PSI - Issue 83

Niccolò Vilotta et al. / Procedia Structural Integrity 83 (2026) 246–255

249

(a) (b) Figure 2: (a) Experimental setup for the static tensile test on the dog-bone specimen, with temperature monitoring by means of an infrared camera; (b) detail of a DMLS welded AISI 316L dog-bone specimen after fracture. The tensile campaign was used both to characterize the global mechanical response of the investigated configurations and to provide the calibration data required for the local constitutive reconstruction. A toe-offset correction was subsequently applied to the tensile-test data prior to stress-strain conversion in order to remove the non-physical initial compliance arising from slack take-up, grip seating, and slight machine-specimen misalignment. The corrected strain origin was determined by back-extrapolating the tangent drawn to the steepest segment of the initial loading branch, namely the region exhibiting the maximum local slope at the onset of the effective elastic response, to the zero-load axis. The resulting intercept was defined as the toe offset and subtracted from the measured elongation, while the load signal was left unchanged. This preprocessing step provides a more rigorous definition of the strain origin and prevents an artificial underestimation of both the initial stiffness and the subsequent strain-dependent tensile response. Let F be the applied load and Δ L the measured elongation. Let ( Δ L t , F t ) indicate the point at which the tangent with the maximum slope is identified in the initial portion of the load-elongation curve. The corresponding local slope is defined as: k t = dF d ( Δ L ) | t (11) Accordingly, the toe offset and the elongation effectively associated with the actual deformation of the specimen can be written as: ΔL toe = ΔL t - F t k t ⁄ (12) Δ L eff = Δ L- Δ L toe (13) The corrected elongation ΔL eff was then used for the subsequent stress-strain conversion, ensuring that the derived tensile response reflected the actual deformation of the specimen from the onset of effective loading. Thermographic monitoring of the specimens’ surface was carried out simultaneously with the tensile tests by means of a compact infrared camera, with a noise equivalent temperature difference of 0.04 K and a thermal resolution of 0.01 °C, focused on the gauge section, (see Figure 2(a)). The temperature field was recorded during loading to track the onset and evolution of irreversible deformation and to compare the thermal response with the mechanical one [16–19]. 3. Results and discussion The mechanical response of the investigated DMLS-manufactured AISI 316L specimens is analysed at both global and local scales. The tensile behaviour of the non-welded and welded configurations is first presented to assess the effect of welding on the overall strength and deformation response and to define the reference quantities used for constitutive calibration. The focus is then shifted to the welded cross-section, where hardness mapping and hardness-based constitutive reconstruction are employed to identify the mechanical gradients across the base material, heat-affected zone, and weld metal. The interpretation is complemented by thermographic evidence collected during loading and by the macroscopic features of the fracture surfaces after failure.