Review of correlation methods for evaluating finite element simulations of impact injury risk

Finite element models have been used to understand human injury responses in various crash configurations. Most of the model validations were limited to qualitative descriptions. Quantitative analysis was needed for the validation of finite element models against experimental results. The purpose of this study is to compare the existing correlation techniques and to determine the best method to use for evaluating finite element simulations of impact injury risk in vehicle crashes. Five correlation methods in the literature were reviewed for systematic comparisons between simulations and tests. A full frontal impact test of a 1997 Geo Metro was simulated. The finite element model of a 1997 Geo Metro was obtained from NCAC finite element model archive. The acceleration and velocity responses of the vehicle seat were extracted from the simulation and compared to the test data. Evaluations of the validation methods were based on the analysis results compared to the suggested criteria. Performance of the different methods showed that the Comprehensive Error Factor method was the best overall correlation method, and therefore was recommended for assessing occupant injury potentials in vehicle accidents.