators take measures to reduce line failures and improve the safety of the power grid.

[0089] By monitoring the vibration mode of the conductor, the system can identify abnormal vibration characteristics, which may indicate potential mechanical failure or structural problems. Continuous monitoring and analysis of vibration patterns can help identify and address potential problems in a timely manner, thereby extending the service life of the line. Understanding the vibration characteristics of the conductor can help operators develop more refined maintenance strategies and improve maintenance efficiency. By integrating multiple data sources (abnormal condition data, meteorological analysis data, and vibration analysis data), the system can provide more comprehensive and accurate line status assessment. Finite element analysis can help identify key areas and factors that cause line failure, thereby enabling targeted prevention and maintenance. Through continuous monitoring and evaluation of line status, operators can more efficiently manage the power grid, reducing unnecessary downtime and maintenance costs.

[0090] In a preferred embodiment of the present application, step 1 above involves constructing, training, and tuning a convolutional neural network model to automatically identify and classify the abnormal conditions of the line from preprocessed images. The abnormal conditions include device damage data, foreign object intrusion data, and temperature anomaly data, which can include:

[0091] Step 11, collect real-time image data of the three-span line through video monitoring equipment and perform image preprocessing to obtain preprocessed images, specifically including: using video monitoring equipment installed near the three-span line, such as cameras, to capture image data of the line in real time, these devices can be set to continuous recording or timed recording to ensure that all important events are captured; pre-process the collected raw images to improve image quality and prepare for subsequent analysis, the preprocessing steps may include denoising, contrast enhancement, edge detection, etc., these operations help highlight key features in the image, such as device damage, foreign matter or temperature abnormality areas.

[0092] Step 12, screen samples containing device damage, foreign matter intrusion and temperature abnormality conditions from preprocessed images, and label samples to create a labeled dataset, specifically including: from the preprocessed images, screen samples containing device damage, foreign matter intrusion and temperature abnormality conditions, this can be done by observing specific patterns or abnormalities in the image; label the screened samples, labeling usually involves adding a bounding box or mask to the abnormal area in the image and assigning it a corresponding label (such as "device damage", "foreign matter intrusion" or "temperature abnormality"), combine all labeled samples into a labeled dataset.

[0093] Step 13, divide the labeled dataset into training set, validation set and test set, specifically including: randomly divide the labeled dataset into three subsets: training set, validation set and test set, the training set is used to train the model, the validation set is used to adjust the model parameters and hyperparameters, and the test set is used to evaluate the final performance of the model; according to the specific requirements of the project and the total amount of available data, reasonably allocate the proportion of the three subsets, the common division ratio can be 70% (training set), 15% (validation set) and 15% (test set).

[0094] Step 14, build a convolutional neural network model and initialize the convolutional neural network model parameters, including weights and biases, specifically including:

[0095] Determine the size of the input image, for example, if the pre-processed image size is 224x224 pixels, the input layer should accept images of this size; add multiple convolutional layers to extract image features, select appropriate convolution kernel size (such as 3x3, 5x5) and stride for each convolutional layer; set the number of output channels for each convolutional layer, i.e. the number of convolution kernels; apply an activation function such as ReLU (Rectified Linear Unit) after each convolutional layer to increase the nonlinearity of the network; insert a pooling layer (such as max pooling) between consecutive convolutional layers to reduce the dimensionality of the feature map and computational complexity, set the size and stride of the pooling window. Add a fully connected layer at the end of the network to integrate the features extracted earlier and perform the final classification, determine the number of neurons in the fully connected layer, usually the number of neurons in the last layer is equal to the number of classes in the classification task; use the softmax function as the activation function of the output layer to get the prediction probability of each class. For the weights of the convolutional and fully connected layers, you can choose a random initialization method such as He initialization to adjust the initial distribution of the weights according to the number of neurons in the layer; the bias parameter is usually initialized to zero or a small number close to zero, because the main function of the bias is to adjust the offset of the output value.

[0096] Step 15, use the training set to train the convolutional neural network model, calculate the prediction value by forward propagation, and optimize the convolutional neural network model parameters by back propagation algorithm to get the trained convolutional neural network model, specifically including: input the image in the training set into the convolutional neural network model, calculate through each layer of the model to get the prediction probability of each class; compare the prediction result of the model with the true label of the image, calculate the loss function (such as cross-entropy loss); use gradient descent algorithm to update the weight and bias parameters of the model according to the gradient of the loss function, repeat the above steps until the performance of the model on the validation set reaches the preset standard or reaches the maximum number of training rounds.

[0097] Step 16, use the test set to evaluate the trained convolutional neural network model, calculate the accuracy; according to the accuracy, adjust the trained convolutional neural network model to get the final convolutional neural network model, specifically including: use the test set to evaluate the trained convolutional neural network model, calculate the accuracy of the model; according to the accuracy, adjust the model, such as changing the network structure, increasing or decreasing the number of layers, adjusting the learning rate and other hyperparameters.

[0098] Step 17, input the preprocessed new image into the final convolutional neural network model to obtain the abnormal condition of the line, the abnormal condition of the line including equipment damage, foreign matter intrusion and temperature anomaly, specifically including: performing the same preprocessing operation on the newly collected three-span line image as step 11; input the preprocessed new image into the final convolutional neural network model to obtain the prediction result of the model on the line abnormal condition, and display the prediction result of the model in a visual or other form for further analysis and processing by the operator.

[0099] In the embodiments of the present application, image data is collected in real time by the video monitoring device, ensuring the timeliness and accuracy of the information. Image preprocessing can remove noise and enhance image features, providing higher quality input for subsequent automatic recognition and classification. By filtering samples of specific abnormal conditions, a dataset focused on line abnormal conditions can be created, improving the relevance and efficiency of model training. Accurate labeling of samples provides reliable labels for supervised learning, which helps the model accurately identify and classify abnormal conditions. By dividing the dataset into training, validation and test sets, the generalization ability of the model can be more effectively evaluated to prevent overfitting. An independent test set can provide an unbiased estimate of the model's performance. A convolutional neural network model specifically designed for identifying line abnormal conditions can better adapt to this specific task. Through training, the model can learn to extract useful features from images and accurately classify abnormal conditions. The backpropagation algorithm can effectively adjust the model parameters to improve the prediction performance of the model. Using an independent test set to evaluate the model's performance ensures the model's reliability in actual applications. Based on the test results, the model can be adjusted to further improve its accuracy and generalization ability. Inputting new images into the trained model can automatically detect the abnormal condition of the line, greatly improving the detection efficiency. By monitoring the abnormal condition of the line in real time, problems can be discovered and addressed in a timely manner, ensuring the safe and stable operation of the power grid.

[0100] In a preferred embodiment of the present application, step 16, using the test set to evaluate the trained convolutional neural network model, calculating the accuracy, can include:

[0101] Step 161, using the trained convolutional neural network model to predict each image in the test set, record the predicted label of each image; compare the predicted label of each image in the test set with the true label, specifically including: loading the test set containing preprocessed images, these images have true labels; using the trained convolutional neural network model to predict each image in the test set, the prediction is completed by forward propagation, that is, the input image passes through each layer of the network to get the output; for each image, convert the model's predicted output (category probability) into a specific category label and record these predicted labels; compare the predicted label of each image with its corresponding true label to determine whether the prediction is correct.

[0102] Step 162, construct a confusion matrix to record the number of true and predicted categories for each category, the confusion matrix includes true positives TP, false positives FP, true negatives TN and false negatives FN, specifically including: creating a confusion matrix with a size equal to the number of categories multiplied by the number of categories, each cell represents a combination of true and predicted labels for a category (such as TP, FP, TN, FN); update the corresponding cell in the confusion matrix according to the comparison results of true and predicted labels in step 161. For example, if the true label of a sample is category A and the predicted label is also category A, then the corresponding position (A, A) in the confusion matrix is increased by 1, indicating that the number of true positives (TP) is increased.

[0103] Step 163, according to the confusion matrix, calculate the accuracy rate, that is, the proportion of all correctly predicted samples to the total number of samples, specifically including: traversing the diagonal elements of the confusion matrix, these elements represent the number of correctly predicted samples for each category (i.e. true positives TP), summing all values on the diagonal to get the total number of correctly predicted samples; divide the number of correctly predicted samples by the total number of samples in the test set to get the accuracy rate, which represents the proportion of correctly predicted samples in all test samples.

