The objective of the EC funded HUMOS2 project was to develop a set of Finite Element models of the human body capable to represent a large range of the population of road users involved in any type of road accident and allowing an accurate prediction of the injury risk. Fifteen partners (INRETS, Chalmers U., TU Eindhoven, ESI, Faurecia, U. Heidelberg, SERAM, Mecalog, U. Mediterranee, LM U. Munich, PSA, REGIENOV, TNO, Volvo, VW) from 4 member states were involved in this project which ended in February 2006. Starting from the outcome of a first phase completed in 2001, methods allowing the personalisation (anthropometry, geometry and position) of a generic human model have been developed. They include a scaling tool enabling to derive any individual model from the original mesh by means of control points and statistical relationships between external and internal dimensions. These were established from geometric data collected on standing and sitting volunteers with a low dose bi-plane X ray system and from geometric data directly measured on bone parts. A positioning tool has also been developed in order to adjust the model for different sitting postures and standing postures as well. The application of these tools required the redefinition of the mesh in several areas. Second, experimental work was conducted on human volunteers in order to identify the influence of muscular tensing on the response of the body to moderate impacts and muscular tissue mechanical properties. These were also analyzed from animal and cadaver testing. Third, a set of reference postures for testing the models, including occupant OOPs and pedestrian postures has been defined. A data base of laboratory biomechanical test results, appropriate for model validation, has been set up. It mainly includes existing cadaver tests results coming from former EC funded projects and and from Heidelberg university and was completed by specific cadaver and volunteer tests performed during the present project. Bone and ligament material properties and failure limits were implemented in as well as injury criteria for most common injuries sustained by road victims. The models, implemented on 3 different FE crash codes, were used to simulated a set of chosen cases in order to evaluate their performance in terms of biofidelity of behaviour and injury prediction capability. Results show that good overall kinematics of the human body can be reproduced in various accidental situations. Body part loading can be calculated and resulting injury risk evaluated. Some improvements are still necessary to improve the robustness of the model in certain cases and extensive testing against real accident cases should be performed in order to fully validate the models. Part of this work is being conducted in the APROSYS EU funded integraed project (A). Only abstract (as above) is available from the conference proceedings. For the covering abstract of the conference see ITRD E212343.
Samenvatting