GPU CABARET Solutions for the CoJen Jet Noise Experiment

Markesteijn, Anton P.; Karabasov, Sergey A.

doi:10.2514/6.2018-3921

Cited by 11 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Poised to deliver greater than 1.5 exaflops of HPC and AI processing performance, the Frontier system will feature future-generation AMD CPUs and Radeon GPUs. Some jet simulations have already been performed on GPUs by Markesteijn, Semiletov & Karabasov (see [128] and the follow-on studies [129] with references therein) and showed promising results in terms of performance. The authors report that the GPU solver was able to handle computational grids up to 80 million grid cells on a conventional desktop computer equipped with a few GPUs, while obtaining acoustic results within several days.…”

Section: Remaining Challenges and Upcoming Opportunitiesmentioning

confidence: 99%

Modelling of jet noise: a perspective from large-eddy simulations

Brès

Lele

2019

Phil. Trans. R. Soc. A.

View full text Add to dashboard Cite

In the last decade, many research groups have reported predictions of jet noise using high-fidelity large-eddy simulations (LES) of the turbulent jet flow and these methods are beginning to be used more broadly. A brief overview of the publications since the review by Bodony & Lele (2008, AIAA J. 56 , 346–380) is undertaken to assess the progress and overall contributions of LES towards a better understanding of jet noise. In particular, we stress the meshing, numerical and modelling advances which enable detailed geometric representation of nozzle shape variations intended to impact the noise radiation, and sufficiently accurate capturing of the turbulent boundary layer at the nozzle exit. Examples of how LES is currently being used to complement experiments for challenging conditions (such as highly heated pressure-mismatched jets with afterburners) and guide jet modelling efforts are highlighted. Some of the physical insights gained from these numerical studies are discussed, in particular on crackle, screech and shock-associated noise, impingement tones, acoustic analogy models, wavepackets dynamics and resonant acoustic waves within the jet core. We close with some perspectives on the remaining challenges and upcoming opportunities for future applications. This article is part of the theme issue ‘Frontiers of aeroacoustics research: theory, computation and experiment’.

show abstract

Section: Remaining Challenges and Upcoming Opportunitiesmentioning

confidence: 99%

Modelling of jet noise: a perspective from large-eddy simulations

Brès

Lele

2019

Phil. Trans. R. Soc. A.

View full text Add to dashboard Cite

show abstract

“…The researchers mentioned that two parallel GPUs accelerated the running time by 150X compared to a serial run on CPUs. A few years later, in 2018, Markesteijn and Karabasov [27] developed a GPU solver that would accelerate LES calculations for simulating subsonic coaxial jet noise at three different operation points. The GPU solver was combined with the penetrable integral surface formulation of the FW-H method.…”

Section: Literature Reviewmentioning

confidence: 99%

Efficient Implementation of the Lattice Boltzmann Method on a GPU for Predicting Near-Field and Far-field Jet Noise

Noah,

Abdelrahim,

Mosa

et al. 2024

Preprint

View full text Add to dashboard Cite

The lattice Boltzmann method (LBM) is a powerful technique for the computational modelling of a wide variety of single and multiphase flows in complex geometries. It is a discrete computational approach based on solving the Boltzmann equation numerically. LBM has recently become one of the new and promising CFD methods to solve many computational fluid mechanics problems. LBM works particularly well when solving incompressible flow problems, but limitations arise when solving compressible flow, especially at high Mach numbers. Compressible LBM was applied for the simulation jet flow in order to demonstrate the ability of the fifth-order equilibrium distribution function to simulate compressible flows efficiently. The near-field flow physics and noise simulations were performed using the compressible LBM. The results from the LBM simulation were then employed in Kirchhoff’s surface integral approach to predict far-field jet noise. Because of the ability of the lattice Boltzmann technique to be used in parallel computing and to improve LBM computational efficiency with respect to the numerical simulations of turbulent flows in predicting far-field noise, the final step in this research was to use the compute unified device architecture (CUDA) for implementing the LBM in a graphics processing unit (GPU), thus creating a hybrid code, LBM-MRT-LES, through the utilization of the Kirchhoff integral method, a powerful tool for simulating aeroacoustics problems.

show abstract

“…Thanks to the GPU implementation of the CABARET solver, a considerable reduction of the flow solution time is achieved in comparison with conventional LES approaches similar to the pervious jet flow simulations [37,38,39]. The simulations are performed on a single computer workstation equipped with two GPU cards (NVidia Titan RTX 24GB).…”

Section: Solution Of the Burgers' Model Of Nonlinear Wave Propagation...mentioning

confidence: 99%