[0104] Step 164, initialize a list to save the accuracy rate of each iteration; for each of the K subsets, denoted as the validation set, perform the following steps:

[0105] Combine the remaining K-1 subsets to form a temporary "test set"; use the trained convolutional neural network model to predict each image in the temporary "test set" and record the predicted label; compare the predicted label of each image in the temporary "test set" with the true label and update the confusion matrix; calculate the accuracy rate of the current iteration according to the confusion matrix and add it to the accuracy rate record, reset the confusion matrix to prepare for the next iteration; calculate the average accuracy rate of K iterations as the final accuracy rate of the model, specifically including:

[0106] Create an empty list to store the accuracy of each iteration; determine the value of K, the number of data subsets, randomly divide the entire dataset into K equal-sized subsets, use a loop structure such as a for loop to iterate through the K subsets, in each iteration, select a subset as the validation set and combine the remaining K-1 subsets into a temporary "test set". Create an empty data structure (such as a list or array) to store the images and labels of the temporary "test set", iterate through all subsets except the current validation set and add their images and labels to the temporary "test set". Use the trained convolutional neural network model to make predictions on each image in the temporary "test set", record the predicted label for each image. Compare the predicted label of each image in the temporary "test set" with its corresponding true label, update the confusion matrix according to the comparison result, for each sample, according to its true label and predicted label, add 1 to the corresponding position of the confusion matrix; calculate the accuracy according to the confusion matrix, the accuracy is the sum of the diagonal elements of the confusion matrix (true positives TP) divided by the total number of samples in the test set, add the calculated accuracy to the accuracy record list initialized earlier.

[0107] After each iteration, reset the confusion matrix to a zero matrix to prepare for the next iteration, this can be achieved by reinitializing the confusion matrix or using a clear operation; after K iterations are completed, calculate the average of all accuracies in the accuracy record list, this average represents the final performance evaluation result of the model under K-fold cross-validation.

[0108] In the embodiments of the present application, by predicting each image in the test set and comparing with the true label, the performance of the model on unseen data can be comprehensively evaluated. The confusion matrix provides an intuitive way to visualize the classification performance of the model, which can clearly show the misclassification between categories; accuracy is one of the important indicators to evaluate the performance of classification models, which can intuitively reflect the proportion of correct predictions, accuracy can be used as a benchmark for performance comparison between different models or different iteration versions. By initializing a list to save the accuracy of each iteration, the performance of the model on different validation sets can be tracked, and the accuracy records of multiple iterations help to analyze the stability and generalization ability of the model; K-fold cross-validation can ensure that each sample in the dataset is used as a test set at least once, thus making full use of limited data resources, through multiple iterations and different data division, the performance robustness of the model under different data distribution can be evaluated, a single data division may introduce accidental errors, K-fold cross-validation can reduce the impact of such accidental errors on model evaluation, by calculating the average accuracy of K iterations, a more reliable and stable model performance estimate value can be obtained.

[0109] In a preferred embodiment of the present application, the above-mentioned step 2, real-time acquisition of meteorological parameters, including temperature, humidity, wind speed, precipitation, air pressure and radiation; based on the meteorological parameters, analyze their influence on the line operation state to obtain meteorological analysis results, including the influence of temperature, humidity and wind speed on the line operation state, which can include:

[0110] Determine reliable meteorological data sources, which can be local weather stations, online weather service APIs or satellite data. According to the selected data source, set up the corresponding data collection system, if it is from the weather station to obtain data, it may be necessary to install data receiving equipment; if it is from online API, it is necessary to write code to request data regularly. Ensure that the system can receive meteorological data in real time, which usually involves obtaining the latest meteorological parameters from the data source regularly (such as every minute, every hour); preprocess the received raw data as needed, such as format conversion, unit unification, outlier processing, etc., to ensure the accuracy and usability of the data.

[0111] Understand the type of line (such as overhead line, cable, etc.) and its material, design specification, operating environment, etc. characteristics, on this basis, identify the meteorological parameters that may have a significant impact on the line operation state, here we mainly focus on temperature, humidity and wind speed; determine the purpose of analysis, that is, to explore how these meteorological parameters affect the operation state of the line individually or jointly; ensure that the collected meteorological data is consistent with the line operation state data in time, so as to accurately analyze the relationship between the two, compare the meteorological data and line state data at different time points, for example, in high temperature, high humidity or strong wind weather, whether the operation state of the line has changed obviously. Analyze how temperature affects the conductivity of the line, for example, high temperature may increase the resistance of the line, thereby reducing the transmission efficiency, investigate the influence of temperature on the thermal expansion of line materials, which may change the tension and sag of the line, thereby affecting the safety distance, study how humidity affects the insulating material of the line, high humidity environment may cause water film on the surface of the insulating layer, reduce the insulation resistance and increase the risk of electric leakage. Analyze whether humidity accelerates the corrosion process of line materials, which may long-term affect the structural strength and conductivity of the line. Explore the influence of wind speed on the mechanical stability of the line, strong wind may cause the line to sway or even dance, increasing the risk of line-to-line contact or ground short circuit, analyze the influence of wind speed change on the wind load on the line, which is directly related to the tension design of the line and the stability of the support structure.

[0112] Record the findings, observations and conclusions in the above analysis process in detail, including the specific changes of line operation state when each meteorological parameter changes, use visualization tools such as charts and curves to show the relationship between meteorological parameters and line state, which is convenient for more intuitive understanding of the analysis results.

[0113] According to the analysis results, targeted line maintenance and operation suggestions under different weather conditions are proposed. For example, in high temperature weather, the inspection frequency of the line may need to be increased to ensure the normal work of the heat dissipation equipment; in high humidity environment, the condition of the insulating material needs to be paid attention to, and the damaged parts need to be replaced in time; in windy weather, the tension setting of the line may need to be adjusted to enhance its wind resistance.

[0114] In a preferred embodiment of the present application, the step 3 of monitoring and identifying the vibration mode and characteristics of the three-span line conductor in real time to evaluate the influence of vibration on the operation state of the line can include:

[0115] Step 31, define the initial temperature T of simulated annealing algorithm o , the final temperature T f , the cooling coefficient a and the iteration number L at each temperature; set the initial solution S o , that is, the initial position set of the vibration sensor;

[0116] Step 32, define the objective function for evaluating the pros and cons of the position set S, wherein the calculation formula of the objective function f(S) is:

[0117]

[0118] Wherein λ1, λ2, λ3, λ4, λ5 and λ6 are weight coefficients; |X(f i )| 2 represents the signal intensity corresponding to each frequency f i ; N is the total number of frequency points, which represents the number of frequency components considered in the calculation; SNR(f i ) is the signal-to-noise ratio of each frequency point (f i ); each sensor is circular with a radius of r i , then the coverage area of each sensor is The total coverage area of all sensors is obtained by summation; K represents the total number of sensors, which represents the number of sensors in the layout; O(S i ,S j ) is the area of the overlapping part between sensors S i and S j ; i≠j represents all different sensor pairs; N t represents the number of iterations in the optimization process, which represents the number of steps the algorithm needs to execute; T maxThis represents the maximum allowable computation time. By optimizing signal strength, signal-to-noise ratio, and frequency response, the system's signal quality is ensured to reach its optimal level. Reducing sensor overlap and optimizing sensor layout improves coverage efficiency. At the same time, it reduces redundant signals and resource waste. By controlling the number of computation iterations and the maximum computation time, computational efficiency is improved and the computational burden on the system is reduced. These optimizations help improve system performance, reduce resource waste, and improve the efficiency of signal reception and processing, ensuring that the best network configuration or sensor layout scheme can be obtained under limited time and resources.

[0119] Step 33: For each iteration at the current temperature T, generate a new solution S. new Specifically, this includes: at the current temperature, the algorithm will perform L iterations. In each iteration, the algorithm will adjust the current solution S. current Generate a new solution S new This typically involves minor adjustments or relocations of the sensor's position.

