Surrogate Modeling Accelerated Shape Optimization of Deployable Composite Tape-Spring Hinges

Abstract

Composite tape-spring hinge (CTSH) is a simple yet elegant mechanical component for various deployable space structures. This paper formulates and addresses cut-out shape optimization of a CTSH, which is seldom touched upon in literature. Both the maximum strain energy stored during the folding process as well as the maximum bending moment during deployment were maximized in a concurrent way, and the multi-objective optimization problem was realized by merging data-driven surrogate modeling and shape optimization. Four different surrogate modeling techniques (radial basis function, kriging, Gaussian process regression, and artificial neural network) are evaluated and compared. The maximum stored strain energy at the fully folded state and the maximum bending moment during deployment for the optimal CTSH are increased by 50 and 35%, respectively, compared to the initial design under a previously developed composite failure criterion as constraint. To ensure reproducibility and foster future research, we publicly share our full implementation with the source codes and trained models with the community.