Accelerating the Design of Morphing Structures and Topology Optimization Surrogates

Dr. Philip Buskohl

Research Engineer
U.S. Air Force Research Laboratory

Seminar Information

Seminar Series
Mechanics & Materials

Seminar Date - Time
May 22, 2023, 11:00 am
-
12:15

Seminar Location
EBU2 Room 479, Von Karman-Penner Seminar Room
In person only


Abstract

Computational design optimization is a key driver for the discovery of innovative solutions, improved efficiencies and mechanistic understanding for many of today’s engineering and fundamental research challenges.  In this seminar, we will discuss two approaches to accelerate design, motivated from applications in morphing structures and topology optimization.  First, we will investigate the discovery of low energy folding paths of morphing origami structures. The interplay between stretching, folding, and facet bending modes in origami structures generates a complex energy landscape of multistable states. However, identifying rigid and deformable folding paths in this high-dimensional and non-convex energy landscape is challenging. To help address this, we leverage the nudged elastic band algorithm to identify low energy folding paths between stable configurations in this landscape.  The nudge elastic band technique is a global method that optimizes each point along the path to follow the gradient of the energy function, while also maintaining separation between the points to provide adequate resolution along the path.  We demonstrate the utility of this approach for discovering sequenced folding motions, distinguishing between rigid and deformable folding paths, and finding low energy folding paths between stable states in multistabilty origami structures. In the second half of the talk, we will discuss another strategy to accelerate design optimization using generative adversarial neural networks as a surrogate model of topology optimization. The neural network surrogate maps an image representation of the design problem boundary conditions to a prediction of the corresponding topology with minimum compliance, resulting in a significant computational speed-up.  Collectively, these tools highlight two potential strategies to accelerate design optimization and point to opportunities to generalize these concepts to other problems.

 

Speaker Bio

Philip R. Buskohl is a Research Engineer in the Functional Materials Division at the U.S. Air Force Research Laboratory.  The Division delivers materials and processing solutions to revolutionize AF capabilities in Survivability, Directed Energy, Reconnaissance, Integrated Energy and Human Performance.  Phil has authored over 40 peer-reviewed papers ranging from the chemical-mechanical feedback of self-oscillating gels, design of reconfigurable origami structures and mechanical computing concepts. His research interests include nonlinear elasticity, topology optimization for material design, and mechanically adaptive materials.