On Design Optimization of a Composite Impact Attenuator Under Dynamic Axial Crushing

Motorization brings two significant challenges to the modern society. Firstly, road and vehicle safety becomes increasingly important, which has notably heightened legislative requirements by introducing more effective protective systems to the vehicle. Secondly, there is an ever-growing concern in environment and sustainability, which largely push up the lightweight standards to reduce fuel consumption. For these reasons, the automotive industry has devoted a substantial effort to deliver more crashworthy vehicles for addressing these two competing issues simultaneously. Over the past two decades design optimization has been developed as a powerful tool to seek the highest possible crashworthiness and lightest possible structure for various vehicles, therefore becoming an important topic of research. In crashworthiness optimization, direct coupling method may be inefficient since iterative non-linear FEA during optimization usually require huge computational efforts and take the high risk of premature simulation failure prior to a proper convergence. As a result, surrogate models (or metamodels) are more often used as an alternative for formulating the design criteria in terms of an explicit function of design variables in advance of optimization, which has proven an effective and sometimes a unique approach. The idea of surrogate modeling is to construct an approximate function based on a series of sampling evaluations, in which design space is typically sampled using the design of experiment methods. Then, the FEA is performed at these sample points to establish surrogate models with a certain confidence of approximation for crashworthiness optimization. This paper provides the results obtained from an optimization procedure on a composite impact attenuator, under dynamic axial crushing, using two different metamodels, such as Radial Basis Function and Kriging. In particular the sizing optimization for some geometric parameters was solved combining the commercial solver LS-DYNA with the optimizer LSOPT. In order to measure the fitness of results and do a comparison between different surrogates, global error parameters were used, such as root mean squared error, maximum residual error and coefficient of determination.