Predicting the probability of ice storm damages to electricity transmission facilities based on ELM and Copula function

Abstract

In this paper, a novel method based on extreme learning machine (ELM) and Copula function is proposed to predict the damages to electricity transmission facilities during ice storms. The ELM is firstly trained based on the historical data of wind speed, freezing precipitation, temperature, as well as the distribution parameters of wind and ice loads. The ELM can then be employed to predict the distributions of the real-time wind and ice loads on electricity transmission facilities. Furthermore, the correlation between wind load and ice load is modeled with Copula functions. On the basis of ELM and Copula function, the joint probability distribution of wind and ice loads can be finally formulated and applied to predict the potential damages to electricity transmission facilities such as transmission lines and towers. The proposed method is tested with a real dataset to demonstrate its effectiveness.