Research on Multi-variable Grey Prediction Model for Icing Thickness

The icing disaster seriously affects the safe operation of the power grid. Through predicting the icing growth trend in the early stage of icing, the deicing and the hidden icing danger could be guided and eliminated. In this paper, a multi-variable grey prediction model for icing thickness is studied. Firstly, the correlation between micro-meteorological parameters such as ambient temperature, relative humidity, wind speed, wind direction and icing thickness is analyzed by grey correlation analysis method. Furthermore, considering the grey correlation analysis results, a multivariate grey prediction model of icing thickness is established based on the icing thickness, ambient temperature and wind speed. The validity of the model is verified by the online monitoring data of an icing process. The prediction error is within 5%, which satisfies the engineering application requirements of the transmission line icing prediction.