Deep Learning Surrogate for the Temporal Propagation and Scattering of Acoustic Waves

Alguacil, Antonio; Bauerheim, Michaël; Jacob, Marc C.; Moreau, Stéphane

doi:10.2514/1.j061495

Cited by 6 publications

(5 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such a method has been shown sufficient to replicate CFD results from high-order numerical codes, as for instance by Alguacil et al. (2021, 2022) for acoustic waves propagation.…”

Section: Optimisation Of Flapping Wing Motions Using Neural Networkmentioning

confidence: 97%

“…, x(t − (m − 1)τ dt)], with x a state vector, m and τ two integers chosen in the following, and dt = 1/250f . Such a method has been shown sufficient to replicate CFD results from high-order numerical codes, as for instance by Alguacil et al (2021Alguacil et al ( , 2022 for acoustic waves propagation.…”

Section: Neural Network Inputmentioning

confidence: 99%

“…The complete time sequence is then computed by calling successively the shift operator while updating the input history. Alguacil et al (2021Alguacil et al ( , 2022 have used such techniques with a CNN to propagate acoustic waves in both space and time, showing high accuracy and generalisation capabilities compared with the LSTM methods proposed by Sorteberg et al (2018). Colombo, Bauerheim & Morlier (2023) developed a similar strategy based on a graph network used as a shift operator to estimate the unsteady aerodynamic forces and structural loading of a high-altitude long-endurance (HALE) concept aircraft subject to incoming flow perturbations.…”

Section: Introductionmentioning

confidence: 99%

“…Alguacil et al. (2021, 2022) have used such techniques with a CNN to propagate acoustic waves in both space and time, showing high accuracy and generalisation capabilities compared with the LSTM methods proposed by Sorteberg et al. (2018).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Discovering optimal flapping wing kinematics using active deep learning

Corban,

Bauerheim,

Jardin

2023

J. Fluid Mech.

View full text Add to dashboard Cite

This paper focuses on the discovery of optimal flapping wing kinematics using a deep learning surrogate model for unsteady aerodynamics and multi-objective optimisation. First, a surrogate model of the unsteady forces experienced by a 3-D flapping wing is built, based on deep neural networks. The model is trained on a dataset of randomly generated kinematics simulated using direct numerical simulation (DNS). Once trained, the neural networks can quickly predict the unsteady lift and torques experienced by the wing, using sparse information on the kinematics. This fast surrogate model allows multi-objective optimisation to be performed. The resulting Pareto front consists of new kinematics that may be very different from the kinematics of the initial dataset. A few arbitrarily chosen kinematics on the Pareto front are thus simulated using DNS and used to enhance the database. The new dataset is used to train again the networks, and this active deep learning/optimisation framework is performed until convergence, obtained after only two iterations. Overall, this method reduced the cost of optimisation by 83 %. Results reveal two distinct families of motions. Kinematics promoting high efficiency are characterised by large stroke amplitudes and relatively low angles of attack, as observed for fruit flies, honeybees or hawkmoths. For those, lift production is driven by quasi-steady effects and the formation of a stable leading edge vortex. Kinematics promoting high lift are characterised by small stroke amplitudes and high angles of attack, reminiscent of mosquitoes. Lift production is driven by the rapid generation of vorticity at the trailing edge.

show abstract

“…Such a method has been shown sufficient to replicate CFD results from high-order numerical codes, as for instance by Alguacil et al. (2021, 2022) for acoustic waves propagation.…”

Section: Optimisation Of Flapping Wing Motions Using Neural Networkmentioning

confidence: 97%

Section: Neural Network Inputmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Discovering optimal flapping wing kinematics using active deep learning

Corban,

Bauerheim,

Jardin

2023

J. Fluid Mech.

View full text Add to dashboard Cite

show abstract

“…A different approach consists in solving the compressible Navier-Stokes or lattice-Boltzmann (LB) equation to solve both the acoustic and aerodynamic flow fields in a single simulation. The LB method has been applied to simulate outdoor noise propagation (Alguacil et al, 2022;Salomons et al, 2016) and the noise of emerging urban air mobility (UAM) vehicles in a futuristic urban environment (Casalino et al, 2019). Nevertheless, they are computationally demanding and could be unaffordable when the acoustic impact of high-frequency noise or large propagation distances are considered.…”

Section: Introductionmentioning

confidence: 99%

Efficient prediction of airborne noise propagation in a non-turbulent urban environment using Gaussian beam tracing method

Fuerkaiti

Casalino

Avallone

et al. 2023

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

This paper presents a noise propagation approach based on the Gaussian beam tracing (GBT) method that accounts for multiple reflections over three-dimensional terrain topology and atmospheric refraction due to horizontal and vertical variability in wind velocity. A semi-empirical formulation is derived to reduce truncation error in the beam summation for receivers on the terrain surfaces. The reliability of the present GBT approach is assessed with an acoustic solver based on the finite element method (FEM) solutions of the convected wave equation. The predicted wavefields with the two methods are compared for different source-receiver geometries, urban settings, and wind conditions. When the beam summation is performed without the empirical formulation, the maximum difference is more than 40 dB; it drops below 8 dB with the empirical formulation. In the presence of wind, the direct and reflected waves can have different ray paths than those in a quiescent atmosphere, which results in less apparent diffraction patterns. A 17-fold reduction in computation time is achieved compared to the FEM solver. The results suggest that the present GBT acoustic propagation model can be applied to high-frequency noise propagation in urban environments with acceptable accuracy and better computational efficiency than full-wave solutions.

show abstract

Establishment and validation of a relationship model between nozzle experiments and CFD results based on convolutional neural network

Yu,

Wu,

et al. 2023

Aerospace Science and Technology

View full text Add to dashboard Cite

Deep Learning Surrogate for the Temporal Propagation and Scattering of Acoustic Waves

Cited by 6 publications

References 21 publications

Discovering optimal flapping wing kinematics using active deep learning

Discovering optimal flapping wing kinematics using active deep learning

Efficient prediction of airborne noise propagation in a non-turbulent urban environment using Gaussian beam tracing method

Establishment and validation of a relationship model between nozzle experiments and CFD results based on convolutional neural network

Contact Info

Product

Resources

About