R. M. Ward

The Magnitude and Origin of Residual Stress in Ti-6Al-4V Linear Friction Welds: An Investigation by Validated Numerical Modeling

The linear friction welding (LFW) process—of the type required for production of bladed discs for the next-generation aeroengines—was modeled using numerical methods developed previously. An elastic–viscoplastic material formulation was considered to allow for residual stress calculations to be included in the numerical solution. A study of the evolution of residual stress during the welding and cooling processes was made. It was evident that residual stresses arose primarily as a consequence of the cooling of the part after joining is completed. The sensitivity of predicted residual stress to the applied forge load was investigated and compared with measurements from X-ray diffraction methods. Only small changes in residual stress were predicted for large changes in forge load, supporting the hypothesis that the welding process is only of secondary importance to residual stress formation, after the cooling process. Finally, a sensitivity study was carried out investigating the accuracy of modeling the welding process with a simpler, viscoplastic material law, and only switching to the more computationally demanding elastic–viscoplastic law for the cooling modeling. Predictions suggested that this was a sufficient modeling method, given that stress during the welding stage is almost uncorrelated to that present once ambient temperature is reached.

Linear friction welding of Ti6Al4V: experiments and modelling

Abstract Linear friction welding of the Ti6Al4V alloy is studied. A new definition of the energy input rate is proposed, based on an integration over time of the in-plane force and velocity; a strong correlation with the upset rate is then found. The effective friction coefficient is estimated to be 0·5±0·1 for varying frequencies and amplitudes, with only a weak dependence on the processing conditions displayed. A model is proposed that accounts for both the conditioning and equilibrium stages of the process, which is shown to be in good agreement with the experimental data. The model is used to study the mechanism by which the flash is formed. A criterion is proposed by which the rippled nature of its morphology can be predicted.