Fatigue behavior and mechanical characterization of austenitic stainless-steel components produced through Wire+Arc Additive Manufacturing

Abstract

Purpose: Additive manufacturing has enabled innovative, cost and resource efficient mass customization of products over a wide range of manufacturing industries. Under the broad spectrum of options for manufacturing of metallic components, Wire + Arc Additive Manufacturing, WAAM, offers the combination of high design freedom and productivity at a comparably low cost. As a maturing technology, the application in the building industry of steel and stainless-steel products becomes increasingly interesting. Research is nevertheless necessary for the identification of the material’s performance under dynamic loading. It has been demonstrated that the manufacturing procedure inherently produces a heterogenous material matrix and wavy surface, which may be susceptible to fatigue. Methodology: Based on a broad literature review concentrated on the material performance of stainless steels as welding consumable, stainless-steel SAE316L is chosen as the material of interest for characterization. Test coupons are extracted from heat-treated wire+arc-additive-manufactured single-bead walls to perform monotonic and fully reversed strain-controlled fatigue tests in both longitudinal and build directions. The material’s microstructural characteristics are documented. Test methods include Wavelength Dispersive X-Ray Fluorescence for chemical characterization; X-Ray Diffraction for phase identification and quantification; and Scatter Electron Microscopy and Optical Microscopy for qualitative sub-grain and grain microstructural characterization. The latter methods are subsequently used to characterize the monotonic and fatigue fracture surface morphology of post-mortem test samples. A constitutive model for the material is described based on the results of the monotonic and cyclic behavior in both orthogonal directions. Inspection of the fracture surface of the tested coupons is realized to validate the assumptions made for the application of the fatigue models at hand. Addressing a commonly cited problem, the macroscopic surface waviness of the printed material is quantified, and its effect as a stress concentrator is numerically modelled and verified through a smaller group of tested coupons retaining the original surface waviness. Finally, suggestions for the extrapolation of these results are done for the design of structural components produced with the studied material and designed for Fatigue Limit State. Findings: Microstructural characteristics include a strong metallurgical texture in the build orientation, limiting the validity of X-Ray diffraction results. Ferrite is quantified at approximate 13% volume ratio through image analysis of SEM micrographs. Possible traces of σ-phase are identified through optical microscopy, specifically in inter-grain boundaries, although further work is necessary to confirm their constitution. Fracture surface analysis indicates mostly ductile material deformation, confirmed by the elongation capacity of tensile test results, and a predominantly smooth fatigue fracture surface populated by ridges with isolated striation locations, indicating crack propagation at a low stress intensity amplitude. The material shows marginally lower performance in terms of quasi-static characteristics when compared to the one cited by the consumable manufacturer for all-weld material. This remains nevertheless consistently higher than the allowed design values published by commercially available standards. A notoriously low elastic modulus is observed, source of which remains unidentified; the elastic modulus is measured between 54% and 58% of that of commercially available hot-rolled stainless-steel alloys. Cyclic material performance corresponds with all-weld predicted performance for ferritic-austenitic stainless-steels, where the first few cycles are characterized by a slight material hardening, followed by a consistent softening throughout the cyclic life of the material. Average fatigue performance is undistinguishable from mean design values proposed for structural stainless steels under standardized testing conditions. At a 95% reliability and 75% confidence interval, the material performance is superior to the design curves proposed by ASME. The theoretical stress concentration obtained through numerical analysis of the surface geometry is modeled as a function of maxima between 3.150 and 3.927 in the build orientation with a 75% reliability for component lengths of 20 mm for the former value and 100 mm for the later value. Similarly, the values 1.944 and 2.413 in the longitudinal direction describe the stress concentration effect for 20 mm and 100 mm component lengths correspondingly. Originality / Value: The research contributes to the characterization efforts made for the improvement of Wire+Arc additive Manufacturing technologies. This work is carried out both at a microstructural and a bulk level, allowing to establish a correlation between the manufacturing parameters, the microstructural morphology and the mechanical behavior of a given material. Design aids are presented at a typically cited reliability level for engineering practice, accompanied by a thorough description of design considerations based on the material’s inherit characteristics as a product of WAAM process.