Function design and realization of transmission line icing monitoring and early warning system combined with micro-meteorology and micro-topography in Guizhou

The icing of transmission lines is one of the main causes of transmission line accidents and power outages, which poses a great threat to the safe and stable operation of transmission lines. Therefore, it is of great significance to establish a set of icing monitoring and early warning system to predict possible icing situations in advance, take corresponding measures, and ensure the normal operation of transmission lines. According to the characteristics of micro-meteorology and micro-topography in Guizhou, this paper designs and implements a set of transmission line icing monitoring and early warning system based on micro-meteorology and micro-topography. Through the experimental analysis of the system, it is proved that the system can effectively monitor and warn the icing of the transmission line, and improve the safety and stability of the transmission line.


Introduction
The icing of transmission lines is a common problem in extreme cold weather in winter.It may cause serious consequences such as tripping of insulators of transmission lines and burning of cables, which poses a great threat to the safe operation of the power grid.The micrometeorology and microtopography characteristics of the Guizhou region have further increased the uncertainty and complexity of transmission line icing.Therefore, it is of great significance to establish a reliable transmission line icing monitoring and early warning system to ensure the safety of power grid operation.
The purpose of this paper is to design and implement a reliable monitoring system for transmission line icing monitoring and early warning under micro-meteorological and micro-terrain conditions in Guizhou, so as to improve the work efficiency of power grid operation and maintenance personnel and the operation safety of the power grid.

System structure
The transmission line icing monitoring and early warning system designed in this paper adopt a distributed architecture.The whole system includes modules such as data acquisition, data preprocessing, feature extraction, model training, prediction and early warning, as shown in Figure 1.The prediction and early warning module mainly uses the trained model to predict the icing situation of transmission lines in the future and performs corresponding early warning processing according to the prediction results.

Data collection
The data acquisition technology route of this system mainly includes sensor type selection, layout scheme, data transmission, data storage, etc.
1. Sensor selection In this system, in order to realize the icing monitoring of transmission lines, it is necessary to select sensors such as temperature sensors, humidity sensors, and wind speed sensors for data collection.This system selects sensors with high accuracy, good stability, strong reliability, and easy maintenance for data collection.

Layout plan
In order to ensure the accuracy and comprehensiveness of the data, the system needs to arrange the sensors reasonably.The sensor layout scheme mainly includes the following aspects: (1) Location selection: It is necessary to select a suitable location for the layout of the sensor.The characteristics of the parameters measured by the sensor should be considered.For example, the temperature sensor should be placed in the air circulation position, and the humidity sensor should be placed in the sunshade position.
(2) Quantity determination: The number of sensors should be considered comprehensively based on factors such as the length of the transmission line, line characteristics, and monitoring requirements, and an appropriate number should be selected for deployment.
(3) Arrangement method: The appropriate arrangement method should be chosen according to the actual situation, and different methods such as fixed type, movable type, and hanging type for arrangement can be selected.

Data Transmission and Storage 1. Data transmission
In order to ensure the real-time and reliability of data, this system adopts wireless data transmission technology for data transmission.Specifically, this system adopts LoRa wireless communication technology, which has the advantages of being able to realize long-distance data transmission, strong signal penetration ability, strong anti-interference ability, low power consumption, and low costs.
2. Data storage For real-time data storage, the relational database MySQL is used.Relational databases take transaction processing as the core, support SQL query language, have good data consistency and transaction processing capabilities, and are suitable for real-time data storage.In terms of historical data storage, the distributed file system Hadoop HDFS is adopted.The distributed file system has the advantages of distributed storage, high availability, and horizontal expansion, and is suitable for the storage of massive data.

Data processing
Preliminary processing of the collected meteorological data, micro-topographic data and transmission line parameter data, including data cleaning, format conversion, outlier processing, etc.At the same time, the data is standardized so that various types of data are comparable.For transmission line icing prediction, relevant features are extracted from the processed data.For example, meteorological data can extract indicators such as temperature, humidity, and precipitation, micro-topographic data can extract indicators such as elevation, slope, and aspect, and transmission line parameter data can extract indicators such as line length, line diameter, and ice coverage level.The extracted features will be used as input for model training and prediction.

Meteorological Forecasting and Modeling
Meteorological data were obtained through meteorological observation stations, satellites, radars, etc., including indicators such as temperature, wind speed, humidity, and precipitation.Quality control was performed on the acquired meteorological data, and unqualified data and abnormal data were excluded.The WRF meteorological model was selected according to the time and space range to be predicted.The parameters of the selected meteorological model were configured, including physical process parameters, numerical parameters, etc.Then the configured weather model was run to obtain corresponding weather forecast results.Then the weather forecast results were subject to post-processing, including interpolation, smoothing, etc., and then compared with measured data.
Model evaluation and adjustment: Through comparison and verification with the measured data, the prediction accuracy of the meteorological model was evaluated, and the meteorological model was adjusted and optimized according to the evaluation results.

Selection and optimization of forecasting algorithms 1. Choice of forecasting algorithm
The neural network is a computational model based on artificial neurons simulating the way the human brain works.In the icing prediction of transmission lines, the prediction of the amount of icing is realized by establishing a neural network model.Commonly used neural network algorithms include multilayer perceptron (MLP), recurrent neural network (RNN), and convolutional neural network (CNN).

