DRL Stabilizes Hypersonic Inlet Unstart Phenomenon
Ameya Jagtap et al. (arXiv v1 submitted Jan 27, 2026) demonstrate a deep reinforcement learning (DRL) active-flow-control strategy to prevent unstart in a canonical two-dimensional hypersonic inlet at Mach 5 and Reynolds number 5×10^6, using a high-fidelity CFD solver with adaptive mesh refinement. The learned controller stabilizes inlet operation across wide back-pressure ranges, generalizes zero-shot to unseen Reynolds numbers and sensor configurations, tolerates noisy measurements, and uses a minimal sensor set for practical deployment.
Key Points
- 1Demonstrates DRL controller stabilizes Mach 5 inlet unstart across wide back-pressure and Reynolds conditions
- 2Uses high-fidelity CFD with adaptive mesh refinement to resolve shocks, boundary layers, and separations accurately
- 3Enables practical deployment via zero-shot generalization, noisy-sensor robustness, and minimal optimally selected sensor set
Scoring Rationale
Strong novel DRL demonstration in high-fidelity hypersonic CFD, limited by single preprint and simulation-only validation.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
