February 20, 2023: Leixin Ma
February 20, 2023
Room: 310 Kelly Hall
Leixin Ma, Department of Mechanical and Aerospace Engineering, University of California, Los Angeles.
Faculty Host: Dr. Todd Lowe
"Physics-Constrained , Data-Driven Approach for Optimal Design of Flexible Structures"
Abstract: Traditional engineering structures are made of rigid structures with simple geometry, such as robotic arms and wind turbines. In contrast, recent advancements in manufacturing open the door to designing and manufacturing soft structures with programmable and unconventional functionalities. An optimal and efficient design of these flexible structures interacting with external loads, such as fluids, remains challenging. In this talk, I will discuss how a combined physics-based modeling and machine-learning approach can help tackle these challenges. The first two challenges are the extraordinarily large structural design space and large structural deformation that are not amenable to conventional analytical tools. I will introduce a symmetry-constrained machine learning technique, combining Variational Autoencoder and Bayesian Optimization, to design soft composite shells of targeted functionalities. These soft composites – without rigid springs and hinges - are manufactured by combining kirigami and pre-stretch and can be scalably fabricated on a 2D plane. Despite fully planar fabrication, they can be programmed to assume a prescribed 3D shape without any external stimulus. The other challenge is the interaction between deformable structures with fluid flow. I will introduce my contribution to developing a physics-constrained data-driven approach to identifying the most critical parameters governing the flow-induced vibration of flexible cylinders. The techniques enable us to reveal new physical insights and construct much simpler structural response prediction models without compromising the prediction accuracy and physical consistency. Finally, I envision how my future investigation plan can fit with and broaden the researchcommunity at Virginia Tech. A physics-constrained and data-driven approach can help design novel programmable soft structures and robots, which could find applications in renewable energy systems and bio-inspired propulsion.
Bio: Leixin Ma is currently a postdoctoral fellow in the Department of Mechanical and Aerospace Engineering at UCLA. She is also senior research personnel at UCLA Clean Energy Smart Manufacturing Innovation Institute. She received her BSc in Naval Architecture & Ocean Engineering from Shanghai Jiao Tong University in 2015, an S.M. degree, and a Ph.D. degree in mechanical engineering from MIT in 2017 and 2021, respectively. Her research interest spans a wide range from dynamics and controls, flow-induced vibration, interpretable machine learning-aided science discovery, and the data-driven design of smart and soft structures. She was awarded an Excellent Bachelor Thesis (Top 1%) of Shanghai Jiao Tong University in 2015 and the Ho-Ching and Han-Ching Fund Award from MIT Office of Graduate Education in 2019. She also served as Teaching Development Fellow at MIT from 2020-2021.