Application of dimension reduction based multi-parameter optimization for the design of blast-resistant vehicle

• DOI: 10.1007/s00158-017-1696-2
• Publication Date: 2017-05-08T23:43:35Z
• Authors: Ran Wei; Xianhui Wang; Ming Zhang; Yunbo Zhou; Liangmo Wang
• Keywords: 8. Economic growth; 0101 mathematics; 01 natural sciences

The design of blast-resistant vehicle provides an appropriate level of protection for the vehicle and occupants against the serious threat from landmine and improvised explosive devices (IED). Mathematically, the objective of this research is to optimize the configuration of light armored vehicle installed multilayer honeycomb sandwich structures (MHSS), shock-mitigating seat and seat belt, and cope with the challenge of highly computational cost on dealing with the large scale, multi-parameter, nonlinear and fluid-structure interaction simulation models. Multi material Arbitrary Lagrangian-Eulerian (MM-ALE) method is used to obtain the high-precision vehicle responses and occupant injuries under blast wave. The baseline model validated by the blast test is built, and the optimization model for blast-resistant vehicle is defined. Then, identifying important design parameters accurately is so difficultly when sufficient samples are not provided due to the expensive computational cost, and it’s inappropriate to screen parameters with inconsistent sequence of variable sensitivities for occupant injuries. Factor analysis based multi-parameter optimization (FAMO) is proposed to reduce the computational cost on improving the blast resistant performance of vehicle. The normal-boundary intersection (NBI) and the R 2 metric are used to obtain optimal compromise solution which noticeable reduced the peak value of occupant injuries, and the physical insights which drive the optimal solution to reduce the occupant injuries is analyzed. Additional studies are conducted on comparing between proposed algorithm and the conventional algorithm, including shape of the Pareto front, optimal compromise solution, design variables and responses.