Research on the prediction method of icing disasters on ultra-high voltage transmission lines based on spatiotemporal feature analysis

The risk of icing on power lines in areas with high rainfall is increasing due to extremely cold weather, which will pose a huge challenge to power supply work. The causes of icing on ultra-high voltage power lines are closely related to weather, terrain, temperature, etc. The high randomness of its occurrence makes it difficult to predict. A prediction method for icing disasters on ultra-high voltage transmission lines based on extemporization feature analysis is proposed in this paper, which analyzes the faults of ultra-high voltage transmission lines from the physical characteristics of power conductors and the changes in the external environment. Considering the mutual influence between the factors that cause icing, key factor extraction is used to sort meteorological, temperature, humidity, geographic and other data, and weight values are assigned to various indicator data based on the sorting results. Construct a multi factor comprehensive impact sequence data prediction model to incorporate the above indicators into the input parameters of sequence prediction. Based on the training of historical data using support vector machine algorithm, the fitting method of the ice cover risk prediction model mentioned above is improved, and the best fitting coefficient is found to improve the prediction accuracy of the model. Due to the comprehensive consideration of various factors such as weather, high precision terrain, and environment, this model predicts the occurrence patterns of icing disasters from both time and space dimensions, and the prediction results of icing disasters are more closely related to the actual situation.