This paper aims to optimise a bulk scale design of a novel auxetic structure, the hourglass structure (HGS), through a multi-objective optimisation model to improve its protection performance. A 3D numerical model of the HGS under an in-plane quasi-static compression was developed and validated with experimental results. Based on the validated numerical model, a series of numerical analyses were conducted by automating the HGS design process with randomly generated design variables. The automated numerical analyses built a dataset of the selected protective performance indicators (namely, peak elastic stress, plateau stress, and energy absorption capacity). Then, the dataset was used to develop a high-accuracy surrogate model using a radial basis function (RBF) neural network. Afterward, the Pareto optimal solutions were searched with the non-dominated sorting genetic algorithm (NSGA-II). The best compromise design out of the Pareto optimal set was determined with the ideal point method. The performance of the optimum design was simulated under both quasi-static and blast loadings to comprehensively explore the protective performance. In addition, a correlation matrix was constructed to investigate the effects of each design parameter on the protective performance indicators quantitatively. The results showed that the obtained optimum design outperformed the baseline structure under both quasi-static and blast loadings. The optimum HGS design displayed a higher and more stable negative Poisson's ratio along with two deformation modes leading to two plateau stress levels. The optimised HGS design is applicable as the core of high-performance protective sandwich structures.
Discussion(0)
No comments yet. Be the first to comment.