Issue 78

G. Palomba et alii, Fracture and Structural Integrity, 78 (2026) 233-247; DOI: 10.3221/IGF-ESIS.78.14

Metallographic preparation and microhardness mapping To reveal the different metallurgical regions, the welded surface, obtained from a section dog-bone specimen, was sequentially polished using 1200 to 4000 grit silicon carbide papers, in order to obtain a mirror surface, and etched with glyceregia (3 parts glycerol, 2-5 parts hydrochloric acid, 1 part nitric acid). Microhardness tests were performed on the treated surface using a Vickers indenter with a load of 3 N and a dwell time of 15 s. Measurements were taken along a longitudinal line crossing the entire weld region, with 0.5 mm spacing between adjacent indents. Applied methodology The adopted methodology integrates experimental hardness testing, mechanical property derivation from hardness measurements through empirical correlations, experimental tests using Digital Image Correlation (DIC) and numerical simulation (FEM), experimentally validated. This multiscale strategy allowed hardness-based reconstruction of the local mechanical behavior in welded joints of AISI 316L stainless steel, combining localized hardness data with full-field deformation analysis. The workflow comprised five main stages: 1. Definition of an experimental benchmark; 2. Microhardness mapping across the welded joint; 3. Conversion of hardness values into local mechanical properties through established correlations; 4. Generation of true stress–strain curves for each metallurgical zone (BM, HAZ, WM); 5. Implementation of these curves in a nonlinear FEM to simulate the tensile response, validated against experimental DIC strain fields. Tensile testing and DIC procedure Quasi-static tensile tests have been carried out using a servo-hydraulic MTS 810 (MTS, USA) universal testing machine mounting a loading cell of 250 kN. Tests have been performed in stress control conditions, using a stress rate of 6 MPa/s. The images were acquired using a Basler acA2440-35uc camera with a full resolution of 2464 × 2056 pixels. Owing to the geometry of the specimens, a two-dimensional 2D DIC approach was considered appropriate, as the flat geometry of the specimens did not require the measurement of out-of-plane displacements. To prepare the specimen surface, a thin layer of anti-reflective white paint was first applied. Subsequently, a random speckle pattern was produced by spraying black paint, ensuring a fine and homogeneous distribution of speckles over the region of interest. Images were recorded at a frame rate of 2 fps, in accordance with the iDICs recommendations for quasi-static mechanical testing, where the low deformation rate allows a reduced acquisition frequency without compromising correlation accuracy. The camera was manually focused to achieve optimal image sharpness and contrast before the test. Prior to the correlation analysis, the images were preprocessed using a Gaussian low-pass filter ( σ = 2.0 pixels) to reduce high-frequency image noise while preserving the quality of the speckle pattern. The DIC analysis was performed using the python open-source software µ DIC [21]. Unlike conventional local subset-based DIC algorithms, µ DIC employs a global finite-element formulation in which the region of interest is discretized into a mesh of finite elements. Consequently, the concepts of subset size and step size are not applicable. Instead, the spatial resolution is determined by the finite-element mesh. For the present analyses, a mesh with element dimensions of 34 × 34 pixels was adopted, based on the speckle quality and the imaging scale determined by the minimum focus distance of 200 mm. By doing so, it was ensured that each element contained a proper number of speckle-pattern points to be correlated at the reference frame, set to be the first one. An example of speckle pattern, of the region of interest encompassing the gauge section of the specimens and of the elements grid is displayed in Fig. 2.

R ESULTS Experimental results using DIC technique

igital Image Correlation (DIC) analysis provided high-resolution, full-field strain maps that revealed distinct differences in the deformation behavior between welded and non-welded AISI 316L specimens. In the adopted DIC coordinate system, displayed in Fig. 3, the y-axis corresponds to the tensile loading direction of the dog-bone specimen, whereas the x-axis is transverse to the loading direction. Therefore, ε yy is hereafter denoted as longitudinal strain, D

236