Methodology for parameter identification on a thermoplastic composite crash absorber by the Sequential Response Surface Method and Efficient Global Optimization. (15th December 2021)

• Record Type: Journal Article
• Authors: Raponi, Elena; Fiumarella, Dario; Boria, Simonetta; Scattina, Alessandro; Belingardi, Giovanni
• Is Part Of: Composite structures. Volume 278(2021)
• Publication Date: 2021-12-15
• Subjects: Crashworthiness -- Composite impact attenuator -- Finite element analysis -- Efficient Global Optimization -- Sequential Response Surface Method -- Surrogate models; Composite construction -- Periodicals; Composites -- Périodiques; 624.18
• DOI: 10.1016/j.compstruct.2021.114646
• Languages: English
• ISSNs: 0263-8223

Abstract

Abstract: Numerical simulations for crashworthiness require the definition of material properties that are not always predictable with standard experimental tests. This paper deals with the numerical optimization of a thermoplastic composite material model. The component is a vehicle impact attenuator made of an innovative All-PP (PolyPropylene) composite material. The peculiar failure mechanism of this material makes the numerical simulation of the collapse a difficult challenge to achieve with a trial and error calibration of the material card. Therefore, an optimization procedure is proposed to determine the material parameters. The optimization is implemented in LS-OPT, where the mean square error between the experimental and numerical load–displacement curves is the objective function to be minimized. Two test cases are considered: (1) optimization of the material card based on the full load–displacement curve from the experimental tests and (2) optimization of relevant parameters of a numerical trigger added to control initial contact instabilities between the impacting rigid wall and the component in the numerical simulations. The optimization strategies Sequential Response Surface Method (SRSM) and Efficient Global Optimization (EGO) are used. The results show that the presented methodology allows characterizing the studied composite material and thus obtaining a more efficient numerical model.