Simulation of the T6 bridge rail system using LS-DYNA3D

Abstract

Full-scale crash testing currently the primary means of evaluating trustworthiness of roadside safety structures, such as bridge rails. However, explicit finite element analysis is rapidly becoming a feasible alternative and offers several advantages to the designer. Over the past ten years, FHWA has supported research involving use of the explicit code LS-DYNA3D for analytical simulation of vehicular impacts of roadside safety systems. Consequently, the goals of this research were the development of a finite element model of the TxDOT T6 bridge rail system using LS-DYNA3D and validation of simulation results against recorded data from full-scale crash testing. The process of modeling the T6 system was a conglomeration of efforts to accurately represent the various components and approximations to reduce the required computational time. When feasible, material testing was conducted to obtain input values for material definitions. The focus of the research approach was to validate subsystem models, primarily the post-base-weld assembly, against TTI test data prior to assembling the entire T6 model. Sections of the T6 installation located outside the impact region and various connections were represented using approximate modeling techniques. Because no test data existed, explicit models of these components were created solely to calibrate simplified models. Also, most initial simulations utilized rigid impactors to evaluate the response of the finite element models. To evaluate the accuracy of the T6 model, a 2000-kg truck model was obtained from the National Crash Analysis Center (NCAC). Impact conditions of the final simulation were based on TTI full-scale crash test 418048-03. Simulation results were compared both qualitatively and quantitatively with recorded data from a full-scale crash test conducted by TTI. Evaluation criteria, such as the overhead vehicle trajectory, deformation of the vehicle and the T6 system, transverse deflections, and the velocity time history of the vehicle, correlated well. However, the exit angle and amount of rolling experienced by the vehicle did not reflect results of the experimental test. It was concluded that the research objective was accomplished within the limitations of LS-DYNA3D'S fillet weld definition, the inability to model reinforced concrete, and the mass distribution of the NCAC truck model.