Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Multiscale constitutive modeling of additively manufactured Al-Si-Mg alloys based on measured phase stresses and dislocation density

: Zhang, X.X.; Knoop, D.; Andrä, H.; Harjo, S.; Kawasaki, T.; Lutz, A.; Lahres, M.


International Journal of Plasticity 140 (2021), Art. 102972
ISSN: 0749-6419
Journal Article
Fraunhofer ITWM ()

To better understand and predict the mechanical properties of additive manufacturing (AM) Al–Si–Mg alloys, developing a physically-based constitutive model is crucial. Among different models, the dislocation-density-based Kocks-Mecking (K-M) constitutive model has been widely used. Unfortunately, two challenges arise when the K-M model is used for the multiphase Al alloys. Firstly, an accurate K-M model demands the separation of phase stresses. Secondly, a thorough analysis of the K-M model involves the measurement of dislocation density during deformation. In-situ neutron diffraction, a powerful method to measure the phase stress and dislocation density in bulk polycrystalline materials under loading, is employed to investigate the AM AlSi3.5Mg1.5 and AlSi3.5Mg2.5 (wt.%) alloys. Based on the present results and reported data of AlSi10Mg, a multiscale constitutive model is developed for different AM Al–Si–Mg alloys. At the microscale, the evolution of dislocation density in the Al matrix with plastic strain can be well predicted by the K-M model. Meanwhile, the developments of the average stresses in different phases with plastic strain can be well captured by the Voce model. The measured microscopic k2/nc values agree with the theoretical value two of the K-M model quite well. Here, nc is the characteristic factor of the microscale Voce model, while k2 is the coefficient associated with the dynamic recovery process in the microscale K-M model. At the macroscale, the mechanical behavior can also be well reproduced by the K-M model and the Voce model. However, the macroscopic K2/Nc ratio is far away from two, where Nc and K2 are the characteristic factor and coefficient of the macroscale Voce model and K-M model, respectively.