Three-Dimensional Shape Optimization Through Design of Experiments and Meta Models in Crash Analysis of Automobiles

Crash analysis is an important step in modern day design of vehicles to satisfy the standards laid down by various agencies. In order to satisfy the standards, the designs should adopt energy absorbing members that are efficient and thus optimization becomes important. Modern design practices now provide procedures for this optimization involving shapes. Because each crash analysis takes considerable computer time, Design of Experiments (DOE) methods are used to generate the response surface on which the optimum is searched using approximate or meta models that consume less computational time. This paper describes methods to achieve shape optimization of the identified members which absorb the energy to a maximum possible value.