Deep Neural Network-Based Prediction of Flow-Induced Noise Around Cylindrical Bodies

-д хадгалсан:

Номзүйн дэлгэрэнгүй
-д хэвлэсэн:	Journal of Marine Science and Engineering vol. 13, no. 11 (2025), p. 2161-2175
Үндсэн зохиолч:	Kim, Minjoon
Бусад зохиолчид:	Im-jun, Ban, Sung-chul, Shin
Хэвлэсэн:	MDPI AG
Нөхцлүүд:	Hydrodynamics Datasets Fluid dynamics Artificial neural networks Sound pressure Signal processing Polar coordinates Fluid flow Pressure Pressure distribution Computer applications Design Cylindrical structures Turbulence models Simulation Fourier transforms Coordinate systems Root-mean-square errors Cylindrical bodies Computational efficiency Geometry Noise Detached eddy simulation Time dependence Turbulence Accuracy Flow velocity Noise prediction Spatial discrimination Pressure dependence Cylinders Inflow Predictions Spatial resolution Neural networks Variables Computing costs Noise generation Computational fluid dynamics Acoustics Environmental
Онлайн хандалт:	Citation/Abstract Full Text + Graphics Full Text - PDF
Шошгууд:	Шошго нэмэх Шошго байхгүй, Энэхүү баримтыг шошголох эхний хүн болох!

Тодорхойлолт
Хураангуй:	Cylindrical bodies generate flow-induced noise when exposed to external flows, which can be predicted numerically using Computational Fluid Dynamics (CFD) combined with the Ffowcs Williams–Hawkings (FW–H) Equation. Accurate prediction, however, requires turbulence models such as Detached Eddy Simulation (DES) with fine spatial resolution and small time steps, in addition to time-dependent surface pressure data and receiver arrangements. These requirements greatly increase computational costs and limit the applicability of such methods during the design stage. To address this challenge, a Deep Neural Network (DNN) model was developed to predict flow-induced noise around a cylinder. Training data were generated from CFD cases using cylinder geometry and inflow velocity as design variables, with multiple receivers arranged in a polar coordinate system. Acoustic signals were computed using Farassat’s Formulation 1A, the time-domain surface solution of the FW–H Equation. The DNN was trained with design variables, receiver coordinates, and octave-band center frequencies as inputs, while the Sound Pressure Level (SPL) served as the output. Model performance was evaluated using the adjusted coefficient of determination (<inline-formula>Radj2</inline-formula>) and the root mean squared error (RMSE). In addition, interpolation capability was tested by varying receiver spacing to examine robustness under sparse data conditions. The results confirm that the proposed framework provides accurate and computationally efficient predictions suitable for early-stage design.
ISSN:	2077-1312
DOI:	10.3390/jmse13112161
Эх сурвалж:	Engineering Database