Asymmetric instability of flow-induced vibration for elastically mounted cube at moderate Reynolds numbers
Journal of Fluid Mechanics, 996, A34.
Postdoctoral Associate · MIT
I study how unsteady flows interact with flexible structures, and develop computational models for predicting vibration, understanding noise, and improving engineering design. During my Ph.D., I worked with Professors Fue-Sang Lien at the University of Waterloo and Earl H. Dowell at Duke University.
Postdoctoral Associate at MIT. In the Department of Aeronautics and Astronautics, I work with Professor Adrian Lozano Duran on machine-learning-based closure models for LES modeling.

About my work
My research centers on fluid-structure interaction (FSI) and flow-induced noise in aerospace and energy engineering, with technical strengths in high-fidelity multiphysics modeling, aeroelasticity and flow-induced vibration (FIV), reduced-order modeling, machine learning, and bio-inspired structural design. I use these capabilities to improve the safety, efficiency, resilience, and acoustic performance of aircraft and renewable-energy systems.
During my Ph.D. at Waterloo and Joint Ph.D. at Duke (2019-2023), I developed computational and reduced-order models for FIV and aeroelastic stability from low-speed to supersonic flow. My work revealed the role of geometric tapering in galloping instability and established reduced-order models for rapid aeroelasticity prediction, attracting interest from the NASA Aeroelasticity Group.
At UBC, I worked with Mitacs, Clear Seas, and Robert Allan Ltd. on FSI, cavitation, and machine learning for safer and quieter propulsion systems.
At UC Berkeley (2024-2026), I developed bio-inspired flexible-wing vehicles by combining high-fidelity multiphysics simulation with physics-based machine learning. I further extended FSI design to internal structural topology for lightweighting, vibration suppression, and aeroelastic resilience.
I am a Postdoctoral Associate at MIT, working with Adrian Lozano Duran to improve the predictive capabilities of LES models for aerospace applications through machine-learning-based closure models.
Education & experienceResearch in focus
From the onset of structural instability to the radiation of sound, I investigate the mechanisms that connect flow, motion, and performance.
Internal structural topology regulates flow-structure energy transfer for passive aeroelastic control. High-fidelity FSI models connect structural architecture to deformation and flow-induced response, guiding bio-inspired design for aeroelastic resilience.
Internal topological design can regulate flow-structure energy transfer for passive aeroelastic control.
Cheng, Zhi; Dowell, Earl H.; Gu, Grace X.
Under review
Structural and fluid reduction, GNNs, and applications in galloping, transonic aeroelasticity, topology design, and active morphing.
02Acoustic and vortex-particle models for wind turbines, propeller singing, integrated propulsion analysis, and digital-twin control.
Selected publications
Journal of Fluid Mechanics, 996, A34.
Journal of Fluid Mechanics, 967, A18.
Journal of Fluid Mechanics, 986, A16.
Engineering Applications of Artificial Intelligence, 182, 116051.
International Journal of Multiphase Flow, 186, 105149.
J. Fluid Mech., 1031, A23 (2026).