Multi-objective optimization for the aerodynamic noise of the high-speed train in the near and far field based on the improved NSGA-II algorithm

With the increased running speed of trains, the aerodynamic noise of trains becomes increasingly obvious. Reducing aerodynamic noises has become one of keys to controlling the noise of high-speed trains. This paper conducted a numerical simulation on the aerodynamic noise of head of the high-speed train. Firstly, this paper established a mathematical-physical model for the three-dimensional turbulent flow field of a high-speed train, adopted standard k-epsilon equation turbulent model and broadband noise source model to compute the aerodynamic noise sources of the high-speed train and applied three-dimensional transient large eddy simulation (LES) to compute the external unsteady flow field of the high-speed train after obtaining noise sources. Based on the unsteady flow field, then this paper applied FW-H equations to compute the far-field aerodynamic noise of the high-speed train. After obtaining the unsteady fluctuation pressure on the surface of the train, this paper computed the radiation characteristics of aerodynamic noises around the high-speed train based on the boundary element method (BEM). Researched results showed: The main aerodynamic noise sources of the high-speed train were at the nose tip of head train; fluid separation and recombination were main reasons for the aerodynamic noise of the highspeed train; vortexes in the position of head train were striped and horseshoe-shaped or hairpin vortexes were mainly in the area of tail train; in addition, vortexes were symmetrically distributed along the longitudinal symmetry plane of train; dipole noises were mainly distributed in the area of head train, whose main energy was decreased with the increased frequency; the quadrupole noise of aerodynamic noises of the high-speed train was mainly distributed in the wake flow area of tail train; when the high-speed train ran at the speed of 300 km/ h, the maximum sound pressure level of far-field observation points was 76.8 dB; additionally, aerodynamic noises in the far field were mainly a broadband noise, whose main energy was within the frequency range of 1250 Hz to 3150 Hz. Finally, the improved NSGA-II algorithm was used to conduct a multi-objective optimization for the head shape. The aerodynamic drag of the high-speed train could be most reduced by 6.74 %, and the dipole aerodynamic noise source could be most reduced by 8.34 dB. The improved NSGA-II algorithm has an obvious effect on the multi-objective optimization of the head shape.