[0120] Step 34, calculate the objective function value f(S) of the new solution. new If f(S) new The objective function value f(S) is greater than the current solution's objective function value. current If the solution is correct, then accept the new solution and update S. current =S new , of which S current Indicates the current solution; if f(S) new )≤f(S current If the current temperature T drops to the termination temperature T, then accept the new solution, complete the inner loop, and lower the temperature. f The simulation annealing process is then stopped, and the final solution, i.e., the final set of vibration sensor positions, is output. Specifically, this includes calculating the new solution S. new The objective function value f(S) new If f(S) new The objective function value f(S) is greater than the current solution's objective function value. current If the solution is f(S), then the new solution is accepted unconditionally, and the current solution is updated to f(S). new )≤f(S current If f(S) new ) less than or equal to f(S) current If the algorithm encounters a new solution, it determines whether to accept it. This allows the algorithm to avoid getting trapped in local optima during the search process. After L iterations, the temperature T decreases according to the cooling coefficient α. When the temperature drops to the termination temperature T... f When the algorithm stops, it outputs the final set of sensor locations.

[0121] Step 35, install multiple vibration sensors at key locations of the three-span line to monitor the vibration data of the conductor in real time, and pre-process the vibration data to obtain pre-processed vibration data, specifically including: according to the final sensor position set obtained by the simulated annealing algorithm, install vibration sensors at key locations of the three-span line, these sensors will monitor the vibration data of the conductor in real time and transmit the data to the analysis system, and pre-process the collected vibration data, such as filtering, denoising, etc., to obtain clearer vibration signals.

[0122] Step 36, extract vibration features of pre-processed vibration data using Fourier transform, vibration features include frequency, amplitude and phase, specifically including:

[0123] Convert time-domain vibration data into frequency-domain data through fast Fourier transform (FFT); FFT is used to calculate the Fourier transform of discrete signals, which can decompose signals into combinations of different frequency sine and cosine waves; the frequency-domain data obtained by FFT can determine the frequency components present in the vibration signal, which correspond to the periodic components in the signal and are the key to identifying vibration patterns. For each frequency component, extract its corresponding amplitude value, which represents the strength or amplitude of the corresponding frequency component in the vibration signal and is an important indicator for evaluating the distribution of vibration energy. In addition to frequency and amplitude, phase information can be extracted from FFT results, which describes the temporal relationship between frequency components and helps understand the waveform structure and synchronization between vibration signals. Organize the extracted frequency, amplitude and phase features into a structured dataset, which can include creating a feature vector, where each element represents a specific feature value (such as the amplitude of a certain frequency).

[0124] Step 37, based on vibration features, establish a vibration analysis model, identify the vibration pattern of the conductor, analyze the impact of vibration on the operation state of the line, and obtain vibration analysis results, including the vibration pattern, frequency and amplitude of the conductor.

[0125] In the embodiments of the present application, the optimal positions of the vibration sensors are determined by the simulated annealing algorithm, which can ensure that the sensor layout can capture the key information of the conductor vibration, improve the efficiency and accuracy of data collection. By installing vibration sensors at key positions, the vibration of the conductor can be monitored in real time, providing data support for timely detection of potential line problems; by preprocessing and Fourier transform of the vibration data, the key features of vibration such as frequency, amplitude and phase can be accurately extracted, which provides an important basis for subsequent analysis. The vibration analysis model established based on the vibration characteristics can effectively identify different vibration modes of the conductor, which helps to understand the dynamic behavior of the line under different environmental conditions. Through the vibration analysis results, the specific influence of vibration on the operation state of the line can be evaluated, including the risk of fatigue damage, loosening or fracture, etc., so that necessary maintenance measures can be taken in time. According to the vibration analysis results, more accurate preventive maintenance plans can be developed to prolong the service life of the line and reduce the occurrence of unexpected power outages. Through continuous monitoring and analysis of the conductor vibration, potential safety hazards can be detected and solved in time, thereby improving the stability and reliability of the entire power system.

[0126] In a preferred embodiment of the present application, step 37, based on the vibration characteristics, establishes a vibration analysis model to identify the vibration mode of the conductor and analyze the influence of vibration on the operation state of the line to obtain vibration analysis results, the vibration analysis results include the vibration mode, frequency and amplitude of the conductor, which can include:

[0127] Step 371, according to the characteristics of the vibration data, the vibration data is labeled, and the corresponding vibration mode label is judged and labeled to form a labeled data set;

[0128] Step 372, divide the labeled data set into training set, validation set and test set, use the training set to train the recurrent neural network model, in the training process, continuously adjust the weights and biases of the recurrent neural network model through the back propagation algorithm and the optimizer to obtain the trained recurrent neural network model, specifically including: divide the labeled data set into training set, validation set and test set according to a certain proportion (such as 70:15:15), select a suitable recurrent neural network (RNN) structure, such as long short-term memory network (LSTM); initialize the weight and bias parameters of the RNN model, use the training set data to train the RNN model, the input is the vibration feature, and the output is the corresponding vibration mode label, in the training process, calculate the gradient of the loss function with respect to the model parameters through the back propagation algorithm, use the optimizer (such as Adam) to update the weights and biases of the RNN model according to the gradient information, repeat the steps until the performance of the model on the validation set reaches the preset standard or the training round number reaches the upper limit.

[0129] Step 373, using the trained recurrent neural network model to classify the new vibration data to obtain the corresponding predicted vibration mode label, specifically including: obtaining new unlabeled vibration data, using the trained RNN model to classify and predict the new data, and outputting the predicted vibration mode label.

[0130] Step 374, by comparing the predicted vibration mode label with the predefined vibration mode label, identifying different vibration modes, including wind-induced vibration and dance, specifically including:

[0131] A set of vibration mode labels is predefined, which is usually created based on historical data. The predefined vibration mode labels include "wind-induced vibration" and "dance" mode. Each label corresponds to a specific vibration mode, which differs in physical characteristics, frequency range, amplitude, etc.; using the previously trained recurrent neural network (RNN) model, the new, unlabeled vibration data is classified and predicted, and the RNN model will output the predicted label corresponding to each data segment, which represents the vibration mode that the model believes the data segment belongs to. Compare the predicted vibration mode label with the predefined vibration mode label. The comparison process can be string matching; if the predicted label completely matches a predefined label, it is considered to have successfully identified the vibration mode; for example, if the predicted label is "wind-induced vibration", and there is a completely identical label in the predefined label set, then it can be determined that the vibration mode corresponding to the data segment is wind-induced vibration; for the predicted label that does not match or has a similarity lower than the threshold, further analysis or classification as "unknown vibration mode".

[0132] Step 375, analyze the correlation between the identified different vibration modes and the line operation state, specifically including: analyzing the trend of current, voltage, temperature and other line operation state data in different time periods, especially under the identified specific vibration mode, whether these data appear significant abnormal fluctuations; for each identified vibration mode, extract the feature parameters of the corresponding line operation state data, such as current peak, voltage valley, average temperature and their fluctuation range, etc. These feature parameters will help reveal the internal relationship between vibration mode and line state; apply Pearson correlation coefficient formula to calculate the correlation coefficient between each vibration mode and the corresponding line operation state data. Pearson correlation coefficient can provide a quantitative index to measure the strength and directionality of the linear relationship between two variables.

[0133] Step 376, according to the specific value of the correlation coefficient, evaluate the risk level of the line under different vibration modes, specifically including:

[0134] According to the specific value of the correlation coefficient, by A risk index R is calculated, where C represents the correlation coefficient; D represents the duration; a, b, y, d, and h are weight coefficients; A represents the amplitude; F represents the frequency; I represents the ideal frequency; M represents another reference frequency; S1 and S2 represent the standard deviation of the Gaussian function;

[0135] The risk index R is a comprehensive indicator, the higher the value of R, the greater the risk of the line under the current vibration mode; a risk threshold is set, which is a critical value for dividing different risk levels, and the risk threshold is set based on historical data; for example, three risk levels can be set: low risk, medium risk and high risk, each level corresponds to a range of R values. Risk levels are divided:

[0136] Low risk, when the R value is lower than a certain set low risk threshold, it is considered that the line has a low risk under the current vibration mode, which means that the operation state of the line is relatively stable and the possibility of problems is small.

