Research On Annual Prediction Method Of Low Energy Of Jiangbei Airport Based On Deep Learning

The Civil Aviation Administration of China has placed Chongqing Airport among the top ten domestic aviation coreareas due to the swift advancement of civil aviation transportation.Nevertheless,in recent years,the airport has been severely hindered by the frequent occurrence of low-visibility weather conditions.Therefore,this thesis will analyze the meteorological influence factors of visibility at Jiangbei Airport,and carry out the research on the forecast methods of low visibility.In this thesis,using the meteorological message data of Jiangbei Airport,the change characteristics of low visibility in this region are discussed.The results show that the frequency of low visibility decreases and visibility is gradually improved.Secondly,through the correlation analysis method experiment,the results show that visibility is negatively correlated with air pressure and wind direction,and positively correlated with wind speed and temperature.Then Verifying the correlation between visibility and meteorological factors,the Pearson correlation coefficient method enhances the dependability of the results.Data from the model algorithm.Three algorithms were fashioned,and their parameters,such as neuron count and activation function,were modified.The model algorithm was evaluated.Using LSTM algorithm combined with convolutional neural network,CNN-LSTM algorithm is built,and the parameters of CNN-LSTM algorithm are further adjusted to optimize the algorithm.Two experiments were conducted to assess the efficacy of the algorithm,comparing the low visibility forecast results of CNN-LSTM for Jiangbei Airport,and evaluating the benefits and drawbacks of each algorithm: LSTM,CNN-LSTM,CNN-LSTM,BP neuralnetwork,LSTM,and GRU.The results show that GRU algorithm has the shortest training time.The MAE,RMSE,MAPE and SMAPE values of CNN-LSTM algorithm are 159.76 m,200.52 m,4.38% and 4.42%,respectively.The CNN-LSTM algorithm has significantly improved the performance of low energy prediction.The LSTM algorithm’s accuracy is augmented by 1%,while its training time is augmented by 50% in comparison.Civil aviation operation departments pay much attention to the accuracy of low visibility forecast.The CNN-LSTMalgorithm,in comparison to the customary forecasting approaches,can enhance the precision of the forecast and reduce the forecast duration.Therefore,the low visibility forecast method based on CNN-LSTM algorithm can provide a certain reference for the low visibility forecast of Jiangbei Airport.