Ph.D. Defense

Shreyas Ashok

(Advisor: Prof. Juergen Rauleder)

"Development of a Lattice-Boltzmann Method Simulation Tool for Real-Time Ship–Rotorcraft Interactional Aerodynamics"

Thursday, July 17^th 

10:00 a.m. 
Montgomery Knight 317 

 Microsoft Teams
 

Abstract

Rotorcraft ship-deck landing operations pose a considerable challenge for pilots because of the highly unsteady aerodynamic interactions between the turbulent ship airwake and rotor wake. For effective piloted flight simulator training, accurate and quick-turnaround rotorcraft interactional aerodynamics models are highly desirable. In particular, real-time capable two-way fully-coupled simulations which resolve the effect of the rotor wake on the ship airwake (and vice versa) may be necessary to capture the interactional aerodynamics. While traditional Computational Fluid Dynamics (CFD) simulations can resolve these interactions, they are computationally cost-intensive and can take days or weeks to calculate on High-Performance Computing (HPC) clusters. Conversely, many low-fidelity methods neglect important physics. A mid-fidelity real-time capable methodology for ship–rotorcraft aerodynamic simulations is highly desirable to improve the fidelity of flight simulation models. 

A candidate methodology for interactional aerodynamics simulations is the Lattice-Boltzmann Method (LBM). The LBM is highly tailored towards efficient parallelization, especially on Graphics Processing Units (GPUs). The high computational efficiency of the LBM makes it a strong candidate for real-time capable ship airwake and ship–rotorcraft interactional modeling.

This thesis presents the development of a GPU-accelerated LBM model for ship–rotorcraft aerodynamic interactions. Simulations of the NATO Generic Destroyer ship airwake correlated well with experimental data. Real-time capable simulations of a notional UH-60 helicopter approaching a ship deck were conducted using both one-way and two-way couplings, and the coupling methods produced significantly different pilot control input frequency spectra. The ship–rotorcraft simulations were validated against new Georgia Tech wind tunnel experiments, showing that the LBM captured relevant interactional aerodynamic phenomena well. Finally, a comparative study between LBM and a comparable Navier–Stokes CFD solver was conducted, and the LBM was 4–6x faster than the Navier–Stokes CFD solver on the same GPU hardware.

Committee

• Prof. Juergen Rauleder – School of Aerospace Engineering (advisor)
• Prof. Marilyn Smith – School of Aerospace Engineering
• Prof. Beckett Zhou – School of Aerospace Engineering
• Dr. Dylan Jude – Aerospace Engineering Consultant, Science and Technology Corporation
• Dr. Andrew Wissink – Director, HPCMP CREATE Applied Surrogates Institute, US Army DEVCOM Aviation and Missile Center