The dramatic reduction in death and serious injury that has been achieved within the existing vehicle crash testing framework has, in some cases, reached a plateau. Consumer and regulatory organisations are studying those crash modes that are predominantly responsible for the remaining types of injury, and new test requirements are under development, with the aim of driving further enhancement in occupant protection. One such proposal involves a moving deformable barrier impacting the front of a stationary vehicle at an angle of 15 ° and with an overlap of 35%, and this oblique impact introduces a lateral component of motion into the vehicle response. In parallel with the development of additional impact modes, a new frontal crash dummy, the Test Device for Human Occupant Restraint (THOR), is being investigated as a replacement for the traditional Hybrid3. The THOR ATD (Anatomical Test Device) provides a more bio-fidelic response, in particular to the lateral component of vehicle motion created by an oblique impact, and it is being considered for introduction into this and other crash test scenarios. THOR is based on a completely new internal structure, and incorporates several new sensors to increase the amount of information available for assessing the severity of the crash from an occupant point of view. Optimal occupant protection relies on a careful matching of restraint specification to the response of the ATD, and a detailed understanding of the internal structure and instrumentation features of THOR is essential in achieving this.

State-of-the-art vehicle development makes extensive use of virtual test methods, and relies on having sufficient confidence in the virtual toolset. Highly fidelic ATD computer models have been used for many years, and a model of THOR is undergoing rapid refinement to meet its increasing relevance within new test regimes.

This paper describes a test and simulation project designed to study the internal structure and behaviour of THOR, and to investigate the quality of the DYNA simulation model. In order to remove as many sources of variability as possible, testing was carried out using the ATD on its own, as well as with a disassembled thorax. The use of CAE models to derive load cases representative of THOR behaviour in a vehicle crash led to a set of test data relevant to in-service loadings. A comparison of simulation and test results allowed a detailed assessment of the quality of the model, and formed a basis for its future development.

It was concluded that the behaviour of THOR is complex and significantly different to that of the Hybrid3 that it replaces, which reinforces the need for detailed understanding of the internal structure and its interaction with vehicle systems. It was also concluded that the existing DYNA model is adequate for guiding development of vehicle systems in respect of protecting THOR, but that some further refinement in specific areas is required.

THOR will be increasingly relevant as it is introduced into the oblique and other crash test cases, and the understanding of its characteristics is developing rapidly. It is hoped that this study will contribute to this knowledge base.