Prediction Of Runway Ice Thickness Based On Grey Neural Network

In recent years,extreme weather conditions such as fog,freezing rain and snow have brought serious risks to the safety and efficiency of civil aviation transport.Among them,runway icing is a key factor affecting the safety and efficiency of civil aviation operation in winter.Aiming at the problem of icing on the runway in winter,based on the field monitoring data under rain and snow weather and the experimental data of simulated icing on the runway,the relevant problems of icing thickness prediction on the runway are studied.The research contents mainly include:(1)Analyze the influence of meteorological factors,environmental factors and road surface materials on the ice thickness,and calculate the correlation between quantifiable factors and the thickness of runway icing according to the grey relational degree model.According to the sorting results of the correlation between various factors and the thickness of ice accretion,it prepares for the establishment of the prediction model of subsequent ice accretion thickness.(2)According to the timing and grey characteristics of the icing process,the grey model is used to predict the thickness of the runway icing.Based on the initial condition of the grey model,a grey prediction model with two-point weighted optimization as the initial condition is proposed.The icing data in different stages are selected to verify the model,and the best grey prediction model of icing thickness is obtained,which prepared for the establishment of the combined prediction model.(3)According to the characteristics of multi-factor and nonlinear change in the icing process,a prediction model of RBF neural network is established.Combining the grey prediction model with the RBF neural network prediction model,the grey neural network combination model based on weight allocation and error correction is established respectively.At the same time,icing data of different stages are selected to verify the prediction effect of the model.The results show that the combined prediction model can effectively improve the prediction accuracy of the model,and provide a theoretical basis for the airport winter de-icing work.(4)In order to be close to the practical engineering application,the icing grade classification model is further established based on the thickness of icing and meteorological factors,so as to provide theoretical support for the airport staff to take corresponding anti-icing or deicing measures in the future.