Thirty-four Shinkansen noise barrier shapes were tested using 20:1 scale models in order to determine the most effective shape against the main noise sources of Shinkansen trains, namely vehicle lower part noise and current collecting system noise. To evaluate the effects of reducing each source and the total noise, a line source unit was placed at the center of the track as lower part noise and a point source unit was placed at the pantograph location as current collecting noise.
      The experiment showed a Y-shaped wall as the most effective noise barrier; this is due to the enhanced double-diffraction control effect on vehicle lower part noise and the large shielding effect of the wall overhanged externally on current collecting system noise (Table 1). If the wall cannot overhang due to restriction of site boundaries, a modified Y-shaped noise barrier wall can be used. In this case, however, no reduction can be expected on current collecting system noise.
      Reduction of low frequency noise has generally been difficult to implement with noise barrier walls or sound absorbing materials. To improve this, a method of installing sound tubes at the head of the noise barrier wall. This method relies on the phenomenon of sound pressure minimization during resonance at the opening of a 1/4-wavelength sound tube, which brings sound pressure at the head of the noise barrier wall close to zero. This model experiment demonstrated that vehicle lower part noise reduction can be expected at frequencies of 1 kHz or below (Fig. 1).
      Plans for field tests on actual Shinkansen sites are now underway to verify the noise-reduction effects of these methods.