Junhui Yan

Banded microstructure in AA2024-T351 and AA2524-T351 aluminum friction stir welds: Part I. Metallurgical studies

Abstract Results from careful investigations of the banded microstructure observed on horizontal transverse cross-sections in all AA2024-T351 and AA2524-T351 aluminum friction stir weld (FSW) joints indicate the presence of periodic variations in (a) the size of equiaxed grains, (b) micro-hardness, and (c) concentration of base metal impurity particles (i.e. constituent particles) that correlate with the observed band spacing. The latter trend is more distinct in AA2024-T351, which has a higher volume fraction of constituent particles resulting in easily recognizable particle rich regions on horizontal cross-sections near the mid-thickness of the joint and well-defined variations in hardness. In AA2524, the trends are more muted but clearly visible. Results from recent numerical simulations of the FSW process enable interpretation of the trends in grain size along the weld centerline in terms of the time–temperature cycle experienced by the material. Specifically, the AA2524 FSW joints having low power and high input energy (i.e. the slow FSW), exhibit micron-size grain structure across both bands. Conversely, the fast and medium FSW in AA2524 have higher maximum temperatures, and a corresponding six-fold increase in grain size.

Banded microstructure in 2024-T351 and 2524-T351 aluminum friction stir welds Part II. Mechanical characterization

Abstract A series of micro-mechanical experiments have been performed to quantify how the friction stir welding (FSW) process affects the material response within the periodic bands that have been shown to be a common feature of FSW joints. Micro-mechanical studies employed sectioning of small samples and micro-tensile testing using digital image correlation to quantify the local stress–strain variations in the banded region. Results indicate that the two types of bands in 2024-T351 and 2524-T351 aluminum FSW joints (a) have different hardening rates with the particle-rich bands having the higher strain hardening exponent, (b) exhibit a periodic variation in micro-hardness across the bands and (c) the individual bands in each material have the same initial yield stress.

Three-dimensional digital image correlation to quantify deformation and crack-opening displacement in ductile aluminum under mixed-mode I/III loading

Fractures in ductile thin-sheet structures, such as a fuselage or automobile panels, often occur under complex loading conditions. In particular, under remote mixed-mode I/III loading conditions, a cracked structure is subjected to a combination of in-plane tension and large out-of-plane tearing deformation, which may lead to crack tip fields consisting of all three fracture modes (modes I, II, and III). Understanding such fracture events in ductile materials is an important component of the structural integrity analysis of load-bearing structures containing ductile, thin sheets. Due to the complex nature of mixed-mode I/III fracture in ductile thin-sheet materials, reports of experimental investigations are very limited in the literature. We configure three-dimensional digital image correlation (3D-DIC) systems to acquire full-field deformations during the loading and stable tearing processes. The full-field deformation measurements are used to characterize the stable crack extension behavior of an aluminum alloy undergoing quasistatic and dynamic mixed-mode I/III loading. Results confirm that 3D-DIC is an excellent methodology for measuring 3-D deformations in the presence of large out-of-plane warping and motion, both dynamically and statically. Data obtained during the fracture process indicate that the introduction of a mode III component into the loading process alters the crack tip displacement and strain fields relative to those measured in the nominally mode I loading. Furthermore, the measured crack-opening displacement (COD) values during quasistatic and impact mixed-mode I/III fracture show that (1) COD is nearly constant for crack extension beyond 2 mm and (2) COD under combined-mode I/III loading is four times larger than observed during mixed-mode I/II or mode I fracture of the same material, indicating that the magnitude of the critical COD is a function of loading mode in highly ductile, thin-sheet materials.

Process–structure–property relationships for nugget and heat affected zone regions of AA2524–T351 friction stir welds

Abstract The effects of weld tool rotational speed ω, welding speed v and z-axis force FZ during friction stir welding of the aluminium alloy 2524-T351 on the resulting process response variables, nugget microstructure, nugget tensile properties and heat affected zone hardness variations were investigated. For the range of conditions examined, the results indicate that ω has the dominant effect on nugget properties and structure, that optimum nugget tensile properties can be obtained by increasing ω to obtain a peak temperature that is just below the incipient local melting temperature, and that excessive values of ω result in low nugget ductility because of localised embrittlement near the weld crown. The study has also shown that the peak weld temperature is inversely related to the measured torque T 0. The T 0–ω relationship appears to be a useful guide for weld modification, as it is indicative of conditions leading to overheating in the nugget region.

Processing and banding in AA2524 and AA2024 friction stir welding

Abstract An attempt has been made to clarify the relationship between the periodicity of process parameters and that of the micro-/mesostructural banding in friction stir welding (FSW). The oscillation of FSW parameters was measured and analysed during baseline experiments without welding and during a series of FSW of AA2524 and AA2024 aluminium alloys; the banded structures in all the welds were also characterised. It is found that independent of the tool runout, there are well defined periodic variations in the process forces, the translational resistance force F x) and Z axis force F z of the welding tool in all joints. The measured band spacing is closely correlated with the oscillation period of process forces F x and F z, exerted on and by the welding tool. The direct and consistent relationship observed between periodicity in variation of F x and F z and the metallurgical structure, e.g. grain size, particle distribution, band width and microhardness, indicate that the formation of banded structures in FSW joints is related to the periodic variation of welding parameters.

