Validation of a geometric diffraction model with respect to level-time history prediction

Abstract

Noise induced wake-up reactions strongly depend on single noise events, which are characterized by the maximum sound pressure level and the steepest slope of rise in level time-history. In this paper the applicability of the CNOSSOS-EU guidelines in terms of assessing the level-time history of moving sound sources is investigated. Generally, major changes in noise level occur due to shadowing effects. Therefore, the focus of investigation lies on the diffraction model. This model distinguishes between three different diffraction cases: the partial diffraction, when the direct ray comes close to an edge, the single diffraction and the multi diffraction. All three diffraction cases are investigated in one scenario of a sound source passing-by a box. The results are compared to numerical simulations using a Discontinuous-Galerkin method. The comparison shows an overestimation of the level by the diffraction model in case of double diffraction, while in the single diffraction case an underestimation occurs. The deviation between the Discontinuous-Galerkin and the geometric model in case of partial diffraction is below 3 dB. In terms of level history, the geometric diffraction model shows an unrealistic discontinuity when the model switches between the single and double diffraction case. The jump is increasing with higher frequencies.