Finite Element Analysis And Optimization Of Beaded, Tapered Crash Box

Crash box equipped at the front of cars, is one of the most important automotive parts forcrash energy absorption. In case of low speed frontal collision, it is expected to be collapsedin a favourable deformation mode, absorbing the crash energy prior to other automotive bodyparts so that the injury of the occupants and damage to the car is minimized. The peak impactforce is the factor that affects the occupants’ safety and also the damage to the car itself. Thisstudy focuses on optimization of a tapered crash box with crash beads as crash fold initiators,to obtain a crash box model with lower peak crash force and keep the crash interface forcestable as well. Tapered structures are considered preferable because of their lower cost andthey are more likely to provide a desirable constant mean load under dynamic loading.Dynamic crushing behaviors of the crash boxes of different cross sections (circular,hexagonal and square) were first studied. A number of finite element models were modelledwith varying taper angles and were used for crashworthiness analyses. Peak crushing force,mean crushing force, crush length and total energy absorbed were recorded from the finiteelement crash simulation output. Parameters like specific energy absorption (SEA), the ratobetween the average and maximum crushing force (Favg/Fpeak) also referred to as “Crush forceefficiency”(CFE) and energy absorbed per unit crush length were determined and compared.Comparing these different crashworthiness parameters, a design selected as a baseline designwas put to further improvement with application of crashbeads as crash fold initiators. Thebeaded crashbox was then used as base design for further optimization.Five design variables (thickness, taper angle, bead length, width and depth) that describethe tapered crash box and beads geometry are optimized to maximize CFE to obtain morestable force deflection curve. To represent complex crashworthiness objective function, asurrogate model method, more specifically, response surface method (RSM) was adopted. Thedesign of experiments (DOE) using Latin Hypercube sampling technique is employed toconstruct the response surface model. SEA was constrained to be higher than that of basedesign and Peak force was constrained to be less than that given by the vehicle crash safetyregulation.Analyses involved in this paper are undertaken using finite element models and, anexplicit finite element solver LS-DYNA and HyperStudy are used to perform all the analysesand the optimisation.