Research On Icing Thickness Prediction Model Of Overhead Transmission Lines

The overhead transmission lines are vulnerable to icing disasters in the operation process, icing may cause line tripping, hardware damage, even cause accidents such as disconnection, pour tower, which influence the safe and stable operation of the power grid seriously.In recent years, Yunnan power grid corporation set up large-scale icing online monitoring system. To use the monitoring system effectively and provide guidance for deicing and melting icing work, based on a lot of online monitoring icing and meteorological data, this paper deduced and verified the equivalent icing thickness calculation model, established the grey prediction model and neural network prediction model of icing thickness, analyzed the effect and applicability of different icing prediction models. The contents of this paper are as follows:(1) Introduced the icing online monitoring system and each component. Deduced the theory of equivalent icing thickness calculation model based on the parabolic equation of overhead lines. For the problem of lacking transmission line basic parameters in practical application, this paper improved the model by calculating the equivalent line length based on historical online monitoring data. Analyzed the online monitoring icing data of Yunnan power grid based on the model, compared with meteorological data and icing field monitoring images, verified the feasibility and accuracy of the equivalent icing thickness calculation model.(2) Based on the single time series of icing thickness, icing thickness grey prediction model and neural network prediction model were established. Verified the models by real icing cases, results show that prediction error increases obviously when meteorological parameters change largely. Therefore, it is necessary to consider the meteorological effects on icing when predicting icing.(3) In order to study the correlation between icing and meteorological parameters, calculated the grey comprehensive relational grade between icing and environmental temperature, relative humidity, wind speed, wind direction based on grey relational analysis, results show that the correlation between environment temperature, wind speed and the ice thickness is bigger. On the basis of grey correlation analysis, established the multivariate grey prediction model and neural network prediction model based on the icing thickness and the environment temperature, wind speed. Take Yunnan power grid corporation lines icing case for instance, verified the prediction effect of models. Results show that the prediction error is small, and meet the demand of icing prediction engineering application. Finally, combine the real-time online monitoring icing thickness with multivariate grey prediction model of icing thickness, established the ice monitoring and early warning model. For the convenient use of line operators, icing monitoring and early warning system was developed using MATLAB software.