Issue 78

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

K EYWORDS . AISI 316L, Hardness correlation, Welding, Local properties, Numerical modelling, DIC.

I NTRODUCTION

A

ustenitic stainless steel AISI 316L is widely employed in critical sectors such as marine, nuclear, and biomedical industries, as a consequence of some essential properties, such as excellent corrosion resistance, good formability and weldability [1], the latter being a fundamental feature to guarantee efficient connections when assembling complex parts [2]. However, welded joints often represent the weakest point in terms of mechanical resistance, due to microstructural heterogeneity across the weld metal (WM), heat-affected zone (HAZ), and base metal (BM) [3,4]. Consequently, a reliable methodology to retrieve local properties of welded joints regions is necessary for an in-depth knowledge of their mechanical response. In this context, Vickers or Brinell hardness measurements are gradually becoming popular, since they allow non-destructive testing of each sub-zone from which deducing mechanical properties [5] such as yield strength, ultimate tensile strength, and even full stress-strain curves [6,7,8]. Standards like ISO 18265:2014 and studies such as those by Lopez and Fatemi [9] and Kamaya [10] offer semi-empirical correlations that relate hardness to Ramberg– Osgood parameters, enabling stress–strain curve reconstruction with negligible alteration of the tested component. The obtained information on local properties can therefore be used for detailed analysis of welded joints. For instance, finite element models describing welded response, can be populated with the experimental data from hardness tests to assess the mechanical response of welded joints accounting for realistic local properties heterogeneity. Previous works compared experimental tensile tests and strain fields data from digital image correlation (DIC) with stress-strain curves derived from hardness measurements, demonstrating good agreement and hence the feasibility of such procedure [6,7,8]. In a key benchmark study, Dong et al. [8] demonstrated also that local hardness-based strain-life parameters, adjusted for weld toe roughness, is able to successfully estimate fatigue initiation life using a strain-based approach. Thus, accounting for mechanical properties heterogeneity, deducible from cost-effective and non-destructive measurements, allows the assessment of accurate stress and strain distributions at critical locations [11], which is essential, as proven by recent literature [12], to avoid improper estimation of deformation and failure in welded joints. Considering the relevance of hardness measurements for subsequent deduction of material properties, researchers are developing alternative and more efficient methods and technologies, as reported in Ref. [13], which provides a comparison between two modern hardness-based methods to get stress-strain data. In particular, a traditional instrumented indentation testing, based on load-depth data, and a novel Profilometry-based Indentation Plastometry method, which uses the 3D residual indent profile and inverse FEM, were analysed and the accuracy in reconstructing the true stress-strain curve of metals was evaluated. The correlation methodology between Vickers hardness measurements and yield stress was further improved by Takakuwa et al. [14], who focused on surface-treated stainless steels, where the treatments induce residual stresses, responsible for alterations on local yield stress. Based on this premises, they established a relation between hardness and yield strength, introducing a correction for residual stress. Although scientific literature more frequently addresses welded joints made of structural or ferritic steels, comparatively few investigations have examined austenitic stainless steel welds along with the associated mechanical property mapping and FE modeling, particularly those involving AISI 316L, which is more commonly studied in its not-welded form [15]. Notably, to the best of the authors' knowledge, no studies have specifically addressed the modeling of the full stress-strain curve for AISI 316L butt-welded joints derived from hardness measurements, especially when combined with advanced full-field validation methods such as 3D-DIC. The integration of full-field experimental measurements with numerical modeling has received increasing attention due to its enhanced capability for accurately characterizing the mechanical heterogeneity of welded joints. Within this framework, the use of DIC to identify local strain distributions and mismatch effects in welds has provided robust validation for FE models that incorporate variations in material properties across different regions [16,17]. Multiple studies have demonstrated that neglecting local strength gradients among the weld metal, heat-affected zone, and base material can lead to substantial inaccuracies in predicting deformation, damage initiation, and fatigue behavior [18]. Concurrently, indentation-based methodologies, including conventional microhardness mapping and instrumented and inverse indentation techniques, have proven effective for non-destructively reconstructing realistic local stress-strain curves of welded steels [19], particularly when integrated with calibrated constitutive models [20]. Within this context, the research activity presented in the current paper is aimed at promoting the application of minimally invasive local measurements for stainless steels welded joints mechanical response prediction, potentially transferable to future non-destructive evaluation procedures. The developed approach integrates microhardness measurements, validated

234