[0137] Medium risk, when the R value is between the low risk and high risk thresholds, the line is identified as medium risk. This indicates that the line is still within the safe operating range, but there are certain potential problems that need to be strengthened monitoring and maintenance.

[0138] High risk, if the R value exceeds the set high risk threshold, it means that the line is in a high risk state under the current vibration mode, which may mean that the line has serious safety hazards and needs to be immediately intervened to prevent possible accidents.

[0139] For lines with different risk levels, corresponding measures need to be developed. For example, low-risk lines can be regularly checked; medium-risk lines may need to increase the frequency of checks or perform certain preventive maintenance; high-risk lines may need to be immediately shut down for maintenance or take other emergency measures.

[0140] In the embodiments of the present application, by annotating the vibration data and training the recurrent neural network model, the vibration mode of the conductor can be more accurately identified. Compared with the traditional manual identification method, this method based on big data and machine learning has higher accuracy and efficiency. Using the trained recurrent neural network model, automatic classification of new vibration data can be realized, reducing the need for manual intervention and improving the automation and intelligence level of processing. By real-time monitoring and identifying the vibration mode of the conductor, abnormal vibration conditions such as wind vibration and dancing can be detected in time, so that timely warning and prevention can be carried out to avoid line faults. Analyzing the correlation between different vibration modes and line operation states can provide important decision support for the operation and maintenance of the power system. Based on this correlation, the risk level of the line under different vibration modes can be evaluated, which helps to develop reasonable maintenance and overhaul plans. Through accurate identification of vibration modes and risk assessment, potential safety hazards can be detected and handled in time, thereby improving the reliability and stability of the entire power system. Through the intelligent monitoring and warning system, unnecessary inspection and maintenance work can be reduced, thereby reducing the operation and maintenance cost of the power system.

[0141] In a preferred embodiment of the present application, step 371, according to the characteristics of the vibration data, the vibration data is annotated, and the corresponding vibration mode label is judged and annotated to form an annotated data set, including:

[0142] Step 3711, determine the category of vibration mode, including wind vibration and dancing, specifically including: sufficient research and analysis of possible vibration modes in the target field, which includes reading relevant literature, communicating with field experts, etc.; on the basis of research, determine two main vibration modes: wind vibration and dancing, which should have obvious distinction in physical characteristics, causes, etc.; record the determined vibration mode categories and their definitions in the document.

[0143] Step 3712, define a unique label for each vibration mode, and convert the characteristics of the vibration data, i.e. frequency or amplitude, into genes in genetic algorithm, which are used to form the chromosomes of individuals, specifically including: design a unique label for each vibration mode (wind vibration and dancing). For example, numbers or specific strings can be used to represent different modes; select features from vibration data that can represent different vibration modes, such as frequency and amplitude; convert these feature values into genes in genetic algorithm, which usually involves mapping data values to a certain coding range, such as binary coding; multiple genes form the chromosomes of individuals, each chromosome represents a possible vibration mode label sequence.

[0144] Step 3713, randomly generate an initial population, wherein each individual represents a vibration mode label sequence, specifically including: determining the size of the initial population according to the complexity of the problem and the computing resources, using a random number generator to generate the initial population according to the structure of the chromosome, each individual is a randomly generated vibration mode label sequence, initializing the fitness and other record information of each individual in the population.

[0145] Step 3714, extract key features for each vibration data sample, including time domain and frequency domain features, specifically including: performing necessary preprocessing on the original vibration data, such as denoising, normalization, etc., calculating key features of the vibration data in the time domain, such as mean, variance, peak value, etc., converting the vibration data to the frequency domain through the Fourier transform method, and extracting key features such as dominant frequency, spectral energy distribution, etc., integrating the extracted time domain and frequency domain features into a feature vector.

[0146] Step 3715, define a fitness function for evaluating the matching degree of each individual with the vibration data features, wherein the calculation formula of the fitness function is:

[0147]

[0148] Wherein, μ f (I j ) is the expected frequency value corresponding to label I j ; σ f (I j ) is the standard deviation of the frequency feature; F(I) represents the fitness value of individual I, ranging from 0 to 1, the higher the value, the better the matching degree; N is the total number of vibration data samples; I j represents the label corresponding to the jth data sample in individual I; and represent the frequency feature, amplitude feature and time domain feature of the jth data sample, respectively; μ a (I j ) is the expected amplitude value corresponding to label I j ; σ a (I j ) is the standard deviation of the amplitude feature; μ t (I j ) is the expected time domain feature value corresponding to label I j ; σ t (I j ) is the standard deviation of the time domain feature; w f , w a and w t are the weight coefficients of the frequency, amplitude and time domain feature matching degrees, respectively, satisfying w f +w a +w t =11.

[0149] In step 3716, according to the fitness function, select the corresponding individual into the next generation, randomly select two individuals as parents, perform cross operation to generate new offspring, randomly mutate the label of the gene, i.e. the vibration mode, of the offspring, repeat the selection, cross and mutation operation until the termination condition is met, when the termination condition is reached, stop iteration, and output the final individual as the labeling result of the vibration data, specifically including: according to the result of the fitness function, select the individual with high fitness into the next generation, which can be realized by roulette selection, tournament selection and other methods, randomly select two individuals as parents, perform cross operation to generate new offspring, the cross point can be randomly selected; randomly mutate the gene of the newly generated offspring to increase the diversity of the population, set the iteration termination condition, such as reaching the maximum iteration number, the fitness reaching the preset threshold, etc., when the termination condition is met, stop iteration, after iteration, output the individual with the highest fitness as the labeling result of the vibration data.

[0150] In step 3717, decode the final individual into the corresponding vibration mode label sequence, and correspond the vibration mode label sequence with the original vibration data sample to form a labeled data set, specifically including: decoding the final individual into the corresponding vibration mode label sequence, which involves converting the encoding in the genetic algorithm back to the original vibration mode label; correspond the decoded vibration mode label sequence with the original vibration data sample to ensure that each data sample has its corresponding label; organize all labeled data samples to form a complete labeled data set.

[0151] In the embodiments of the present application, the labeling process of vibration modes is optimized by genetic algorithm, which can more accurately identify and label different vibration modes (such as wind-induced vibration and dance). This method can effectively utilize the characteristics of vibration data such as frequency and amplitude, thereby improving the accuracy of labeling. This method realizes a certain degree of automation, reduces the need for manual intervention and manual labeling, and through the iterative optimization process, the algorithm can automatically find the label that best matches the characteristics of the vibration data, thereby improving the efficiency of labeling. Since genetic algorithm is used to find the optimal solution by simulating natural selection and genetic mechanisms, this method has strong robustness in dealing with complex and nonlinear problems, and through continuous iteration and optimization, the algorithm can learn the internal laws and characteristics of the data, thereby having good generalization ability and being able to handle different types of vibration data. This method is not only suitable for labeling of wind-induced vibration and dance, but also can be extended to more types of vibration modes by defining new labels and fitness functions, so that new vibration modes can be easily included in the labeling system. Through accurate vibration mode labeling, a high-quality labeled data set can be formed, which is crucial for subsequent machine learning, pattern recognition and other tasks, and helps to improve the performance and accuracy of related applications.

[0152] The fitness function provides a comprehensive evaluation method by combining the matching degree of frequency, amplitude and time domain characteristics, ensuring multi-dimensional matching between individuals and vibration data characteristics, thereby improving the accuracy of labeling. By adjusting the weight coefficients w f , w a and w t , the importance of certain features can be emphasized or attenuated according to actual needs, and this flexibility enables the fitness function to adapt to different application scenarios and data characteristics. The use of Gaussian function to calculate the similarity between features and labels makes the function have a certain tolerance to small changes in features, enhancing the robustness of the algorithm, which means that even if the data has a certain degree of noise or fluctuations, the algorithm can still effectively identify the vibration mode label that best matches it.

[0153] In a preferred embodiment of the present application, the device damage data, foreign object intrusion data and temperature anomaly data in step 4 above are sorted to form an abnormal condition data set, which is used as the boundary condition of finite element analysis, which can include:

[0154] According to the type of anomaly (such as equipment damage, foreign matter intrusion, temperature anomaly), the data is classified, and a corresponding label is assigned to each type of data, the cleaned and labeled data is integrated into an anomaly condition data set, which will contain information such as description, time, location, severity of various anomaly conditions; according to the anomaly condition data set, the boundary conditions that affect the finite element analysis are extracted, such as structural changes caused by equipment damage, external forces caused by foreign matter intrusion, material property changes caused by temperature anomalies, etc.

