This conference paper reveals cutting-edge findings on the application of GPU-accelerated large-eddy simulations (LES) to the NASA Fan Noise Source Diagnostic Test (SDT) benchmark.
Why this matters
Reducing aircraft noise is crucial for improving environmental impact and passenger comfort. This study provides:
- Detailed insights into noise generation mechanisms in high bypass ratio turbofan engines
- Utilization of advanced computational techniques to improve simulation accuracy and efficiency
- Data that supports the development of quieter, more efficient engine designs
Key highlights:
- Comprehensive simulation domain: Our simulations include the full annulus of the fan, OGVs, nacelle, and the entire test section
- Simulation accuracy: We compare LES results with experimental data, demonstrating excellent agreement in terms of aerodynamic performance, unsteady flow fields, and far-field noise
- GPU acceleration: Leveraging GPU acceleration, we achieve substantial computational efficiency, with results for a 140-million-cell mesh obtained in approximately 6 hours using 40 standard GPUs.
Our work highlights the potential of GPU-accelerated simulations to advance aeroacoustics research, providing valuable insights into noise generation mechanisms in modern turbofan engines.
This conference paper was first presented at the AIAA AVIATION 2023 Forum - Turbomachinery and Core Noise IV session - AIAA 2023-4299.