The glass sponge is a porous lightweight structure in the deep sea. It has high toughness, high strength, and high stability. In this work, a super-depth-of-field microscope was employed to observe the microstructure of the glass sponge. Based on its morphological characteristics, two novel bio-inspired lightweight structures were proposed, and the finite-element analyses (FEA) of the structures were carried out under compression, torsion, and bending loads, respectively. The structure samples were fabricated using stereolithography 3D-printing technology, and the dimension sizes of the samples were equal to those of the corresponding FEA models. Mechanical tests were performed on an electronic universal testing machine, and the results were used to demonstrate the reliability of the FEA. Additionally, lightweight numbers (LWN) were proposed to evaluate the lightweight efficiency, and a honeycomb structure was selected as the reference structure. The results indicate that the lightweight numbers of the novel bio-inspired structures are higher than those of the honeycomb structure, respectively. Finally, the proposed structures were optimized by the response surface, BP (Back Propagation) and GA-BP (Genetic Algorithm optimized Back Propagation) method. The results show that GA-BP model after training has a high accuracy. These results can provide significant guidance for the design of tube-shaped, thin-walled structures in the engineering.

中文翻译：

---

受玻璃海绵（Porifera，Hexacinellida）启发的轻质结构的优化设计及其机械性能。

玻璃海绵是深海中的多孔轻质结构。它具有高韧性，高强度和高稳定性。在这项工作中，采用了超景深显微镜来观察玻璃海绵的微观结构。根据其形态特征，提出了两种新颖的受生物启发的轻质结构，并分别在压缩，扭转和弯曲载荷下进行了结构的有限元分析（FEA）。使用立体光刻3D打印技术制作结构样本，样本的尺寸大小等于相应的FEA模型的尺寸大小。在电子通用测试仪上进行了机械测试，结果被用来证明FEA的可靠性。另外，提出了轻量化数（LWN）来评估轻量化效率，并选择蜂窝结构作为参考结构。结果表明，新型生物启发结构的轻量化数分别高于蜂窝结构。最后，通过响应面，BP（反向传播）和GA-BP（遗传算法优化反向传播）方法对拟议的结构进行了优化。结果表明，训练后的GA-BP模型具有较高的准确性。这些结果可为工程中的管状薄壁结构的设计提供重要指导。结果表明，新型生物启发结构的轻量化数分别高于蜂窝结构。最后，通过响应面，BP（反向传播）和GA-BP（遗传算法优化反向传播）方法对拟议的结构进行了优化。结果表明，训练后的GA-BP模型具有较高的准确性。这些结果可为工程中的管状薄壁结构的设计提供重要指导。结果表明，新型生物启发结构的轻量化数分别高于蜂窝结构。最后，通过响应面，BP（反向传播）和GA-BP（遗传算法优化反向传播）方法对拟议的结构进行了优化。结果表明，训练后的GA-BP模型具有较高的准确性。这些结果可为工程中的管状薄壁结构的设计提供重要指导。