[0155] In a preferred embodiment of the present application, the meteorological analysis results obtained based on the analysis of meteorological parameters in step 4 above are summarized to form a meteorological analysis data set, which can be used as environmental loads for finite element analysis, which can include:

[0156] The meteorological parameters are analyzed to extract features closely related to conductor vibration, such as average wind speed, maximum wind speed, wind direction change frequency, etc.; the analyzed meteorological features are integrated into a meteorological analysis data set, which will contain parameter values under various meteorological conditions and their corresponding time stamps, according to the meteorological analysis data set, the required environmental loads in finite element analysis are determined, such as wind load, temperature load, etc., these loads will be used to simulate the dynamic behavior of the conductor under different meteorological conditions.

[0157] In a preferred embodiment of the present application, the vibration mode, frequency and amplitude of the conductor in step 4 above are sorted to form a vibration analysis data set, which is used to simulate the dynamic behavior of the conductor in finite element, which can include:

[0158] Key features such as dominant vibration mode, main frequency component, maximum amplitude are extracted from the preprocessed vibration data; the extracted vibration features are integrated into a vibration analysis data set, which will be used to describe the vibration state of the conductor under different time and conditions, according to the vibration analysis data set, the corresponding dynamic behavior parameters are set in the finite element model to simulate the actual vibration of the conductor, which helps to more accurately evaluate the performance and safety of the conductor under various conditions.

[0159] In a preferred embodiment of the present application, in step 4 above, a finite element model of the three-span line is established, and finite element analysis is run to simulate the response of the three-span line under different meteorological conditions and vibration loads, to identify the key areas and factors that cause line failure, including:

[0160] Determine the specific structure of the three-span line, including the conductor, insulator, tower geometry, material and connection method, specifically including: obtaining detailed parameters of the conductor, insulator and tower from design drawings, technical specifications or site investigation; determine the three-dimensional geometry of the conductor, insulator and tower, including length, diameter, cross-sectional shape, etc.; determine the materials used for each component, such as the conductor may be aluminum or copper alloy, the insulator may be ceramic or glass, and the tower may be a steel structure; understand the connection method between the conductor and the insulator, the insulator and the tower, such as ring connection, bolt connection, etc.

[0161] According to the specific structure of the three-span line, create a three-span line geometric model, assign corresponding material properties to each component in the three-span line geometric model, including elastic modulus, density and Poisson's ratio; define the connection relationship between each component, including hinged or fixed, specifically including: according to the collected data, create the geometric model of the conductor, insulator and tower in ANSYS software; give each component the corresponding elastic modulus, density and Poisson's ratio and other material properties, set the connection relationship between each component in the model to ensure that the model conforms to the actual structure.

[0162] Mesh the three-span line geometric model, set the boundary conditions of the three-span line geometric model, including fixed constraints and load application positions, specifically including: mesh the geometric model to generate finite element meshes for calculation, determine the fixed constraints of the model, such as the fixed bottom of the tower; at the same time, set the position of the load application, such as the wind load application point on the conductor.

[0163] According to the meteorological analysis dataset, define different meteorological conditions, including wind speed, wind direction and temperature, and apply them as environmental loads to the three-span line geometric model, specifically including: import the previously collected meteorological analysis dataset into the finite element analysis software; define different wind speed, wind direction and temperature conditions according to the dataset, apply these meteorological conditions as environmental loads to the corresponding positions of the model, such as wind loads applied to the conductor and tower, and temperature loads affecting the properties of the materials.

[0164] Use the vibration analysis dataset to determine the vibration mode, frequency and amplitude of the conductor, and apply the vibration load to the conductor part, specifically including: import the vibration analysis dataset into the software, extract the vibration mode, frequency and amplitude of the conductor from the dataset, and apply these vibration loads to the conductor part of the model to simulate the actual vibration of the conductor.

[0165] Set the solution parameters, including time step and iteration number, start the solver and begin running the simulation analysis to obtain the simulation results, including stress distribution contour, deformation graph and vibration response curve, specifically including:

[0166] Determine the requirements and objectives of the analysis, which includes identifying the physical phenomenon to be simulated (such as stress distribution, deformation conditions, vibration response, etc.), and the information expected from the simulation results (such as maximum stress values, deformation amounts at specific locations, vibration frequencies and amplitudes, etc.).

[0167] Set the time step, which is an important parameter in finite element analysis, determining the degree of time advancement in the simulation process. Set the number of iterations, which refers to the number of numerical calculations performed by the solver within each time step. After setting the time step and the number of iterations, the next step is to start the finite element analysis solver, which involves the following steps:

[0168] Select the corresponding solver type according to the nature of the problem (linear, nonlinear, static, dynamic, etc.); click "Start" or similar button to officially start the solver for simulation calculation, and monitor the progress and status of the calculation through the ANSYS software interface or log file during the calculation process, which includes viewing the current iteration number, residual norm, calculation time, etc. If you find that the calculation is abnormal (such as not converging, too slow, etc.), you can adjust the parameters or interrupt the calculation in time. After the calculation is completed, ANSYS software will generate a series of simulation results for users to view and analyze, including:

[0169] Stress distribution cloud map, which displays the stress distribution of each part of the model in a graphical manner, helping to identify areas of stress concentration.

[0170] Deformation map, which shows the deformation of the model under the action of force, including displacement, strain, etc.

[0171] Vibration response curve, which displays the response of the model under vibration load in a chart, such as amplitude-frequency curve, time-displacement curve, etc.

[0172] According to the simulation results, identify the key areas of stress concentration, large deformation or abnormal vibration response, including: carefully observe the stress distribution cloud map, deformation map and vibration response curve, identify the areas of stress concentration, large deformation or abnormal vibration response, mark these key areas on the model or result map for subsequent in-depth analysis.

[0173] Analyze the response characteristics of the key areas under different weather conditions and vibration loads to determine the factors leading to line failure, including: accurately extract detailed data of the key areas from the simulation results, which usually include:

[0174] Specific stress values, locate the key areas in the stress distribution cloud map, and record the maximum, minimum and average stress values of these areas, which help to understand the degree and range of stress concentration.

[0175] Deformation data, including displacement and strain values, extracted from deformation maps. These data can reflect the stability and deformation of the structure under stress.

[0176] Vibration response parameters, such as frequency and amplitude, extracted from vibration response curves. These parameters can help analyze the dynamic characteristics and vibration stability of the structure.

[0177] After extracting the data, compare the response characteristics of the key areas under different weather conditions and vibration loads. This includes:

[0178] Weather condition comparison: compare stress values, deformation, and vibration response of the key areas under different wind speed, wind direction, and temperature conditions. Observe how these parameters change with weather conditions and whether there are obvious trends or rules.

[0179] Vibration load comparison: analyze the response of the key areas under different vibration frequencies and amplitudes, paying special attention to the occurrence of resonance and its impact on structural stability.

[0180] During the comparison process, use charts, curves, or data tables to clearly show the response differences under different conditions.

[0181] After comparing the response characteristics under different conditions, analyze the specific impact of these responses on line faults. This includes:

[0182] Stress concentration impact: high stress areas may cause material fatigue, cracking, or fracture, leading to line faults. Analyze the potential link between stress concentration and faults.

[0183] Deformation impact: excessive deformation may cause poor contact, misalignment, or fracture between line components, evaluate the impact of deformation on line performance and reliability.

[0184] Vibration impact: sustained vibration may cause line loosening, wear, or fracture, analyze the correlation between vibration response and line faults.

[0185] Finally, based on the results of the comparative analysis, determine the main factors that cause line faults. These factors may include:

[0186] Specific weather conditions: such as strong winds, extreme temperatures, etc., which may cause stress concentration or excessive deformation of the line.

[0187] Specific vibration load: certain vibration frequencies or amplitudes may cause resonance of the line, thereby accelerating fatigue and damage of the line.

[0188] Defects in structural design: such as insufficient stiffness in certain areas, unreasonable design of connection points, etc., which may also be the main factor causing faults.

