Research On Short-term Prediction Models Of Ice Thickness Of Transmission Line

Transmission line icing often poses a serious threat to the safe operation of power grid.Therefore,the online monitoring and forecasting and early warning technology for icing are the key to ensure the safe and stable operation of transmission lines and the requirement for building intelligent transmission system.In this paper,we use the machine learning intelligent algorithm in computer science to study the intelligent prediction of ice coating.The concrete work is listed as follows:1)The mechanism of icing and the icing patterns of transmission lines were studied,and the icing patterns of transmission lines were analyzed.The corresponding icing lines were described from different classification standards.Types and Characteristics of Ice Coating.2)The icing data in on-line monitoring database is analyzed.Through the screening and pretreatment of these data,the complete icing process data can be selected to predict the experiment.The gray correlation analysis method was used to analyze the equivalent thickness of icing ice,and the key influencing factors of the icing were obtained,and these key factors were used as the input.3)The ELM-based prediction model of transmission line icing was studied.The genetic algorithm was used to optimize the parameters of ELM network,and an improved equivalent thickness forecasting model for limiting learning machine was proposed.At last,it is proved that the prediction accuracy of this model is higher than that of BP neural network and GRNN neural network.4)Firstly,starting from the analysis of icing history data,the principal component analysis method is used to reduce the dimension of the experimental data.Then the support vector machine(SVM)based on the thought evolutionary algorithm is used to carry out the icing thickness Prediction,more ideal experimental results.