بهینه‌سازی ستون میانی خودرو (B-pillar) از جنس کامپوزیت جهت دستیابی به چیدمان مناسب لایه‌ها در تست برخورد از کنار به کمک الگوریتم ژنتیک

In present paper, the b-pillar of the national car, samand, is modeled with laminated carbon-epoxy composite to improve the side crash behavior. then, a side crash simulation is applied with the aid of finite element simulation according to fmvss 214. as the change in stacking sequence of the lamina may lead into strength improvement, the results are optimized for the minimum of intrusion using neural networks and genetic algorithm. in order to train neural networks for identification of b-pillar crash behavior, experimental design techniques such as taguchi and hammersley has been utilized. after verifying modeling results and using them to identify structure behavior through neural networks and conducting optimization, results indicated that proper stacking sequences are occurred in a band of [40°,50°] layer orientations with the minimum of intrusion. therefore, the intrusion of passenger cabin, which is of high important parameters in side crash, is properly enhanced