Numerical Simulation And Feature Extraction Of Fluctuation Pressure Acting On The Surface Of High Speed Train

Railway can be regarded as an artery of the economy for a country,and it is a key infrastructure as well as a popular transportation method,playing an important role in the development of our society.In recent years,along with the implement of the proposal for fast progress in railway,the highest speed in operation for the CRH EMUs in China has exceeded 300 Km/h,while the highest speed in experiments for CRH EMUs has been more than 450 Km/h.According to related research,when the speeds exceed about 300 km/h,flow-induced noise becomes much more dominant than rolling noise,which will restrain the advance of speeds for high speed trains.When the Mach number is low,fluctuation pressure acting on the train surface is one of the sources for aerodynamic noise.In order to control and reduce the aerodynamic noise when trains are running,it is desperate to conduct research on the fluctuation pressure.In this paper,there are two main parts,one is the Large Eddy Simulation about fluctuation pressure acting on the train surface,the other one is feature extraction of the simulated statistics.The author adopts commercial software package Fluent to realize the simulation part when the train is running in the open air at the speed of 200 Km/h,300 Km/h,350 Km/h respectively and then employs the advanced data process methods to analyse and try to extract features of the simulated results.Empirical Mode Decomposition can choose the number of decomposed layers adaptively based on the given signal and transfer the original signal into a sum of finite series of Intrinsic Mode Functions.Empirical Mode Decomposition is used in this paper to separate the average pressure and the fluctuation pressure from the pressure acting on the noise of the head train.In this way,the fluctuation pressure in this position could be extracted successfully.Subsequently,Hilbert Spectrum is employed to illustrate the energy distribution of extracted fluctuation pressure in the Time-Frequency domain.In addition,comparisons are made among the fluctuation pressure levels at the three different train speeds.In order to study the energy of extracted fluctuation pressure quantitatively,referring to the number of decomposed layers in EMD,db6 Wavelet is adopted to decompose the fluctuation pressure into different orthogonal and nonoverlapping frequency bands.This study shows that as train speeds increase,so do the amplitudes for each measuring points.The tendencies for the change rules of pressure acting on the head train and the tail train are similar.Amplitudes of pressure for measuring points on the head train are larger than that of the measuring points in symmetrical position on the tail train.The surface pressure can be seen as the sum of average pressure and fluctuation pressure oscillating around it.The fluctuation pressure will also augment with the increase of train speeds.The main energy of fluctuation pressure distributes under 1500 Hz,while the global energy may occur at any frequency.Main energy of the fluctuation pressure has a tendency to a higher frequency domain at a higher train speed.