Research On Aircraft Wake Vortex Recognition Based On Multi Radar Detection

Aircraft wake vortex is a pair of persistent vortices with the same strength formed behind the wing when the wing generates lift.The aircraft crash event caused by the wake vortex has great influence.Scholars at home and abroad have conducted in-depth research on the evolution and detection of aircraft wake.Aiming at the problems of insufficient data acquisition density and incomplete coverage in single radar detection,focusing on the research of lidar detecting wake vortex data,the method system of multi-radar detecting wake vortex was established,the visualization system of wake vortex data was constructed,and the identification of wake vortex was studied.The main research work of thesis includes:By analyzing the principle of lidar detection,the lidar was selected to met the precision of detecting aircraft wake vortex,and the characteristics of aircraft wake vortex are analyzed,including the radial velocity model,vortex core position,vortex core radius,vortex spacing,vortex circulation,wake vortex dissipation model,etc.,and the scanning mode and location of lidar were determined.A framework method of multi-radar detection of aircraft wake vortex was proposed.Secondly,aiming at the problems of repeated data,noise data and null data in the wake vortex data detected by lidar,the wake vortex database was established by cleaning the data and fusing the data of multiple radars.On this basis,the visualization method of wake vortex data was introduced,and the visualization system of wake vortex was constructed.Then,the identification method of aircraft wake vortex based on random forest(RF)was put forward,and the calculation method of the eigenvalues of aircraft wake vortex data was introduced.Based on the established aircraft wake vortex database,the eigenvalue calculation software was designed to extract the eigenvalues of aircraft wake vortex data,and the eigenvalues data set of aircraft wake vortex was made.Through model training,Grid search determines the optimal parameters of the random forest model,and the identification results show that this method can effectively identify the wake vortex of aircraft.At last,the Light GBM model was used to identify the aircraft wake vortex data,and the importance of the eigenvalues was sorted.The results were compared with those of KNN,SVM and RF models,and the results show that the Light GBM model and RF model have a good performance in identifying the aircraft wake vortex data.The method of multi-radar detection of aircraft wake vortex,visualization of aircraft wake vortex system and identification of aircraft wake vortex proposed in this thesis are innovations and improvements of wake detection research methods.The research results of this thesis provide real and reliable data support for the study of aircraft wake evolution mechanism,and provide theoretical basis for the optimization of wake spacing and the establishment of dynamic wake spacing.