Abstract
Due to the superior cold-forming and impact absorbing performance, dual phase (DP) steels are increasingly employed to replace conventional steels in the automotive industries, to increase safety and reduce environmental impact. Extensive researches on DP steels have been executed, showing that martensite/ferrite (M/F) interface damage plays a major role on governing failure of commercial DP steel grades. Nevertheless, the underlying M/F interface microstructure deformation behavior has not been fully understood yet, because of too many complex interactions and effects of unknown subsurface features of plasticity mechanisms in experiments on bulk samples. To study individual plasticity mechanisms at the M/F interface, experiments on isolated features with well-controlled load cases have been performed in literature. From scanning electron microscopy (SEM) digital image correlation (DIC) and a novel slip identification procedure, substructure boundary sliding (SBS) and M/F misorientation appear to have an important effect on plastic deformation near the interface. Unfortunately, it is still difficult to discern certain plasticity mechanisms, since the entire state of the sample is difficult to obtain.Furthermore, a link to damage cannot be easily made, because local stresses are generally not available. Numerical simulations, in contrast, allow for obtaining information of the entire specimen of investigation, and thus, can provide additional insights on the M/F interface deformation behavior.
Therefore, an experimental-numerical integration framework is developed in this thesis. This integration still faces a few challenges, such as recreating an accurate geometry, applying correct boundary conditions (BCs), and making a consistent comparison. In this thesis, these challenges are tackled, and the resulting experimental-numerical integration framework is used to gain more insights on the M/F interface behavior in DP steels.
Uniaxial nano-tensile tests on DP steel samples with an M/F interface in the tensile direction, with varying M/F misorientations and morphological complexity, are modelled. A 3D geometry is constructed by interpolating the microstructure of the front and rear of the samples, derived from SEM and electron backscatter diffraction (EBSD) images. This geometry is discretized using finite elements. Thin layers are inserted in martensite, to which softer material properties are assigned, to capture discrete strain bands within martensite, which are hypothesized to be triggered by SBS or habit plane lath slip. A classical crystal plasticity (CP) model is adopted to describe individual phase behavior, for which material properties are estimated from literature, by interpolating values for the used carbon content and martensite volume fraction. The exact load case applied to the specimen may play an important role for the resulting material behavior. For this reason, BCs, captured from SEM DIC displacement data, are applied at the gauge section ends. From the simulation results, 2D strain and total slip fields are derived.
The integration framework is validated, by showing that a consistent comparison of the experimental and numerical results is possible for one of the more complex samples.
The experimental-numerical integration is in the first place used to validate the numerical model, by determining the level of agreement between simulations and experiments. A qualitative match of strain fields is found to be possible, and even a correlation between identified active slip systems is achieved for both samples. By performing simulations with variations in important model aspects, the influence of
these on the agreement with experiments is studied, to obtain an optimal numerical model configuration.
Additionally, mechanisms behind the activation of specific slip systems are analyzed, and it is investigated to what extent the numerical model can capture these mechanisms. For low M/F misorientation, the martensite localization seems to have an important influence on the deformation in ferrite. The results suggest that both slip transfer and Schmid factor influence which slip systems activate in ferrite. CP
FEM seems to capture this behavior in a global sense, but also appears to have some limitations, as local slip patterns do not match exactly. For high M/F, the situation is more complex but the microstructure appears to have a stronger effect than the martensite deformation.
| Date of Award | 21 Apr 2022 |
|---|---|
| Original language | English |
| Supervisor | Johan P.M. Hoefnagels (Supervisor 1) |