P 979 - Deformation and failure behaviour of steels for automotive engineering under crash-like multi-axial loading
The loads occurring in crash scenarios are locally very different and complex. In principle, the mechanically loaded structures are subject to very different stress states and load velocities. These different load …
Content
- Publication
- 2016
- Authors
- Prof Gumbsch
The loads occurring in crash scenarios are locally very different and complex. In principle, the mechanically loaded structures are subject to very different stress states and load velocities. These different load patterns have different effects on the deformation and failure behaviour of the materials used. As the deformation behaviour influences the failure, it is necessary to know both of the aforementioned properties precisely in order to realistically simulate complex structures. systematic investigations. Therefore, one of the main objectives of this project is to experimentally characterise certain application-relevant steels in detail and to propose or provide corresponding suggestions for the efficient modelling of deformation and failure behaviour depending on the main influencing variables of multiaxiality and strain rate mentioned at the beginning.
Show full description
In addition, thermal effects occur in real crash situations due to both high strain rates and the different multiaxial loads, which in turn can be explained by the deformation energy partially converted into heat and by transient heat conduction processes. As a result, adiabatic temperature increases occur locally in the constriction area, which influence the deformation and failure behaviour and must also be taken into account. The aim of the project was therefore also to systematically investigate the deformation and failure behaviour of steels for automotive construction under crash-like multi-axial loads by means of experimental and numerical investigations. For the three lightweight steels HX340LAD, HCT980X+Z110MB and 22MnB5 AS with different strengths, both static and dynamic tests were carried out on flat tensile, 0° inclined shear tensile, notched tensile, punched tensile and Nakajima specimens. The deformations occurring in the test piece under dynamic load were recorded The deformations occurring in the test piece under dynamic load were recorded using state-of-the-art high-speed video cameras, whereby 3D video recordings were made for the nakajima tests. The optical method of grey value correlation analysis was used to evaluate local strains up to the onset of failure. High-speed infrared measurements were used to document the adiabatic temperature increases occurring in the highly deformed zones during crash loads. The exact failure location and the failure mode (shear/tensile fracture) were determined by video analysis and fractography. The maximum strains measured at the failure location immediately before the fracture were plotted as failure curves in failure diagrams as a function of the stress multiaxiality calculated with FE for both static and crash-type loads. Overall, the results for the analysed materials show decreasing failure curves from the shear tensile to the notch tensile range. In the biaxial tensile range, the values for the failure strain are then significantly higher again. This is due to different stress states in biaxial tensile loading (plane stress state) compared to notch tensile loading (triaxial stress state), which are also reflected in significantly different lode parameters. With increasing strain rate, moderately higher failure strains occur for the flat tensile tests for HX340LAD and HCT980X+Z110MB; for 22MnB5 AS there were initial indications of significantly lower failure strains. In the shear range, the failure strain is generally lower with increasing strain rate. This is probably due to the very large strains occurring under shear in a narrow, highly deformed shear zone. The high local deformation work implemented there leads to large adiabatic heating, which was measured at over 200 K at crash-like test speeds for the microalloyed steel HX340LAD and apparently promotes slippage. High-speed tensile tests are therefore not sufficient to calibrate a failure curve under dynamic loading, because the shape of the failure curve changes depending on the strain rate and the lower shear strain values under crash-like compared to static loading must be taken into particular account in crash simulations. The strength class of the steel also plays a decisive role in the influence of the strain rate on the failure curve. In contrast to the other two materials, the highest-strength steel 22MnB5 already shows lower failure strains in the tensile range with increasing strain rate. Since adiabatic effects play an important role in the precise description of the deformation behaviour under dynamic crash-like loads, thermomechanically coupled material models were tested and their behaviour investigated. Due to the specifications of the industrial companies involved, fully coupled thermomechanical modelling had to be dispensed with, as an efficient description of the deformation behaviour without the time-consuming implementation of the heat conduction equation is important. Therefore, a pseudo-thermomechanical model without heat conduction with a strain rate-dependent Taylor-Quinney coefficient was proposed and implemented and extensively tested with a partner (SME) involved in the project. The deformation behaviour could be well adapted to the global force-displacement curves of the different materials with the pseudo-thermomechanical model. The local strain distributions also show good agreement with the local strain measurements from ARAMIS in most cases. Only the bending- and biaxial-dominant simulations of the Nakajima tests show partially larger deviations from the experimentally determined local strain measurements, whereas the global force-displacement curves of these calculations show very small deviations at the same time. Overall, the deformation behaviour could be simulated satisfactorily for all tested materials at all tested strain rates and multiaxialities. The failure behaviour was described using a strain-based model, which is based on damage evolution, which in turn depends on the current stress state. The damage evolution was not coupled with the evolution of the current stress state in order to allow separate non-interacting adjustments of the deformation and failure behaviour. Despite this simplifying assumption, the global failure behaviour could be well adapted to the majority of the tests. A failure curve was determined for each investigated strain rate within a material, which is necessary for the calibration of the damage model. The deviations of the simulated local failure strains from the experiments are acceptable in most cases, although for biaxial-dominant loads the difference to the ARAMIS measurements is somewhat greater compared to the other triaxiality ranges. Overall, the experimentally recorded material behaviour of the investigated materials could be described well with the proposed and used models, although the ductility of the lightweight steels is very different. The thermal behaviour can also be described well with the simplified pseudo-thermomechanical model for the majority of the load conditions investigated, although shear-dominant multiaxialities can be partially problematic due to the strong localisation.
Zum Zoomen hovern
€30.00excl. VAT
Payment by card (Stripe) or invoice (DE, AT, CH). Download link by email: automatically after ordering by card, after order approval for invoice (valid for 30 days). Terms
P 979 - Deformation and failure behaviour of steels for automotive engineering under crash-like multi-axial loading—
€30.00
