Research On Rapid Identification And Prediction Of Aircraft Wake Vortex

Aircraft wake vortex is a kind of vortex effect produced by wing lift.Wake vortexes will have an important impact on the flight safety,efficiency,and fuel consumption of subsequent aircraft,and are crucial to the safety of civil aviation flights.Therefore,separation measures must be taken to reduce the impact of wake turbulence and ensure air traffic safety.With the development of the civil aviation industry,the increase of aircraft wingspan,passenger capacity and flight speed has brought new challenges to wake vortex separation and subsequent aircraft safety.Therefore,the optimization of aircraft wake vortex separation has become an important research field.In recent years,the rise of artificial intelligence and the development of deep learning algorithms have provided a new way for the realization of wake vortex dynamic separation system based on lidar detection.This paper takes the aircraft near-ground wake vortex as the research object,and realizes rapid wake vortex identification and prediction by designing a wake vortex detection model and using a deep learning algorithm.The main work includes:Firstly,the detection principle of lidar is studied.According to the evolution characteristics of aircraft wake vortex,the laser radar scanning mode of aircraft wake vortex and the near-earth detection model are determined,and the simulation of aircraft wake vortex detection is carried out.The on-site detection of aircraft wake vortex was carried out,and the on-site detection and data set establishment of aircraft wake vortex radial velocity data in Shenzhen Baoan Airport and Shuangliu Airport were completed by using lidar,and the characteristics of the constructed data sets were deeply analyzed.Then,in terms of rapid identification of aircraft wake vortices,a lightweight deep learning framework is proposed to identify aircraft wake vortices in the wind field of airport arrival and departure segments.The framework uses a patch embedding module as the input representation,combined with a depthwise separable convolution module and a hybrid attention mechanism to enhance the models attention to the spatial location of the wake,and encode the environmental factors that affect the wake behavior into the model.Model test results show that the proposed network can identify wake vortices in lidar faster and more accurately than traditional methods.Finally,in terms of rapid aircraft wake vortex prediction,a space-time model is constructed based on lidar technology,and convolutional long-short-term memory networks are used to predict aircraft wake vortices in space-time.The neural network model uses the currently observed aircraft wake radial velocity field to infer and predict the aircraft wake radial velocity field for the next lidar scan.Experimental results show that the proposed model can effectively extrapolate the spatiotemporal evolution of the aircraft wake vortex.In this paper,by exploring the shape and motion of the aircraft wake vortex,we can better understand the characteristics of the wake vortex,and provide a more scientific and accurate reference and guarantee for the subsequent safe flight of the aircraft.The radar detection aircraft wake vortex model,the aircraft wake vortex identification model,and the aircraft wake vortex prediction model provide technical support for the rapid reduction of aircraft wake dynamic intervals.has practical value.