Impact protection mechanism and failure prediction of modular hierarchical honeycomb system with self-locking effect

Abstract

Modular system with the self-locking effect has garnered increasing attention in the field of impact protection, owing to its inborn modular controllability and low-cost maintainability. Based on the deconstruction of honeycomb structures, a modular hierarchical honeycomb protection system (MHHS) is developed in this study, offering easy assembly and transportation. The protective performance and deformation behaviors of the modular system are evaluated through drop weight impact tests at 20 J and 60 J, verifying the validity of the finite element simulations. Compared to the integrated honeycomb structure, the modular system reduces peak force by 50% on average while enhancing dynamic specific energy absorption by 54.7% (20 J) and 217% (60 J). The collision durations of the modular system are approximately 2.8 times and 5.5 times longer, indicating less structural stiffness and more structural elasticity. The self-locking effect of the modular system emerges from interactions between the bolts’ bidirectional three-point bending deformation and transverse compressive deformation of components, promoting tighter deformation coupling. Two structural failure criteria are established based on the multi-peak and multi-wave characteristics of response curves, enabling effective dataset preprocessing. The XGBoost model is trained to predict binary classification outcomes for bi-objective analysis based on the modular system performance failure scenarios. The trained model effectively addresses the impact inverse problem, reducing testing costs by 86.1% while maintaining 80% accuracy against simulation benchmarks. These results demonstrate the potential for intelligent assembly applications of the machine learning-guided modular system in practical engineering fields.