Abstract
A high-fidelity numerical framework is developed to predict the hydrodynamic noise generated by tidal-stream turbines under realistic flow conditions. The approach couples large-eddy simulations (LES) using the actuator-line method (ALM) with Amiet’s aeroacoustic theory to quantify underwater sound emissions from turbine blades. The model accounts for both trailing-edge and turbulence-interaction noise sources. Simulations are performed for a full-scale 1 MW, 18 m-diameter tidal turbine operating at various tip-speed ratios and inflow turbulence intensities. The results show that higher tip-speed ratios lead to amplified hydrodynamic noise, while increased turbulence broadens the affected frequency range. Cylindrical sound propagation is applied to represent shallow-water effects, and predictions are extended to a nine-turbine array configuration. The array increases noise levels by up to 8 dB downstream and over 50 dB laterally compared with a single turbine. The coupled LES–Amiet framework provides a computationally efficient and physically consistent method to predict underwater noise from tidal energy devices. These findings improve the understanding of the flow-acoustic coupling mechanisms and support the design and environmental assessment of future marine energy systems.