Multi-objective optimization study of driver's airbag parameters

Build a driver side restraint system model of an SUV according to MPDB (50% overlapping progressive deformation barrier impact test on the front) working conditions. Evaluation of the system using the Thor 50M dummy revealed poor performance in terms of chest compression and neck tension Fz, both of which exceeded the requirements set by C-NCAP regulations. To address this issue, we utilized a high-dimensional multi-objective optimization technique to optimize the airbag design parameters. To optimize the dummy injury values, we conducted an analysis of each airbag parameter using the univariate test method. We studied the effect of each parameter on the injury values of the dummy's head, neck, and chest, and selected parameters that exhibited high sensitivity. Specifically, we selected the dummy chest compression, head HIC15, acumulative 3ms acceleration value of the head, and neck tension Fz as our optimization objectives. To perform the sampling simulation, we used uniform Latin hypercube sampling for the parameters. We constructed an agent model for the four optimization objectives using the response surface modeling method, followed by a high-dimensional multiobjective optimization. We substituted the local optimal solutions of the optimized airbag parameters into the simulation model to verify their effectiveness in improving the dummy injury values. The simulation results demonstrate that the optimized airbag parameters reduced the dummy's chest compression by 13.53%, head HIC15 by 31.28%, neck tension Fz by 30.41%, and acumulative 3ms acceleration value of the head by 19.76%. These improvements enhance the occupant protection provided by the restraint system. Our study shows that optimizing the airbag design parameters can provide new ideas and techniques to enhance vehicle safety.