Reduction in Aerodynamic Resistance of High-Speed Train Nose Based on Kriging Model and Multi-objective Optimization

In order to minimum the aerodynamic resistance of both the head and tail cars, an approach for aerodynamic design of a high-speed train nose based on Kriging model and multi-objective optimization is proposed. The aerodynamic resistances of the head and tail cars are chosen the dependent variables. 5 independent variables are related to 5 typical curves, which are the vertical outline of the front nose, vertical outline of the window, lateral outline of the nose, the first auxiliary outline and the second auxiliary outline. Optimal Latin-hypercube design (OLHD) is used to generate the design matrix of the independent variables of the train nose. The dependent variables for every set of independent variables are obtained using computational fluid dynamics. The multi-objective optimization is solved by Non-dominated Sorting genetic algorithm II (NSGA-II). Compared to the initial value which is calculated using the original train model, the aerodynamic resistance of the head car decreased by about 6.5%, and the resistance of the tail car decreased by 5.0%. The magnitude of the decrease in the resistance of the head car is larger than that of the tail one. The vertical outline of the window is the most sensitive to aerodynamic resistance for both head and tail cars. The first auxiliary outline has a contradictory effect on the aerodynamic resistance of the head and tail car.