Effect of initial base metal temper on mechanical properties in AA7050 friction stir welds

Abstract The effect of initial base metal temper on mechanical properties in AA7050 friction stir welds was investigated. AA7050 plates, 6·4 mm thick, with three different heat treatment conditions (T7451, T62 and W), were friction stir welded using nearly identical welding parameters, followed by post-weld aging approximating a T7451 heat treatment. The microstructure, transverse hardness profiles and transverse tensile properties were characterised for these three welds. Experimental results show that preweld heat treatment conditions of AA7050 base metal have significant effect on the mechanical properties of the friction stir welds. Friction stir welding of AA7050 in the W condition, followed by post-weld aging, can change the fracture location from HAZ to weld nugget and increase tensile and yield strengths and elongation in transverse tension, relative to welding in T62 or T7451 conditions.

Identification of Strain-Rate Sensitivity With the Virtual Fields Method

The experimental identification of parameters governing the elasto-visco-plastic constitutive behavior of materials is a key issue which usually relies on performing simple mechanical tests for which a closed-form solution for the corresponding equivalent mechanical problem is available. These tests, such as tension or compression on prismatic specimens or torsion on thin tubes, usually lead to uniform states of stress and strain and therefore, the identification can be performed from a few strain data obtained through strain gages or extensometers. However, in order to fully characterize the material behavior, multi-axial tests are often necessary, requiring costly testing machines and difficult specimen design to obtain uniform stress states in some area of the specimen. Moreover, the models describing the elasto-plastic or visco-plastic behavior of materials are governed by several parameters which cannot be directly determined from these experiments in all cases.

Local Stress-Strain Response of an Axial X100 Girth Weld Under Tensile Loading Using Digital Image Correlation

An important design, construction and maintenance concern for pipelines is the integrity of flaws in the girth welds. Numerous fitness for purpose codes are available to assess weld flaws, many of which were calibrated with reference to wide plate test data. Often, wide plate tests are conducted on girth welded pipe in the as-received condition, i.e. without application of a pipeline coating. The area adjacent to the weld is thus subjected to a thermal cycle due to the heat generated from the welding process. In some pipe materials this thermal cycle might be sufficient to induce strain aging. It is not clear how the welding process changes the behaviour of the area next to the weld. The results of such wide plate experiments are very important in assessing the acceptable flaws in a girth weld under a strain-based design. Therefore, it was important to understand the extent of the aging, specifically the stress-strain behaviour on either side of the girth weld. This paper presents results of cross-weld tensile tests, which utilized a two-dimensional digital image correlation (DIC) technique to determine displacement, and thus infer strain. The local strains were mapped to global stress to obtain local constitutive properties every 12.5mm along the length of the specimen. The DIC test results were very consistent and were also similar to results obtained from standard circumferential tensile tests at corresponding locations. The strength of the specimens, as defined by the relative strength of their stress-strain curves, was found to be highest in the girth weld region, to drop in the HAZ, and then to reach a plateau in the base metal. It was also shown that strain localization in one of the HAZ regions was clearly visible during the loading process and the near-HAZ regions had a stress-strain response with a yield stress value higher than the base metal. This behaviour was observed at 12.5mm away from the girth-weld centerline in both the transverse and longitudinal directions. The reason for this slight change of behaviour can be attributed to the effect of heating supplied to this part during welding (strain aging). The described DIC technique is very promising in obtaining local strain fields within very small areas of the tested specimens.Copyright

Modeling of Stable Tearing with Crack Tunneling in Specimens of Different Thickness

Stable tearing with crack tunneling in ductile materials has been commonly observed, but a quantitative understanding of the 3D complicated crack tunneling phenomena is very limited. In particular, the correlation between the driving force and fracture toughness during stable tearing with crack tunneling has not been well evaluated. In the current study, modeling efforts have been made to simulate stable tearing events with crack tunneling and slanting under remote tension in A2024-T3 plate specimens of two thickness values (2.286mm and 6.35mm). It is observed that values of CTOD, stress constraint and the Lode stress parameter vary in the specimen thickness direction. For the thinner specimen, a higher stress constraint and a lower CTOD in the midsection of the crack front are found, and the critical CTOD decreases approximately linearly with an increasing constraint and is weakly dependent on the Lode stress parameter. The variations of CTOD, the stress constraint and the Lode parameter in the two plate specimens in the thickness direction are very similar, but the magnitudes of these parameters near the midsection of the crack front increase significantly in the thicker specimen. The approximate linear correlations in the two specimens between constraint and CTOD are not identical, probably due to the coupled effects of constraint and the Lode stress parameter in this midsection region of the crack front.