Using multibody-system modeling to make accurate predictions of vehicle impacts on road restraint systems

Abstract Road restraint systems have the aim to prevent vehicles from leaving the road, thus to elude obstacles (lightening posts etc), to protect the environment (water protection areas etc) or to prevent fall hazards at bridges (protecting people and housing underneath bridges). With regard to these different protective aims, the European standard EN 1317 classifies vehicle restraint systems (VRS) into different containment levels for ‘temporary’, ‘normal’ or ‘higher containment’. To classify and certify a VRS to these levels, EN 1317 requires carrying out standardized full-scale impact tests, causing substantial expenses. To reduce such costs, EN 1317-part 5 explicitly allows replacing those full-scale impact tests by computational simulations under certain limitations, in particular if a VRS is only subject to modifications. Due to the high requirements on reproducing the real impact tests by simulations, the modeling of the VRS as well as of the vehicle demands for great accuracy and high skills. Already minor changes on the model of the VRS or vehicle may cause significant changes in results. While it seems at the moment that FEM models are the preferred choice for this, the paper will show that the method of MBS provides an equally or even better simulation approach. The MBS models are assembled using macro-objects from former experience [e.g. Neuenhaus et al. 2007 [12] ] offering simulations with comparably short runtime. In particular favorable for parametric studies, the use of MBS models allows the modification of system parameters directly by changing viscoelastic element parameters. To illustrate the potential and possible accuracy of using MBS, the results from real full-scale impact tests are faced with the results from simulation runs. To show the generality of the method, one of the examples represents a rather flexible (soft) VRS, and the other one represents a much more stiff construction.