[0189] In the embodiments of the present application, through simulation analysis, the key areas in the line where faults may occur can be identified in advance, so that the monitoring and maintenance of these areas can be strengthened in actual operation, effectively preventing the occurrence of line faults. Through finite element simulation, engineers can consider the line response under various complex environments in the design stage, thereby optimizing the structural design of the line and improving the weather resistance and stability of the line. By accurately simulating the influence of different meteorological conditions and vibration loads on the line, a more accurate maintenance plan can be developed, unnecessary maintenance work can be reduced, and maintenance costs can be reduced. Identifying and handling potential problems in key areas in a timely manner can significantly improve the overall safety of the line and reduce safety accidents caused by line faults. After understanding the response of the line under different environments, once a sudden situation occurs, the problem can be located faster and effective measures can be taken. Finite element simulation analysis is an important part of modern intelligent management of power systems, and through accurate data analysis and simulation, it can provide strong support for the construction of smart grids.

[0190] In a preferred embodiment of the present application, the three-span line geometric model is meshed, including:

[0191] The particle swarm optimization algorithm is set with particle number, speed range and position range; the learning rate and iteration number are set for the gradient descent algorithm, specifically including: selecting a suitable particle number, too many particles will increase the calculation burden, and too few particles may affect the globality of the search; a reasonable range is set for the speed of each particle to ensure that the movement of the particle in the search space is neither too fast nor too slow, too fast movement may cause the particle to miss the optimal solution, and too slow movement will affect the search efficiency; according to the possible value range of the meshing parameters, a limit is set for the position of the particle, which ensures that the particle will not exceed the reasonable parameter space during the search process; the learning rate determines the step size of parameter update in the gradient descent process, a suitable learning rate can ensure the convergence speed while avoiding oscillation; according to the complexity of the problem and the required optimization accuracy, a suitable iteration number is set, too many iterations may waste computing resources, and too few iterations may lead to incomplete optimization.

[0192] A set of initial particles is generated, each particle representing a meshing scheme, and the position of the particle represents the meshing parameters, including mesh size and distribution, specifically including: generating the initial position of each particle randomly within the set position range, these positions represent different combinations of meshing parameters, and an initial speed is randomly assigned to each particle within the set speed range.

[0193] The grid division scheme represented by each particle is evaluated by an evaluation function, the speed and position of each particle are updated, the evaluation of fitness and the updating of particles are repeated until the convergence condition is met, to obtain the global final particle, specifically including: for the grid division scheme represented by each particle, the fitness value of the grid division scheme is calculated using the evaluation function, the speed and position of each particle are updated according to the current position and speed of each particle, and the historical best position and global best position of itself, the process of evaluating fitness and updating particles is repeated until the preset convergence condition (such as reaching the maximum number of iterations) is met. The calculation formula of the evaluation function F(x) is:

[0194]

[0195] wherein, denotes the summation symbol, and denotes the summation of all n elements (or units), and the contribution of each element is calculated and totaled; denotes the weight coefficient of the stress part; denotes the stress calculation, which is the calculation method of the stress, wherein F i is the external force applied on the i-th element, A i (x) is the cross-sectional area of the i-th element, which depends on the optimization parameter x, and the definition of stress is the external force divided by the cross-sectional area; is the nonlinear correction term of the stress part, wherein α is the amplification coefficient, and β is the nonlinear index controlling the stress-strain relationship; w i (x) is the weight coefficient related to the deformation amount of the i-th element; is the calculation method of the deformation amount, wherein K i (x) is the stiffness coefficient of the i-th element; is the nonlinear correction term of the deformation amount part, wherein γ1, γ2 and γ are nonlinear indices; x represents the optimization parameter. Each particle in the particle swarm optimization algorithm updates the speed according to its current position and speed, combined with the personal historical optimal position and global optimal position, and the speed updating formula is as follows:

[0196]

[0197] wherein, is the speed of the i-th particle after the k+1 iteration; w is the inertia weight, which is used to control the speed of the particle; is the speed of the i-th particle at the k-th iteration; c1 and c2 are learning factors, which are used to balance the influence of the personal optimal and the global optimal; r1 and r2 are random numbers, usually in the range of [0, 1]; is the personal optimal position of the i-th particle at the k-th iteration; g k is the position of the global optimal particle; is the position of the i-th particle at the k-th iteration; the position of the particle is updated according to the velocity update, and the position update formula is as follows:

[0198]

[0199] wherein, is the new position of the i-th particle after the k+1-th iteration; is the position of the i-th particle at the k-th iteration; is the velocity of the i-th particle after the k+1-th iteration.

[0200] The grid division parameters are extracted from the global final particle as the initial scheme of gradient descent, specifically including: after the particle swarm optimization process is completed, the particle with the highest fitness value, i.e. the global optimal particle, is found, and the grid division parameters, including the size and distribution of the grid, etc., are extracted from the global optimal particle.

[0201] The initial grid scheme is disturbed, the gradient of the optimization function with respect to the grid division parameters is calculated by the finite difference method, the grid division parameters are adjusted in the direction of gradient descent, and the gradient is repeatedly calculated and the grid parameters are adjusted until a preset number of iterations is reached, to obtain the optimized grid division parameters, specifically including:

[0202] Based on the grid division parameters provided by the global optimal particle, a series of similar grid division schemes are generated by slightly disturbing the grid division parameters;

[0203] A set of grid division parameters to be optimized at present is determined, denoted as θ, wherein θ can be a multi-dimensional vector containing grid size, distribution and other parameters, and a small disturbance ∈ is set for parameter change when calculating the gradient, wherein the calculation formula of the optimization function F(θ) is:

[0204]

[0205] wherein, V j is the volume of the j-th grid element; l max,j is the length of the longest side in the j-th grid element; A j is the surface area of the j-th grid element; N(θ) is the number of grids; α and β are weight coefficients; c j is the center coordinate of the j-th grid element; μ is the center coordinate of the region of interest; σ is the standard deviation of the Gaussian function; ‖·‖ represents the Euclidean distance; the vector for recording the gradient has the same dimension as θ and is initialized to zero; for each parameter θ i in the parameter vector θ

[0206] The current parameter θ iA positive perturbation is performed to obtain a new parameter vector where Other parameters remain unchanged; calculate the optimization function value f(θ + ) of the perturbed scheme; perform a negative perturbation on the current parameters θ i to obtain a new parameter vector θ - , where Calculate the optimization function value f(θ - ) of the perturbed scheme, and use the central difference formula to calculate the partial derivative (i.e. one component of the gradient) at the parameter θ i

[0207]

[0208] According to the calculated gradient vector and the set learning rate η, update the grid partition parameters:

[0209]

[0210] Ensure that the updated θ new is still within a reasonable parameter range, and if necessary, perform truncation or adjustment, repeat the steps until the preset number of iterations is reached, and output the final grid partition parameters θ new .

[0211] By optimizing the grid partition parameters, the three-span line geometry model is meshed, specifically including: using the grid partition parameters optimized by the gradient descent, performing actual meshing operations on the three-span line geometry model, and checking the quality of the meshed grid to ensure that the requirements of numerical simulation or analysis are met.

[0212] In the embodiments of the present application, the particle swarm optimization algorithm can effectively explore the entire solution space through parallel search of multiple particles, avoid falling into a local optimal solution, and thus find a better grid partition scheme in the global range. The gradient descent algorithm further refines the global optimal solution found by the particle swarm optimization algorithm, updates the parameters by calculating the gradient and descending along the gradient direction, and can significantly improve the accuracy and quality of the meshing. Combined with the two algorithms, the global-local strategy reduces unnecessary computational load. The particle swarm optimization algorithm quickly locates to the potential optimal solution area, and the gradient descent algorithm performs efficient and fine search in the area, thereby improving the overall calculation efficiency.

