Multi-objective optimization of additive manufactured functionally graded lattice structures under impact

A multi-objective optimization procedure to effectively design gradient lattice structures under dynamic loading is presented. It aims both at maximized energy absorption characteristics and a lightweight design. The reduction of the peak crushing force and a maximized specific energy absorption are simultaneously optimized. Design variables like the relative density and density gradient of the lattice structure are investigated, considering different topologies of the lattice unit cells. To model and simulate the lattice structures efficiently, a simplified beam-based simulation finite element model is introduced and validated against experimental tests. They cover the range from dynamic equilibrium up to compaction wave propagation as impact deformation phenomena. Based on large numbers of simulations, an artificial neural network is used to predict the energy absorbing characteristics for the given design parameters and to find the optimal configuration of the lattice structures. The neural network is trained using a multi response adaptive sampling algorithm allowing parallel simulation with automatically generated finite element models during each iteration step. Subsequently, a multi-objective genetic algorithm is used to find optimal combinations of the design parameters of lattice structures under different impact velocities and cell topologies.