Abstract:

Under the meteorological factors of icing, the changes of icing thickness, shape and distribution on transmission conductor with time cumulation affect the safety operation of power grid system. Based on prediction models, this paper analyzes the association between various stages from icing growth to conductor deicing, and discusses the advantages of models and the possibility of mutual combination. The changes of icing from micro to macro in the whole icing circle affect the growth process of conductor icing. The prediction models can be combined based on the whole cycle. Firstly, the initial data is denoised to solve the data divergence, and the dimensionality reduction by the principal component analysis method can improve the prediction accuracy. Secondly, the combination and intercrossing mode of support vector machine, hybrid swarm intelligence optimization algorithm, genetic algorithm in the model focus on the identification and modeling of icing process. Thirdly, the application of thermal deicing techniques such as AC ice melting and eddy self-heating ring in the de-icing stage helps to form a dynamic closed-loop system for icing monitoring. Finally, an outlook is made on the research direction of icing prediction for transmission lines.

Key words: time cumulation, meteorological factor, overhead transmission line, icing, prediction model, algorithm