2023•IET conference proceedings.Requires access
Icing risk prediction of overhead lines based on MLP neural network
Min Sun, Qi Wang, Yu-E Sun, Yaowen Yang, Wen‐Ming Chen, Wei Ding, Li Zhang, Huan Huang, Yong Song
Open publisher page 0 citations
Abstract
The icing of overhead lines threatens the safe and stable operation of the power grid. Predicting the icing risk of overhead lines can provide decision-making reference for the power grid to carry out anti icing work in time. In this paper, online monitoring of overhead line icing and micro meteorological data pre-processing, using SPSS data analysis platform, MLP neural network model is established. Micro meteorological parameters are used to distinguish overhead lines icing, so as to predict the risk of overhead line icing in the next 24 hours. The results show that the absolute error of the future 24h icing risk prediction of the non icing line is not more than 7%, and the model can accurately predict the future 24h icing risk of the current non icing overhead lines.