Research And Application On Flight Tracks Clustering For Airport Noise Prediction

With the rapid development of civil aviation, the noise pollution problem around the airport is more and more serious. The flight tracks determine airport noise distribution pattern. Therefore, flight tracks play a key role in noise prediction and noise assessment. The massive historical flight tracks data around the airport is accumulated with the airport operation. Research on noise prediction, noise assessment and flight route optimization with the massive data directly will result in the research methods complicated and achieve an unsatisfactory result. Thence, research on flight tracks clustering is important for improvement and control of airport noise.The existing flight track point pair chose method could not achieve one to one match in the space, and it has a strong influence on the clustering results. A flight tracks similarity measure model is proposed based on the vertical distance of track points. 2D and 3D clustering of the flight tracks can be achieved by the K-medoids clustering algorithm. And DB index and Dunn index are used to evaluate the clustering results. The clustering results are input into NoiseMap, the rationality and availability of the model has been verified.In order to estimate and forecast the scope and area of the noise in its totality, a flight tracks similarity method based on the measure of area between in the flight tracks is presented. The proposed model is combined with flight tracks data, aircraft speeds,meteorology data and aircraft engine thrust. The multi-noise factors flight tracks clustering model for airport noise prediction is built in this paper. The flight tracks clustering results are put into INM, The experiments show that the noise influence area and noise value made by the flight tracks within the same cluster are very similar.This paper analyzes the application in airport noise prediction and noise isoline drawing,using flight tracks clustering results by the proposed research on multi-noise factors flight tracks clustering model for airport noise prediction based on the airport operation and airport noise data. For experimental verification and analysis, flight tracks data of the second quarter of 2014 Beijing capital airport is selected. Experimental results show that the proposed model can reduce the input number of flight tracks into INM, and the calculation of noise isoline is more efficient without influence of the accuracy of noise prediction.