Crashworthiness analysis and multi-objective optimization of a commercial vehicle frame: A mixed meta-modeling-based method

In this study, a mixed meta-modeling-based optimization method has been proposed and applied to a commercial vehicle for crashworthiness design subjected to the frontal crash. A full-scale finite element model of the commercial vehicle has been built and validated by a crash test. The front frame parts have been separated to build a sub-model for crashworthiness optimization. Sensitivity analysis has been performed to find the design factors contributing most to crash performance by using design of experiments. With the reduced dimensions of design space, meta-models of crashworthiness criteria (i.e. specific energy absorption, peak crush force, and peak crush acceleration) have been built by using polynomial response surface and radial basis function networks, respectively. The meta-models with higher global fidelity in design space have been adopted to formulate the multi-objective optimization problem of crashworthiness design, which has then been solved by using Non-dominated Sorting Genetic Algorithm-II. The obtained Pareto front has been discussed and validated with that achieved by Strength Pareto Evolutionary Algorithm 2. The normalized optimal solution from the Pareto front has resulted in 11.15% increase in specific energy absorption and 13.2% decrease in peak crush force for the frame and has led to an obvious improvement in occupant protection and energy absorption for the whole vehicle, verifying that the proposed method is effective for vehicle crashworthiness optimization.