Design Optimization of Sandwich Composite Armors for Blast Mitigation using Bayesian Optimization with Single and Multi-fidelity Data 2020-01-0170
The most common and lethal weapon against military vehicles is the improvised explosive device (IED). In an explosion, large cabin’s penetrations injure the lower body, and high accelerations harm the spine. Critical penetrations and accelerations can cause the death of the vehicle’s occupants. In the military industry, there is an increasing interest to improve the blastworthiness of their vehicles. This investigation presents a multi-fidelity Bayesian optimization (BO) approach to design sandwich composite armors for blast mitigation. BO is an efficient methodology to solve optimization problems that involve black-box functions. The black-box functions of this work are the responses of the finite element (FE) simulations of the composite armor. The main two components of BO are the surrogate model of the black-box function and the acquisition function that guides the searching process. In this investigation, the surrogate models are Gaussian Process (GP) regressions and the acquisition function is the multi-objective expected improvement (MEI) function. Information from low and high fidelity FE models trains the GP surrogates. The low fidelity FE model assumes elastic behavior of the sandwich composite. The high fidelity FE model considers the nonlinear behavior of each layer of the armor. The sandwich composite is made of four layers: steel, carbon fiber reinforced polymer (CFRP), aluminum honeycomb and CFRP. The design variables are the thickness of each layer and the fiber orientations of the fibers of the CFRP laminas. The optimization problem includes constraints over the thickness of each layer and fiber orientation. Two objectives are minimized, the cabin’s penetrations and the reaction force at the armor’s supports. The minimization of the reaction force works as a mechanism to reduce vehicle’s acceleration. The results show that the multi-fidelity BO approach produces armor designs with performance similar to optimal designs from high fidelity BO but at a lower computational cost.
Homero Valladares, Andres Tovar
Purdue University, Indiana University Purdue University Indianapolis