Research On Transmission Line Ice Coating Based On Improved Particle Swarm Optimization

Overhead transmission lines are important channels for transmitting electrical energy in the power grid.In recent years,transmission lines in the power industry at home and abroad have been repeatedly subjected to ice disasters,which have brought harm to the safe operation of the power grid.Serious accidents may even cause large-scale power outages and difficult recovery of the power grid.Under such a background,we can be able to keep abreast of trends icing by monitoring the transmission line and forecasting icing conditions based on monitoring data,so early warning can be achieved when icing reaches a certain thickness,it can provide evidence for anti-icing and de-icing of transmission lines.In this paper,the conditions of ice coating on the transmission line,the mechanism of ice coating and the growth process are firstly described.The ice coating on the transmission line is classified,and various meteorological factors affecting the ice coating of the wire are comprehensively analyzed.On this basis,the classic ice coating model is introduced.Then the heat balance of the wire surface is analyzed.The Makkonen numerical analysis model based on thermal equilibrium analysis is used to extract the ice-influencing factors of the model.Secondly,the thickn(ess of the ice on the transmission line was predicted by the extreme learning machine neural network and the support vector machine regression model.Through comparison and analysis of the simulation results of the example,it is found that the support line vector based on the prediction of the ice-covered prediction of the transmission line is smaller than the prediction based on the neural network of the extreme learning machine,and the training time is short.Therefore,the prediction model based on support vector machine linear regression has higher precision and is more suitable for transmission line ice prediction.In the prediction process,the penalty parameters and kernel parameters in the support vector machine have a great influence on the accuracy of the algorithm.We use the improved particle swarm optimization algorithm to normalize the penalty parameters and kernel parameters in the support vector machine.The results show that the improved particle swarm optimization algorithm can optimize the relative prediction error and further improve the prediction accuracy.It can be seen from the analysis of the results of two lines of ice-covered prediction examples that the support vector machine has better prediction ability than the extreme learning machine neural network when the historical data sample is small,and is more suitable for processing small sample problems,and supports the improved particle swarm optimization algorithm.After the vector machine is optimized,the accuracy of the ice thickness prediction is improved.