Optimize the technical route
In transmission line icing prediction, feature selection algorithms are used to determine the most relevant meteorological and topographical features.Techniques such as grid search are used to optimize parameters to improve prediction accuracy.The voting method and the weighted average method are used to combine multiple prediction models to improve prediction accuracy.The icing prediction of transmission lines needs to be updated in real-time to reflect the impact of meteorological and terrain changes on the amount of icing.Therefore, it is necessary to introduce real-time data into the forecasting model and optimize the forecasting results through model update technology.

Early warning mechanism design 1. Determination of early warning indicators
The determination of early warning indicators is the first step in the design of the early warning mechanism.According to the real-time monitoring of icing conditions and historical data, appropriate indicators are selected to evaluate the current icing risk.The commonly used indicators include line ice thickness, ice area, ice weight, etc.

Classification of early warning levels
The division of early warning levels is an important link in the design of early warning mechanisms.Through the definition and division of different early warning levels, early warning signals can be made more accurate and effective.In general, the early warning level can be divided into three levels, namely the first level early warning, the second level early warning and the third level early warning, which are divided according to the severity of the icing situation.

Release of early warning signals
The release of early warning signals is the last step in the design of the early warning mechanism.Its purpose is to transmit early warning information to the operation and maintenance personnel in a timely manner so that they can take corresponding measures to deal with possible problems.The release of early warning signals can be achieved in a variety of ways, such as text messages, emails, voice calls, etc.

User query and report generation
Background development: According to business requirements, background code needs to be written to handle user queries and report generation requests.The popular background framework Spring Boot is used for rapid development.
Front-end development: Users need to perform query and report generation operations through the front-end interface, so it is necessary to design a beautiful and easy-to-use front-end interface.The popular front-end framework Vue is used for rapid development.
Report generation: According to user needs, different types of reports can be generated, such as tables, charts, etc.It is realized by using open-source report generation tools such as JasperReports or BIRT.
Data visualization: In order to allow users to understand the data more intuitively, some data visualization functions can be designed.It is realized by using an open-source data visualization library such as D3.js or ECharts.

System page
We use back-end Java development technology and front-end Vue technology to realize platform development.The specific page is as follows.
Figure 2. Some screenshots of the system.

Accuracy Verification
The data used in this paper are historical data from January to March 2023, among which February is the most severe month for winter transmission lines in Guizhou.The input data used in this paper include meteorological data such as temperature, humidity, precipitation, wind speed, wind direction, and terrain height, as well as data such as the height, model, and spanning number of transmission lines.The output data is ice thickness.
In order to evaluate the predictive ability of the proposed icing prediction model, the experimental data are randomly divided into a training set and a test set, in which the training set accounts for 70% and the test set accounts for 30%.The model is trained using the training set and the performance of the model is evaluated using the test set.In this paper, metrics such as root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient (Correlation coefficient) are used to evaluate the performance of the proposed prediction model.Among them, RMSE and MAE represent the average error and mean absolute error between the predicted value and the actual value, respectively, and the correlation coefficient represents the degree of correlation between the two.
Table 5-1 shows the performance indicators of the icing prediction models trained with different algorithms, where MLP stands for multi-layer perceptron neural network, SVR stands for support vector regression, RF stands for the random forest, GBDT stands for gradient boosting decision tree, XGBoost stands for Extreme Gradient Boosting Tree, and LightGBM stands for Lightweight Gradient Boosting Tree.It can be seen that the model proposed in this paper has better performance than other models, especially in RMSE and MAE indicators.In order to test the accuracy of the icing warning system, we used a 500 kV substation in Guizhou Province for actual monitoring.The substation has a total of 12 transmission towers, and each tower is equipped with 2 weather stations and 1 temperature sensor for real-time collection of meteorological and temperature data.
In the actual monitoring process, we processed the collected data every 5 minutes to obtain the ice status on each tower.According to the different levels of icing, we divided the icing into three levels: light, moderate, and severe, and sent the monitoring results to relevant staff through text messages and emails.
During the monitoring period, there were 2 severe icing situations.By comparing the actual monitoring results with the prediction results of the early warning system, we found that the early warning system can give an early warning 24 hours before the occurrence of icing, and can accurately predict the level of icing.At the same time, the false alarm rate of the early warning system is very low, which can meet the needs of practical applications.
We conducted an online test of the system, which uses real-time data for forecasting and early warning.The test results show that the prediction accuracy of the system has reached more than 80%, indicating that the system has certain accuracy and reliability.The system can send early warning signals in time, and can update the forecast results in time to provide users with accurate forecast and early warning information.

Conclusion
Based on the icing characteristics of transmission lines in Guizhou's microclimate and micro-topography, this paper designs a system for monitoring and early warning of transmission line icing, which can realize real-time monitoring, prediction and early warning of icing on transmission lines.The experimental results show that the designed system can effectively improve the safe operation level of transmission lines, and provide strong technical support and guarantee for the safety management and control of transmission lines.

Figure 1 .
Figure 1.Structure diagram of transmission line icing monitoring and early warning system.The data acquisition module is mainly responsible for obtaining real-time and accurate transmission line icing-related data from multiple data sources, including micro-meteorological data, microtopography data, transmission line status data, etc.The data preprocessing module mainly cleans and processes the collected raw data, removes outliers and noise interference, and ensures data quality and accuracy.The feature extraction module mainly performs feature extraction and selection on the processed data, extracts representative features, and screens, and optimizes the features.The model training module is mainly responsible for constructing and training the icing prediction model, including data set division, model selection and parameter optimization.The prediction and early warning module mainly uses the trained model to predict the icing situation of transmission lines in the future and performs corresponding early warning processing according to the prediction results.
Table1.Performance indicators of icing prediction models trained by different algorithms