Introduction Of Beadings Into A Crash Tube Using A Robust Optimisation Approach

Deterministic optimisations result in optimum designs for the specified load case but in unstable designs when load cases change. Limitations or uncertainties have to be taken into account leading in design optimisations which will then be insensitive to small fluctuations of geometric or input parameters. Such a stochastic analysis is often referred to as robust design optimisation. This paper presents a Robust Design Optimisation for the illustrative example of a cylindrical metal crash tube. The goal of the optimisation is to attain a crash structure which is insensitive to small fluctuations in the magnitude of the crash load, the direction of the velocity, as well as the mass of the impactor. The tilting of the impactor is modelled with a stochastic approach. The geometry of the crash tube shall be insensitive to uncertainties by optimising the wall thickness and by introducing beadings into the tube. Robust design optimisation was conducted with LS-OPT, combining ALTAIR HYPERMORPH for small geometrical adaptions and LS-DYNA for finite element method simulations. To reduce the computational effort, the optimisation is based on metamodels as this allows for mathematical optimisation. Robust design optimisation allows for results less sensitive to small system fluctuations and hence the geometric tube tolerances can be reduced or the field of application (e.g. vehicle class) can be increased. For the presented crash tube this advantage results in more stable crash results for scenarios departing from the standard. This in turn leads to increased safety levels of life protection systems. The use of robust design optimisations on the presented crash tube leads to geometries being 6% heavier than the ideal optimum results, but being more robust towards geometric or load fluctuations.