Virtually destroyed

LinkedIn +

To understand and quantify crash performance of a subject vehicle, a number of standardized test procedures have been developed. Although such test procedures can approximate the crash performance over a range of impact configurations, they have limitations when it comes to replicating the diversity and complexity of crash events in the real world. Options for expanding the understanding of crash performance over a wider range of impact configurations can come from the adoption of CAE. Certainly in the area of pedestrian protection, the changes to the EuroNCAP test protocol have extended the role of CAE in the assessment of vehicle performance. But is CAE a panacea for replacement of the physical test?

First we must ask the question, what are we looking to replace? The physical crash test is not ideal for assessing crash performance. At best it is an approximation of what may happen in the real world, translating a structural loading defined in terms of impact velocity, barrier type, overlap, etc. into occupant injury. For example, to determine injury risk from a crash test, the measurements from the ATD must be translated to a risk of injury. For decades, work has been performed on human injury from blunt trauma in the automobile field. Simulated automobile crashes and/or impact tests are performed (replications of impacts in which injury outcome is known or can be approximated), and the response of the biofidelic surrogate (cadavers, porcine subjects, etc.) is taken to represent the response of a human in that crash scenario. This response may be used in the development of numerical relationships between measurable engineering parameters and risk of injury for that crash scenario.

The crash test is therefore a carefully defined system and the CAE is a representation of that system. The issue is that the crash test is not one system, but a construct of multiple complex systems strung together in order to translate one input (the crash event) into a useful output (the likelihood of injury). A problem then arises if part of the system is changed – this is the argument for increased CAE in that it enables multiple different crash events to be explored. For example, the loading applied to the vehicle structure is dependent on the crush strength of the deformable barrier face. Tests have shown that the strain rate of the aluminum honeycomb used in the barrier face is rate dependent. Calibration has enabled CAE models to approximate the behavior observed in reality for a set of defined inputs – typically done by taking an approximation of the material crush strength that is higher than the quasi static crush strength, but lower than the peak crush strength experienced at the start of the impact event. This presents two issues: one is that the model is an approximation of a physical reality – hence useful for exploration of the behavior and furthering understanding of different failure mechanisms of the structure and/or injury causing mechanisms; the second is that changes in the inputs (for example, a change in impact velocity or mass) would require a recalibration of the model, necessitating physical testing.

The limitation of CAE is therefore one of calibration. CAE is a useful tool for furthering our understanding and identifying trends – and hence can even support development – but for regulatory purposes and quantifying performance across different input conditions, the need to calibrate the model would have to be accounted for. The approach therefore has to be to complement the CAE with physical tests – benchmarking the CAE results against physical tests to validate their usefulness in assessing performance as noted above with the reference to EuroNCAP.

September 13, 2016

Share this story:

Comments are closed.