Embedded Files
Welcome to the Computational Laboratory for Aerodynamics and Aeroacoustics Research (CLAAR) at the Georgia Institute of Technology! Our laboratory is a part of the Guggenheim School of Aerospace Engineering and affiliated with the Vertical Lift Research Center of Excellence (VLRCOE) at Georgia Tech.
Led by Prof. Beckett Zhou, this laboratory focuses on the development of efficient numerical methods to help elucidate complex flow physics and noise generation mechanisms of both fixed-wing and rotary-wing configurations and optimize their aerodynamic and acoustic performance. By forging strong synergies between simulation, optimization and data-driven methods, the over-arching research goal of our team is to enable the use of these methods in the early stage of an aircraft design cycle to help investigate novel noise reduction concepts, unconventional configurations, and explore previously uncharted design spaces, in the quest for zero-emission and low-noise aircraft concepts required for future flight.
The main research areas of our laboratory include: high-fidelity scale-resolving methods for turbulent flow and noise prediction, full-vehicle time-domain acoustic scattering, efficient adjoint-based methods for aerodynamic and aeroacoustic optimization, broadband noise modeling, data-driven turbulence and noise modeling.
Hybrid RANS/LES, Wall-Modelled/Resolved LES and Lattice-Boltzmann Method
In-depth investigation of flow physics and noise generation mechanisms
Full space-time Galerkin discretization of the energetic weak form for the boundary integral equation
Scattering of broadband, transient and rotating noise sources
Complete vehicle acoustic simulation in seconds on GPU
Coupled CFD-CAA discrete adjoint via algorithmic differentiation
Machine-accurate and dual-consistent adjoint gradient
Highly efficient design sensitivity evaluation and accurate enforcement of design constraints
RANS-based fast Random Particle-Mesh (fRPM) method for turbulent boundary layer noise prediction
Random Vortex Particle Method (RVPM) for blade-wake/blade-vortex interaction noise prediction
Physics-informed field-inversion machine learning to optimally enhance turbulence and noise models with limited high-fidelity data
Multi-fidelity noise model development using active and transfer learning
Page updated
Google Sites
Report abuse