The gravity of automotive side impact problem has been recognized by the biomechanics community for years. Several proposals have been made for improving occupant safety. However, most of the improvements were assumed by doing tests with dummies, which are basically not considered as being fully biofidelic. Therefore, the question of how effective they are in real world crashes is raised again. Although cadaver tests will yield more human-like responses, the drawback is not only the difficulty of doing the tests but also the inability of cadavers to sustain injuries frequently seen in the real-world accidents. The purpose of this research is to develop a finite element human model with detailed descriptions of the internal organs. We hope to achieve the goal of being able to predict the risk of the real-world human injuries in side impact crashes. Currently, the model for a human thorax was developed and validated against data from lateral pendulum impact tests which were conducted at Wayne State University several years ago. Good correlation with in-vivo injuries were also obtained.
In this model, a human bony skeleton with major muscle layers and internal organs and tissues in the thoracic cavity was constructed for a 50th percentile male. Material properties for the model are mostly elastic except for the lung and heart in which nonlinear compressive stress-strain relationships were used. Gross responses of thoracic impact force versus time and force versus deflections were obtained. They were found to have a good correlation with pendulum test data. Further analysis was done to determine the effect of pendulum impact speed on the thorax. Higher compression of the heart, surface velocity and compression of the lung, stretch of the ligamentum arteriosum, and shear stresses at the isthmus of the aortic arch were found. A qualitative analysis of injury revealed good correlation with animal test results and with hypotheses of injury mechanism proposed by other researchers. The correlation also serves as additional evidence of the validity of the model.