[0213] In a preferred embodiment of the present application, in step 4 above, the safety state of the line is evaluated according to the key areas and factors causing the line fault, and the potential fault type is determined, specifically including:

[0214] ​Collect geological survey reports of key areas to understand soil types, groundwater levels, rock structure, etc. Collect historical climate data, especially records of extreme weather events (such as heavy rain, strong winds, freezing, etc.). Obtain traffic flow data of key areas through traffic monitoring systems, including peak hours, average speed, etc. Analyze the impact of geological conditions on line stability, identify potential geological risks such as subsidence, landslides, etc. Evaluate potential impacts of climate conditions on line facilities such as corrosion, aging, etc. Study the impact of traffic flow on track wear, signal system load, etc. Based on the above analysis, determine the main factors leading to failure, such as geological instability, frequent extreme weather, high traffic load, etc.

[0215] Measure the flatness and wear of the track in the field, check the operation logs of the signal system, test the accuracy and response time of the signal lights, visually inspect the electrical equipment, and measure key parameters such as voltage and current. Record the results of each inspection, including numerical data, photos or video evidence, and mark any problems or potential hazards found. Compare the field investigation data with historical failure records to assess the overall safety status of each key area, identify areas with obvious safety hazards or performance degradation, and classify key areas into different risk levels (such as low risk, medium risk, high risk) based on the assessment results. Obtain the characteristics of key areas such as geological conditions, climate impact, and traffic flow, and identify the key factors that may trigger line failures in these characteristics. Based on the characteristics of the key areas, list the types of failures that have occurred in this area in the past; summarize all possible failure types, such as track deformation, signal failure, electrical short circuit, etc.; for each failure type, study its possible causes; for example, track deformation may be caused by geological subsidence, frequent passage of heavy vehicles, etc.; determine the specific conditions that cause the failure to transition from a potential state to an actual occurrence. For example, an extreme weather event (rainwater accumulation) may trigger an electrical short circuit. Draw the complete development path from failure triggering to the final impact on line operation, which helps to understand how the failure escalates step by step and the possibility of intervention at different stages.

[0216] Collect failure records of key areas over a period of time (such as 5 years, 10 years), clean and classify the data to ensure accurate records of each failure type; count the number of occurrences of each failure type in historical data, calculate the frequency of failure occurrence, such as the average number of occurrences per year; combine the failure frequency and the latest safety status assessment results to assess the likelihood of each failure type occurring in the future.

[0217] Determine the dimensions of the analysis, including line safety, operational efficiency, passenger experience, etc., set specific evaluation indicators for each dimension, such as delay time caused by failure, number of passenger complaints, etc.; estimate the impact of each failure type on each dimension based on historical data and simulation scenarios; for example, signal failure may cause train delays and passenger dissatisfaction, and electrical short circuit may cause safety problems. Aggregate the impact assessment results of each failure type on each dimension, use weighted scoring to rank the comprehensive impact of failures to identify the most critical failure types.

[0218] As shown in Figure 1 A three-span line state intelligent evaluation system based on numerical weather prediction and machine learning, comprising:

[0219] An identification module for automatically identifying and classifying abnormal conditions of the line from preprocessed images by constructing, training and tuning a convolutional neural network model, the abnormal conditions including equipment damage data, foreign object intrusion data and temperature anomaly data;

[0220] A meteorological analysis module for obtaining meteorological parameters in real time, the meteorological parameters including temperature, humidity, wind speed, precipitation, air pressure and radiation; based on the meteorological parameters, analyzing their impact on the line operation state to obtain meteorological analysis results, the meteorological analysis results including the impact of temperature, humidity and wind speed on the line operation state;

[0221] An evaluation module for real-time monitoring and identifying the vibration mode and its characteristics of the three-span line conductor to evaluate the impact of vibration on the line operation state;

[0222] A finite element analysis module for organizing the equipment damage data, foreign object intrusion data and temperature anomaly data to form an abnormal condition dataset, and using it as the boundary condition of finite element analysis; aggregating the meteorological analysis results obtained based on the meteorological parameters to form a meteorological analysis dataset, and using it as the environmental load of finite element analysis; organizing the vibration mode, frequency and amplitude of the conductor to form a vibration analysis dataset for simulating the dynamic behavior of the conductor in finite element; establishing a finite element model of the three-span line, running finite element analysis, simulating the response of the three-span line under different meteorological conditions and vibration loads to identify the key areas and factors causing line failure; evaluating the safety state of the line according to the key areas and factors causing line failure, and determining the potential failure types.

[0223] It should be noted that the system is a system corresponding to the above method, all implementation manners in the above method embodiment are applicable to this embodiment, and the same technical effects can also be achieved.

[0224] The embodiment of the present application also provides a computing device, comprising a processor, a memory storing a computer program, the computer program being executed by the processor to perform the method as described above. All implementation manners in the above method embodiment are suitable for this embodiment and can achieve the same technical effects.

[0225] The embodiment of the present application also provides a computer readable storage medium storing instructions, which, when executed on a computer, cause the computer to perform the method as described above. All implementation manners in the above method embodiment are suitable for this embodiment and can achieve the same technical effects.

[0226] The above is the preferred embodiment of the present application, it should be pointed out that, for those skilled in the art, without departing from the principles of the present application, can make a number of improvements and refinements, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A three-span railway line status intelligent assessment system based on numerical weather prediction and machine learning, characterized in that, include: The identification module is used to automatically identify and classify abnormal conditions of the circuit from preprocessed images by building, training and tuning a convolutional neural network model. Abnormal conditions include equipment damage data, foreign object intrusion data and abnormal temperature data. The meteorological analysis module is used to acquire meteorological parameters in real time, including temperature, humidity, wind speed, precipitation, air pressure, and radiation. Based on the meteorological parameters, it analyzes their impact on the line's operating status to obtain meteorological analysis results, including the impact of temperature, humidity, and wind speed on the line's operating status. The evaluation module is used to monitor and identify the vibration modes and characteristics of the conductors across three spans in real time to assess the impact of vibration on the line's operating status. This includes defining the initial temperature T for the simulated annealing algorithm. o Termination temperature T f The cooling coefficient α and the number of iterations L at each temperature; set the initial solution S. o S represents the initial set of positions of the vibration sensors; an objective function is defined to evaluate the quality of the position set S; for each iteration at the current temperature T, a new solution S is generated. new ; Calculate the objective function value f(S) of the new solution new If f(S) new The objective function value f(S) is greater than the current solution's objective function value. current If the solution is correct, then accept the new solution and update S. current =S new , among which, S current Indicates the current solution; if f(S) new )≤f(S current If the current temperature T drops to the termination temperature T, then accept the new solution, complete the inner loop, and lower the temperature. f The simulation annealing process is then stopped, and the final solution, i.e., the final set of vibration sensor positions, is output. Multiple vibration sensors are installed at key locations on the three-span line to monitor the conductor vibration data in real time. The vibration data is preprocessed to obtain preprocessed vibration data. Fourier transform is used to extract the vibration characteristics of the preprocessed vibration data, including frequency, amplitude, and phase. Based on the vibration characteristics, a vibration analysis model is established to identify the conductor vibration mode and analyze the impact of vibration on the line operation state to obtain the vibration analysis results, which include the conductor vibration mode, frequency, and amplitude. The finite element analysis module is used to organize equipment damage data, foreign object intrusion data, and temperature anomaly data into an abnormal condition dataset, which serves as the boundary conditions for finite element analysis. It also summarizes meteorological analysis results based on meteorological parameters to form a meteorological analysis dataset, which serves as the environmental load for finite element analysis. Furthermore, it organizes the vibration modes, frequencies, and amplitudes of conductors to form a vibration analysis dataset, used to simulate the dynamic behavior of conductors in finite element analysis. Finally, it establishes a finite element model of a three-span line, runs finite element analysis, and simulates the response of the three-span line under different meteorological conditions and vibration loads to identify key areas and factors leading to line faults, including: [details omitted]. The specific structure of the three-span line is defined, including the geometry, materials, and connection methods of the conductors, insulators, and towers. Based on the specific structure of the three-span line, a geometric model of the three-span line is created. In the geometric model, corresponding material properties are specified for each component, including elastic modulus, density, and Poisson's ratio. The connection relationships between the components are defined, including hinged or fixed connections. The geometric model of the three-span line is meshed, and boundary conditions are set, including fixed constraints and load application locations. Based on the key areas and factors leading to line faults, the safety status of the line is assessed, and potential fault types are determined. The meshing of the geometric model of the three-span line includes: The particle swarm optimization algorithm is configured with particle number, velocity range, and position range; the gradient descent algorithm is configured with learning rate and number of iterations. An initial set of particles is generated, each representing a mesh partitioning scheme. The particle's position indicates the mesh partitioning parameters, including mesh size and distribution. Each particle's mesh partitioning scheme is evaluated using an evaluation function, and the velocity and position of each particle are updated. The fitness evaluation and particle updates are repeated until the convergence condition is met to obtain the global final particles. Mesh partitioning parameters are extracted from the global final particles as the initial scheme for gradient descent. The initial mesh scheme is perturbed, and the gradient of the optimization function with respect to the mesh partitioning parameters is calculated using the finite difference method. The mesh partitioning parameters are adjusted along the gradient descent direction. The gradient calculation and mesh parameter adjustment are repeated until the preset number of iterations is reached to obtain the optimized mesh partitioning parameters. The optimized mesh partitioning parameters are then used to partition the three-span railway geometry model.

