Abstract: In order to reduce the air frictional noise at the air-carriage interface of high speed train, a Koch type snowflake surface structure is designed to simulate the fractal structure of snowflake and analyzed with the theory of computational fluid dynamics (CFD) and the method of bionic design for the surface of high speed train. The compressible N-S equation is solved using unstructured hybrid grid and model and the noise distribution is solved based on broadband noise source. Simulation results demonstrate that the air frictional noise is greatly reduced because of the secondary vortex generated in the Koch snowflake like surface structure. A kind of multi scale surface structure consists of Koch snowflake cells on two scales and a sort of compound surface texture combining cells of Koch snowflake and circular dents are designed in order to further explore the influence of surface texture on the decrease of the noise of high speed train; that multi scale surface texture is better than single texture but compound surface texture is worse than the better one of the single texture in reduction of the air frictional noise, which provide an effective guidance for the study of reducing the air frictional noise of high speed train.

Key words: high-speed train;Koch snowflake surface texture;multi scale surface texture;compound surface texture;noise reduction