Krystoffer Mroz

Validation of a Human Body Model for Frontal Crash and its Use for Chest Injury Prediction

Whole-body kinematics of the finite element human body model THUMS was evaluated by means of sled tests. A model of a crash test dummy (Hybrid-III 50%-ile) was used to validate the test environment by matching the model predictions to the experimentally measured dummy sled test responses. Once the environment was validated, the THUMS model was placed in the sled model and the post mortem human subject (PMHS) sled tests were replicated. Two test configurations were used for the evaluation. One configuration was high impact velocity sled tests with an advanced restraint system. The other configuration was low impact velocity sled tests with a basic restraint system. The test velocities were 48 km/h and 29 km/h respectively. The evaluation was carried out by an objective rating method that compared predictions from the model to results from the mechanical tests. The method assessed the peak level, peak timing and curve shapes of the predictions relative to the test results. The mathematical crash dummy model predictions matched the mechanical counterpart responses well, confirming that the test environment was replicated. The match between the THUMS model predictions and PMHS responses was not as good.

Evaluation of finite element human body models in lateral padded pendulum impacts to the shoulder

Lateral impacts are of great concern for occupant safety. In order to design side protective systems, it is of importance that the timing of the body and the head should be well predicted. Today, experimental and numerical Anthropometric Test Devices (ATDs) are used as human substitutes to predict the human kinematics. As a complement to the ATDs, numerical Human Body Models (HBMs) are used as research tools. The objective of this study is to compare the loading and kinematics of the shoulder complex in three different HBMs with published biological experiments. This study also compares the models with each other and with two numerical ATDs. The results indicate that no HBM can be used for detailed prediction of the kinematics of the human shoulder complex. However, in the presented statistical analysis, all HBMs show a better overall correlation to experiments compared to the numerical ATDs.