Optimization Of A Car Frontal Crashworthiness Based On Multiple Collision Modes

Vehicle safety performance is one of the most important standards to value a car.Among all of traffic accidents, frontal impact happens most frequently, with heaviercasualties suffered than other conditions. For this reason, it is highly important to study onfrontal impact. The Automobile Safety Regulations and the New Car Assessment Program insome country has stipulated the front collision test modes, which include100%rigid barrierimpact,30°angular impact and40%offset deformable barrier impact. Each of the three kindsof test modes is different with each other in contents, methods and study priorities, so testson these three collision modes is necessary to evaluate the vehicle frontal impact safetyperformance,.Crashworthiness optimization is a complex issue. Using the traditional way with FiniteElement method to optimize crashworthiness relies on repeated trials and error process and ittakes a lot of time for computer simulation. Furthermore, the FE method couldn’t ensure tofind the optimal solution. Another way using approximate model instead of FE modelbecomes more economic and practicable, and the optimization issue can be done by usingmulti objective optimization algorithm. In addition, the influence of destabilizingfactors in the manufacturing process on the target performance need to be considered toraise the level of quality and engineering robustness of the targets.In this paper, virtual test models in three collision modes (100%rigid barrier impact,30°angular impact and40%offset deformable barrier impact) are established with FEmethod, and the thickness of nine parts are selected as variables, which influence the vehiclecrash safety a lot. Also six responses are selected as targets of the optimization, they areB-pillar peak acceleration in frontal full frontal impact (a1), B-pillar peak acceleration duringthe first20ms and intrusion of the dash panel in30°angular impact (a2and C2), B-pillar peakacceleration and intrusion of the dash panel in40%offset impact (a3and C3) and the totalmass of the nine parts. A3levels and9factors orthogonal experiment is arranged by using design of experiment method, and the responses of each sets of experiment are computed byusing nonlinear finite element software–LS-DYNA. Instead of FE models, highprecision approximate models are built based on these simulation results by using quadraticpolynomial. Furthermore, the engineering robustness is combined with6sigma conceptwhich is usually used in quality engineering, and a multi objective optimization issue aboutcrashworthiness in multi front collision modes based on6sigma robustness has been built.Then the issue is solved by using multi objective genetic algorithm and the most optimalsolution is selected. Finally, the crash safety in the three collision modes is compared byusing FE simulation, and the results show that the effects have been significantly improved.The difficulty of this paper is regarding the crash safety in three different conditions asthe optimization target, because these targets are related to each other. The innovation of thispaper is solving the crash optimization issue with approximate model and multi objectiveoptimization algorithm, and taking robustness into consideration. The new method forcomplex crash optimization issue will improve the development efficiency, and it has astrong value in engineering; it also has some reference value to the optimization of crashsafety performance of other vehicles.