Optimization of a driver-side airbag using kriging based approximation model

Abstract

The performance of an occupant protection system in the proto-design stage of a new car is often evaluated by CAE (Computer Aided Engineering) instead of the real test. CAE predicts and recommends the appropriate design values; hence reducing a number of the real tests. In this research, the optimization procedure of a protection system, such as airbag and load limiter, is suggested for frontal collisions. The DACE modeling, known as one of the kriging interpolations, is introduced to obtain the surrogate approximation model of the system, followed by the tabu search method to determine the global optimum. A mathematical problem is solved to check the usefulness of the suggested method. To overcome the limitation of existing CAE method having uncertainties of parameters, a distribution of combined injury probability is investigated using the Monte-Carlo simulation on the optimum design obtained from the suggested method.