Predicting faulty lines in power grids based on icing thickness prediction

Abstract

Blackouts are mainly caused by internal faults. At present, the primary method of power system security analysis is reductionism. However, reductionism has great limitations in the establishment of the model and ignores a lot of actual electricity system parameters. With the acceleration of industrialization, many of the world's large-scale blackouts are caused by natural disasters such as bad weather, but this aspect has not received due attention. In this work, we propose a faulty line prediction method FLP (Faulty line prediction in power grids based on icing thickness prediction). FLP considers the impact of natural disasters such as bad weather on the safety of lines and then predicts the probability of line failure based on the prediction of ice thickness. First, FLP constructs a model according to historical icing data to predict the future icing thickness of a line; Second, FLP analyzes the probability of line failure based on the predicted ice thickness. Experimental results show that FLP has better performance than similar methods in predicting ice thickness and the probability of faulty lines.