Compared with the forward design method through the control of geometric parameters and material types, the inverse design method based on the target stress‐strain curve is helpful for the discovery of new structures. This study proposes an optimization strategy for mechanical metamaterials based on a genetic algorithm and establishes a topology optimization method for energy‐absorbing structures with the desired stress‐strain curves. A series of structural mutation algorithms and design‐domain‐independent mesh generation method are developed to improve the efficiency of finite element analysis and optimization iteration. The algorithm realizes the design of ideal energy‐absorbing structures, which are verified by additive manufacturing and experimental characterization. The error between the stress‐strain curve of the designed structure and the target curve is less than 5%, and the densification strain reaches 0.6. Furthermore, special attention is paid to passive pedestrian protection and occupant protection, and a reasonable solution is given through the design of a multiplatform energy‐absorbing structure. The proposed topology optimization framework provides a new solution path for the elastic‐plastic large deformation problem that is unable to be resolved by using classical gradient algorithms or genetic algorithms, and simplifies the design process of energy‐absorbing mechanical metamaterials.

The macroscopic mechanical performance of additive manufactured structures is essential for the design and application of multiscale microlattice structure. Performance is affected by microstructure and geometrical imperfection, which are strongly influenced by the size of the struts in selective laser melting (SLM) lattice structures. In this paper, the effect of size on microstructure, geometrical imperfection, and mechanical properties was systemically studied by conducting experimental tests. A series of AlSi10Mg rod-shaped samples with various diameters were fabricated using SLM. The uniaxial tensile test results show that with the decrease in build diameter, strength and Young’s modulus of strut decreased by 30% more than the stable state. The main reasons for this degradation were investigated through microscopic observation and micro X-ray computed tomography (μ-CT). In contrast with large-sized strut, the inherent porosity (1.87%) and section geometrical deviation (3%) of ponysize strut is greater because of the effect of thermal transform and hydrogen evolution, and the grain size is 0.5 μm. The discrepancy in microstructure, geometrical imperfection, and mechanical properties induced by size effect should be considered for the design and evaluation of SLM-fabricated complex structures.