Abstract
The presented work aims at including partner protection more comprehensively and concentrates on developing the fundamentals of a methodology to optimise vehicles structurally to improve general passive safety on the level of the whole society. Compatibility is regarded as a problem not only of one certain vehicle but also of the fleet it is encountering during operation. All vehicles in the fleet have a certain probability to be the crash partner of a specified car. Therefore, compatibility of a vehicle can be measured quantitatively through the weighted sum over all injury severity scores of all involved occupants, statistically taking into account all possible impacts with other vehicles in the fleet. Compatibility of a vehicle against a certain fleet is termed its fleet compatibility.
Taking fleet compatibility as objective function, a vehicle can be structurally optimized to achieve minimal overall injury severity in a real life accident environment. To actually calculate fleet compatibility, a numerical fleet model and an accident environment model have to be built. Fleet compatibility can be calculated by numerically crashing the vehicle in question (subject vehicle) against the surrogates of the fleet.
The fleet model was built based on a statistical analysis of the actual German car fleet. It was broken down into several vehicle categories. The actual fleet was approximated by a linear combination of five different vehicles; each of them was chosen to represent a well defined part of the fleet.
Due to limitations of the available computational capacity, highly refined FE vehicle models of the surrogates can not efficiently be used. Through further development of the beam element method, which was originally developed by Relou, the simplified models maintain all required features relevant for frontal impact and allow a radical reduction in the computational effort.
Among these simplified models of the surrogates, a model of the B-segment was chosen for the role of the subject vehicle. It was parameterized in its frontal structure to allow optimization. All the crash partners are termed the object vehicles.
The accident environment model has to cover the nearly infinite variability of real crash situations with a limited number of computational crash configurations. Simultaneously it should provide the probability of each crash figuration, which serves as the weighting factor in the overall injury severity. Based on the GIDAS (German In-depth Accident Study) database, the accident environment in Germany can be described comprehensively and the model can be built properly.
A structural optimization with the objective function fleet compatibility of the subject vehicle is run. Crash simulations between the subject vehicle and all its potential crash partners were undertaken, the results evaluated, parameters adjusted, and run again until a termination criterion was met. Simultaneously, through stochastic analysis, the robustness of the design settings was evaluated. Eventually, a trade off was undertaken in order to ensure that the optimized design has not only a good system performance but also a high robustness against uncontrollable deviations of design variables.
As a result, the structural optimization of the subject vehicle can achieve an almost 12 % reduction in the overall crash severity. The achievement of the nearly 12 % reduction in the overall crash severity means that the present subject vehicle is a little more aggressive to its fleet than needed, especially for its more vulnerable crash partners.
Keywords: Compatibility, Robust Design, Structural Optimization, Crash Simulation