Materials Science Forum Vols. 524-525

Abstract: In this study we present a direct comparison between residual strain measurements carried out on the same inertia friction weld using ENGIN-X at ISIS, UK and the new strain scanner SALSA at ILL, France. ENGIN-X is a time of flight (TOF) instrument, which receives neutrons from a neutron spallation source, while the SALSA Strain-Imager, a high resolution diffractometer, is based at a research reactor source with a continuous neutron flux and is operated with a constant wavelength. The purpose of this study was to demonstrate a confidence in crosscomparing future strain measurements to be performed at ENGIN-X and SALSA. Measurements were carried out on medium size inertia friction welded nickel superalloy test-piece, which show no significant crystallographic texture across the weld line. The results demonstrate that, even though residual stresses determined on SALSA only rely on a single peak analysis (in this case the (111) reflection), the results show excellent agreement with the measurements carried out on ENGIN-X, where strain is determined from multi-peak Rietveld analysis.

Abstract: Investigations on welded joints of a low strength steel and of an age-hardened aluminium- alloy have been performed to compare the effects of different post-weld treatment methods. The experimental results show, that the methods which are working with help of ultrasonic activation lead to a higher depth of penetration of the compressive residual stresses as shot peening does. The cold worked surface area is deeper and the hardening effect can be found in deeper layers. Furthermore the shape of the weld toe is changed strongly due to the cold working of the surface with the consequence of a change of the weld toe geometry. Anyway the fatigue strength improvement is comparable to that which is generated by shot peening (steel) or less than after shot peening (aluminium). The reason is, that the shape of the treated weld toe profile must not be necessarily smoother due to the ultrasonic hammer peening. Depending on the treatment parameters also sharp defects can be produced at the boarder of the treated zone which may compensate the beneficial effect of the compressive residual stresses.

Abstract: This study looks at the effect of laser welding on residual stresses in sheets of moulded magnesium alloy (AZ91). The modifications are significant even though they are localised along the weld line. The welding process produces a gradient of residual stresses from the weld line to the base metal and throughout the thickness of the weld zone. The distribution of these residual stresses has been qualitatively explained by the kinetics of cooling and by the nature of the multiphase material. Finally, it has been demonstrated that a pronounced crystallographic texture modifies local distribution of residual stresses. These results show that there is a complex coupling between the metallurgical, thermal and mechanical aspects generated by high power welding procedures.

Abstract: Residual stresses in CO2 laser beam welded AA6056 Al-sheets of 3.2 and 6.0 mm thicknesses were studied using neutron and high-energy X-ray diffraction. The influence of the temper T4 and T6 before welding, the effect of sheet thickness in T6 temper, and the effectiveness of post-weld heat treatments T6 and T78 for the reduction of residual stresses in the sheets welded initially in the T4 temper were examined. It was found that tensile longitudinal stresses were significantly higher when welded in T6 than in T4. With the increase in the sheet thickness from 3.2 to 6.0 mm, the residual stress profile was affected more than the stress level when welded in T6. For 3.2 mm sheet, the post-weld heat treatments T6 and T78 did not lead to a significant reduction in residual stresses when welded in T4.

Abstract: In this study the distribution and magnitude of residual stresses in a T-joint of aerospace grade aluminium alloy weldment was determined using neutron diffraction. A 2 mm thick AA 6013 sheet (as clip) was laser beam welded to a 6 mm thick AA 6056 base plate (as skin) to resemble the “short distance” welded clip-skin joints of an airframe. The total length of the weld was 120 mm and it was welded using 3.3 kW Nd:YAG laser source and 12% Si containing wire from one side only. No post weld heat treatment was applied after the welding. Start (run-in) and end (run-out) locations of the T-joints are generally considered as high risk areas with respect to solidification cracking and crack initiations under external loadings. It is of interest to investigate the weld residual stress fields at these locations to develop optimum joint design. Therefore, strain measurements have been performed not only in the middle of the weld seam but also at the run-in and run-out locations of these short distance welds. Higher longitudinal tensile residual stresses are detected at the run-out locations than the run-in locations. The measurements in the clip showed that the clip has a longitudinal tensile stress peak away (about 8 mm) from the weld seam.

Abstract: Microscopic and macroscopic deformation of a polycrystal due to an applied load can be modelled using crystal plasticity implemented within the Finite Element (FE) framework. However, while macroscopic predictions can readily be validated against conventional monotonic and cyclic stress-strain curves, verification at the microscopic level is harder to achieve, since it involves calibrating the predictions for stresses and strains in individual grains, or in grains grouped by certain criteria (e.g., orientation). In this paper an elasto-plastic polycrystal finite element model is introduced, and its calibration is performed at a mesoscopic level via comparison with neutron diffraction data obtained experimentally. Time-of-flight (TOF) neutron diffraction experiments carried out on ENGIN-X instrument at ISIS involved in situ loading of samples of C263 nickel-based superalloy. In order to compare the numerical predictions of the FE model with these experimental data, the corresponding mesoscale average elastic strains must be extracted from the results of the simulation by employing a ‘diffraction post-processor’. This provides a much improved technique for the calibration of FE formulation and enhances the confidence in the model. The FE diffraction post-processing procedures are discussed in detail, and comparison between the model predictions and experimental data are presented.

Abstract: Residual stresses are of great importance during the entire production cycle of high-grade steels. The use of modern tools based on the finite element method is steadily increasing to optimize heat treatment processes. As for industrial purposes it is often not possible to measure the entire set of material data a sensitivity analysis shows the relative influence of material properties related to phase transformation on the residual stresses during hardening. Subsequently the application of the numerical heat treatment model is shown on two examples: The magnitudes of residual stresses during the quenching of a forged bar in different quenching media are compared. The paper concludes with a numerical simulation of the heat treatment of a die used for extrusion processes. Phase distribution and residual stresses after gas quenching of the tool are presented.

Abstract: Finite element prediction of residual stresses in a 3-bead letterbox-type repair weld is investigated in the present study. The repair is performed on a 2¼CrMo low alloy ferritic steel plate, containing a machined central groove where three weld beads are deposited using AL CROMO S 225 2¼CrMo electrodes. The proposed simulation procedure, which is based on decoupled thermal and mechanical analyses and the “birth and death of elements” technique, is evaluated through comparison of predicted stresses with neutron diffraction testing data. Parametric studies include modelling aspects such as 2-D plane strain versus 3-D analysis, re-melting of weld material during sequential bead deposition, melting of base plate near the fusion line and annealing. It is concluded that numerical results come, in general, in satisfied agreement with the experimental data.

Abstract: Welding, which is a largely used process in the mechanical manufacturing, well known to induce high-level residual stresses. The level of residual stresses is of great importance for the lifetime of welded components used in mechanical engineering industry. The use of the ultrasonic method for the evaluation of the residual stresses is based on the acoustoelastic effect, which refers to the change in velocity of the acoustic waves propagating in a strained solid. In the case of welding, the microstructure modifications observed in the heat affected zone (HAZ) and the melted zone (MZ) also induce variations of the velocity of the acoustic waves. The superposition of the two effects, stresses and microstructure, results in over-estimating the levels of stresses. This work which was completed in collaboration with CETIM is a contribution to this problem. The experimental study was carried out on P460HLE and P265 steels welded sheets. The results obtained by the ultrasonic Lcr wave technique were compared with those obtained by the hole drilling technique. This work confirms the possibility of evaluating the residual stresses induced by welding using the ultrasonic method.