2001 年 67 巻 662 号 p. 2478-2486
In order to simulate the flow under high speed trains, a pair of mirror image models are placed symmetrically, with the symmetric plane assumed to be the moving ground. At the bogie section whose aeproacoustic noise is larger than that from any other sources at lower parts of the train, the shear flow separated at the leading edge is unstable and disturbance grows finally to touch the bottom of the cavity of the section. The cavity tone at the bogie section is modeled as being generated by the pressure fluctuation caused by the interaction of vortices with the trailing edge of the bogie section. When the Green's function adapted to the shape is introduced accompanying the ground regarded as an infinite plane wall, the result is nearly equal to the formula of sound generated by the mirror image models. Mirror image models are verified to be suitable, only if the shear flow at the open end of the cavity reproduceds the real flow under the train.