2. The intelligent assessment system for the status of three-span railway lines based on numerical weather prediction and machine learning as described in claim 1, characterized in that, By constructing, training, and optimizing a convolutional neural network model, abnormal conditions of circuits can be automatically identified and classified from preprocessed images. These abnormal conditions include equipment damage data, foreign object intrusion data, and abnormal temperature data, including: Real-time image data of the three-span line is collected by video surveillance equipment and the images are preprocessed to obtain the preprocessed images. Samples containing equipment damage, foreign object intrusion, and abnormal temperature conditions are selected from the preprocessed images, and the samples are labeled to create a labeled dataset. The labeled dataset is divided into a training set, a validation set, and a test set; Construct a convolutional neural network model and initialize its parameters, including weights and biases; The convolutional neural network model is trained using a training set. Predictions are calculated through forward propagation, and the parameters of the convolutional neural network model are optimized through backpropagation to obtain the trained convolutional neural network model. The trained convolutional neural network model is evaluated using a test set, and the accuracy is calculated. Based on the accuracy, the trained convolutional neural network model is adjusted to obtain the final convolutional neural network model. The preprocessed new image is input into the final convolutional neural network model to obtain the abnormal conditions of the line, including equipment damage, foreign object intrusion, and temperature anomalies.

3. The intelligent assessment system for the status of three-span railway lines based on numerical weather prediction and machine learning according to claim 2, characterized in that, The trained convolutional neural network model was evaluated using a test set, and its accuracy was calculated, including: The trained convolutional neural network model is used to predict the label of each image in the test set, and the predicted label of each image is recorded. The predicted label of each image in the test set is compared with the true label. Construct a confusion matrix to record the number of true positives and predicted positives for each category. The confusion matrix includes true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN). Based on the confusion matrix, calculate the accuracy, which is the proportion of all correctly predicted samples out of the total samples. Initialize a list to store the accuracy of each iteration; for each of the K subsets, denoted as the validation set, perform the following steps: Merge the remaining K-1 subsets to form a temporary "test set"; use the trained convolutional neural network model to predict each image in the temporary "test set" and record the predicted label; compare the predicted label of each image in the temporary "test set" with the true label and update the confusion matrix; calculate the accuracy of the current iteration based on the confusion matrix and add it to the accuracy record; reset the confusion matrix to prepare for the next iteration; calculate the average accuracy of K iterations as the final accuracy of the model.

4. The intelligent assessment system for the status of three-span railway lines based on numerical weather prediction and machine learning according to claim 3, characterized in that, Based on vibration characteristics, a vibration analysis model is established to identify the vibration modes of the conductor and analyze the impact of vibration on the line's operating state, thereby obtaining vibration analysis results. These results include the conductor's vibration modes, frequencies, and amplitudes, including: Based on the characteristics of the vibration data, the vibration data is labeled, and its corresponding vibration mode label is determined and labeled to form a labeled dataset; The labeled dataset is divided into a training set, a validation set, and a test set. The recurrent neural network model is trained using the training set. During the training process, the weights and biases of the recurrent neural network model are continuously adjusted through the backpropagation algorithm and the optimizer to obtain the trained recurrent neural network model. The trained recurrent neural network model is used to classify new vibration data to obtain corresponding predicted vibration mode labels; By comparing predicted vibration mode labels with predefined vibration mode labels, different vibration modes, including breeze vibration and swaying, are identified. The correlation between different vibration modes and line operating conditions was analyzed and identified; Based on correlation, the risk level of the line under different vibration modes is assessed.

5. The intelligent assessment system for the status of three-span railway lines based on numerical weather prediction and machine learning according to claim 4, characterized in that, Based on the characteristics of the vibration data, the vibration data is labeled, and its corresponding vibration mode labels are determined and labeled to form a labeled dataset, including: Determine the category of vibration mode, including light oscillation and galloping; Define a unique label for each vibration mode and convert the characteristics of the vibration data, i.e. frequency or amplitude, into genes in a genetic algorithm. Genes are used to form the chromosomes of an individual. An initial population is randomly generated, where each individual represents a vibrational pattern tag sequence; Key features, including time-domain and frequency-domain features, are extracted from each vibration data sample. Define a fitness function to evaluate the degree of matching between each individual and the characteristics of the vibration data; Based on the fitness function, select the corresponding individuals to enter the next generation, randomly select two individuals as parents, perform crossover to generate new offspring, randomly mutate the genes of the offspring, i.e. the labels of vibration patterns, and repeat the selection, crossover and mutation operations until the termination condition is met. When the termination condition is met, stop the iteration and output the final individual as the labeling result of the vibration data. The final individual data is decoded into a corresponding vibration pattern label sequence, and the vibration pattern label sequence is matched with the original vibration data samples to form a labeled dataset.

6. The intelligent assessment system for the status of three-span railway lines based on numerical weather prediction and machine learning according to claim 5, characterized in that, A finite element model of a three-span railway line was established, and finite element analysis was performed to simulate the response of the three-span line under different meteorological conditions and vibration loads, in order to identify the key areas and factors leading to line faults. This also included: Based on the meteorological analysis dataset, different meteorological conditions, including wind speed, wind direction and temperature, are defined and applied as environmental loads to the geometric model of the three-span line. Using the vibration analysis dataset, the vibration modes, frequencies, and amplitudes of the conductor are determined, and vibration loads are applied to the conductor section. Set the solution parameters, including the time step and the number of iterations, start the solver, and begin the simulation analysis to obtain the simulation results, including stress distribution contour maps, deformation diagrams, and vibration response curves. Based on the simulation results, identify key areas with stress concentration, large deformation, or abnormal vibration response; Analyze the response characteristics of key areas under different meteorological conditions and vibration loads to identify the factors leading to line failures.

7. A method for intelligent assessment of the status of three-span railway lines based on numerical weather prediction and machine learning, characterized in that, This method is used to execute the system as described in any one of claims 1 to 6, the method comprising the following steps: By building, training and tuning convolutional neural network models, abnormal conditions of lines can be automatically identified and classified from preprocessed images. These abnormal conditions include equipment damage data, foreign object intrusion data and abnormal temperature data. Meteorological parameters are acquired in real time, including temperature, humidity, wind speed, precipitation, air pressure, and radiation. Based on these meteorological parameters, their impact on the line's operating status is analyzed to obtain meteorological analysis results, including the effects of temperature, humidity, and wind speed on the line's operating status. Real-time monitoring and identification of vibration modes and characteristics of conductors in three-span lines are used to assess the impact of vibration on line operation. Data on equipment damage, foreign object intrusion, and abnormal temperature were compiled to form an abnormal condition dataset, which was used as the boundary conditions for finite element analysis. Meteorological analysis results based on meteorological parameters were summarized to form a meteorological analysis dataset, which was used as the environmental load for finite element analysis. Vibration modes, frequencies, and amplitudes of the conductors were compiled to form a vibration analysis dataset, used to simulate the dynamic behavior of the conductors in finite element analysis. A finite element model of a three-span line was established, and finite element analysis was run to simulate the response of the three-span line under different meteorological conditions and vibration loads to identify key areas and factors leading to line faults. Based on the key areas and factors leading to line faults, the safety status of the line was assessed, and potential fault types were determined.

8. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to perform the method as described in claim 7.

Images