Investigations on prediction of ship noise using the FWH acoustic analogy with incompressible flow input

Artikel i vetenskaplig tidskrift, 2022

Ship noise predictions using FWH acoustic analogy with incompressible flow solution inputs are investigated for a model scale container vessel with a cavitating propeller. Numerical predictions of cavitation and hull pressure pulse predictions are validated first, comparing simulations performed for the tunnel test section and experimental measurements inside a large-size cavitation tunnel. The predictions agree well including the sheet cavitation development, tip vortex cavitation (TVC) bursting, convex shaped sheet cavitation closure line and the traveling re-entrant jet underneath the sheet cavity triggers the TVC bursting behavior. Noise predictions are performed within a large open simulation domain instead of the cavitation tunnel test section. With incompressible solutions, noise levels are predicted based on two different placements of Permeable/Porous Data Surfaces (PDS) where one encloses the cavitating propeller, rudder and downstream wake (PDS−L1) and one encloses the whole ship as a rectangular box (PDS−L2). The FWH noise predictions with impermeable surfaces (S−FWH) are also studied. Differences between predictions using PDS−L1, PDS−L2, and S−FWH are discussed. To compare calculated noise source level (Ls) at different noise receivers at varying distances, normalization assuming spherical spreading acoustic wave is used. In certain combinations of receiver point and method of acoustic computation, the predicted Ls agreed well comparing to experimental measurements, including the prediction with PDS−L2 and receiver close to the PDS and direct probed incompressible hydrodynamic pressure at similar receiver locations. However, with increasing distance to the receiver, the predicted Ls increases for higher frequencies and levels out at unrealistically high levels. To study this phenomenon, a free-field monopole representing the cavity structure dynamics is tested with different combination of PDS and receiver placements, using both incompressible and compressible input. This analysis gives a clear indication that the origin of this erroneous effect is the combination of the FWH acoustic analogy with an incompressible solver.