Time domain model of underwater sound propagation for offshore wind turbine

This paper introduces a numerical method for modeling underwater sound propagation in the time domain, specifically applied to offshore wind turbines. Given that the majority of wind turbines are situated in ocean depths less than 100 m, termed shallow water, the paths of underwater sound propagation become intricate, particularly toward specific receptor positions. Dealing with the complexities of multi-path reflection waves and scattering of the edges, especially in the shallow water regime, incurs significant computational costs in the time domain. To analyze sound propagation in moving inhomogeneous media within the time domain, a numerical technique was implemented. Treating the top and bottom of the ocean as acoustically hard surfaces and considering the horizontal direction as an infinite boundary, the perfectly matched layer is implemented. To address the high computational demands, GPU parallelization was employed, significantly reducing computation time, particularly for large-scale simulations. The validity of the algorithm is further verified through underwater sound propagation experiments. The results demonstrate that the differences in transmission loss between measurements and simulations at the target frequencies remain within 10 dB.