Crashworthiness optimisation of a composite energy-absorbing structure for subway vehicles based on hybrid particle swarm optimisation

Abstract

To improve the crashworthiness of subway vehicles, a composite energy-absorbing structure (EAS) is designed by coupling a thin-walled metal tube and aluminium honeycomb structures. On this basis, the effectiveness of the composite structure and aluminium honeycomb finite element model (FEM) is validated by conducting trolley impact tests and quasi-static compression experiments. Based on the verified FEM, the surrogate models, including polynomial response surface (PRS), Kriging, radial basis function (RBF) and supported vector regression (SVR), are established to get the relationship between crashworthiness indexes and design variables. Then, the most accurate model is employed for crashworthiness optimisation through comparing the accuracies of these four models. Owing to the mutually effects among components of the composite EAS and the relatively complex mathematical formulae acquired by the high-precision surrogate model, a hybrid particle swarm optimisation (HPSO) algorithm is put forward. The performance of the HPSO is tested by a typical engineering optimisation case. The results indicate that the HPSO algorithm presents various advantages (e.g. strong global search capability, high optimisation accuracy, etc.). By optimising the mathematical model of the composite EAS, the optimal configurations of the structure are obtained, which confirms that the HPSO algorithm has favourable applicability and performance in the crashworthiness optimisation of the EASs for subway vehicles.