Human Body Model development for safety

SNIC 2019/3-222


SNIC Medium Compute

Principal Investigator:

Johan Iraeus


Chalmers tekniska högskola

Start Date:


End Date:


Primary Classification:

20301: Applied Mechanics

Secondary Classification:

30599: Other Medical and Health Sciences not elsewhere specified

Tertiary Classification:

10106: Probability Theory and Statistics



The work of the Injury Prevention research group aims at preventing injuries due to extreme mechanical loading to the human body. This can for instance be loading caused in vehicle crashes. The research is prioritized based on accident statistics provided by the Field Data Management group within the division. The research includes injuries and injury mechanism at the microscopic level up to the response of the whole human body during externally applied mechanical loads. The ultimate aim of the research is to develop principles for new or improved protective systems. This can for instance be airbags, whiplash protection systems or child restraints, but also other protective devices such as helmets or motorcycle clothes. In order to reach the aims we need to create and validate tools such as Human Body Models (computer models of the human body). In general, two sources of validation data is used; (1) tests on biological materials and (2) reconstructions of real life crashes documented in databases. Replicating tests on biological data (material up to full body) is normally not very computer intensive. However, the reconstruction of real life crashes poses some problems, as many parameters of the physical crash are normally unknown. To overcome this problem parameter studies or stochastic simulations can be used. However, as both the human body models as well as models of the vehicles are quite detailed, there is a need for quite large computational resources. A third type of simulations that is currently under development is the simulation of pre-crash kinematics. The nature of this type of crashes are quite long durations. To able a smooth transition form pre.vrash to in-crash the models are very similar. The price is that the pre-crash simulations are very computer intensive, a single job can consume as much as 1000+ core hours.