Numerical and random forest modelling of the impact response of hierarchical auxetic structures

Major sources of concern when auxetic protective structures are deployed in service of mission-critical applications encompass the triggering of high impact stress and the weakening of the structures' elastic strength in response to the impact events. The current prevailing approach to assessing the impact resistance of these structures broadly hinges on mechanics-informed nonlinear finite element (FE) analysis. However, this method is computationally expensive and ill-suited for tackling the implementation of automated condition monitoring schemes. To address these issues, first, this paper proposes a hybrid hierarchical auxetic structure named HybridHierarchical Re-entrant Honeycomb (HHRH) that is endowed with impact stress-reducing capabilities. Next, using explicit FE, the investigation uncovers the interplay between the key geometric features of this HHRH auxetic structure and the impact performance under low, intermediate, and high crushing velocities. The outcome of the nonlinear explicit FE simulations is then unified with random forests (RF) scheme towards the establishment of intelligent auxetic structural systems. The results revealed that the developed HHRH maintained the auxeticity of the regular re-entrant auxetic and exhibited superior performance in some crushing strain regions. Moreover, the HHRH structure manifests up to an 85 % reduction in peak stress and the proposed reinforcement boosts the auxetic property by up to 23 % when compared to the regular re-entrant auxetic structure under high-velocity applications. Regarding the established data-driven RF-enabled machine learning model, its predictive strength with optimally-tuned hyperparameters is demonstrated to excellently capture the nonlinear multi-modal crushing stress response at various crushing strains, velocities, and geometric variations. Data Availability: The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.