Residual stresses and distortion in additively-manufactured SS316L-IN718 multi-material by laser-directed energy deposition: A validated numerical-statistical approach

In the present work, a numerical-statistical approach was used to study the distortion and residual stress in the additive manufacturing (AM) of SS316L-IN718 multi-material structure. Also, the structure was fabricated by laser-directed energy deposition (L-DED), and experimental measurements via thermocouple and micro X-ray diffraction were performed to validate the numerical results. In a good agreement with the experiments, the thermomechanical modeling of the process showed that through-thickness residual stresses (build direction) are distributed as a large accumulation of compressive residual stress in the mid-length of the structure-substrate interface and the maximum tensile residual stress at its two free ends. The critical development of longitudinal residual stresses (scanning direction) at the interface of the structure due to the sudden change in properties has caused the formation of the maximum tensile residual stress (up to 475 MPa) in this area and immediately changed its nature to compressive that can easily lead to localized cracking. Besides, Von Mises stress distribution indicated the concentration of stress at the interfaces of structure-substrate and base alloys, in such a way that it causes yielding in the SS316L section. On the contrary, less (plastic) strain occurred in the IN718 section due to higher yield strength. The results of the statistical modeling using multiple response analysis in the response surface method (RSM) also determined the desirable minimum level of longitudinal strain and stress of the structure (about 0.013 and 320 MPa, respectively) in the laser power range of 300–380 W and the powder per unit length of 5–25 g/m. In addition to minimization, it was demonstrated that sharp changes in longitudinal strain and stress at the interface of multi-material structure can be effectively reformed to mild changes by designing compositional gradients through thermodynamic calculations, owing to the approximation of the properties of adjacent layers.