Development of a constrained ageing technique for residual stress reduction and its application in extra-large 7050 component manufacture

Abstract

The production of extra-large aluminium alloy components, such as wing panels, ribs, stringers, etc., is important for the aviation industry. Producing qualified extra-large components with required shape and mechanical properties has been challenging using traditional manufacturing processes. This is attributed to the residual stresses (RS) generated in the process, which cannot be sufficiently reduced before machine finishing, thus causing issues with structural integrity and component distortion. To deal with this problem, a novel RS reduction technique was conceived to maximize the RS reduction in the manufacturing process. The technique is called constrained ageing (CA). The work in this thesis concentrates on studying the mechanism and verifying the effectiveness of the CA technique in reducing the RS in a 7050 extra-large T-joint component, which is currently experiencing RS induced distortion problems in the aviation industry. Constrained aging is a newly developed technique that fixes a component’s shape at the ageing temperature (120 °C for 6 hours and subsequently 177 °C for 7 hours for 7050), enabling both the precipitation hardening and the stress relaxation process. This technique is expected to optimise ageing for RS reduction and determine the required shape whilst improving the mechanical properties of components. Both experimental and modelling work have been employed in this thesis to theoretically study the CA mechanism and verify its effectiveness. To enable a fundamental understanding of the CA mechanism, firstly, a set of uniaxial CA tests (also called stress relaxation ageing (SRA) tests) were carried out. Different initial stresses and pre-strain values were applied to the samples to achieve stress relaxation curves under various loading conditions. The yield strength and microstructure evolutions were examined from tensile tests and TEM, respectively. A set of physically based constitutive equations were developed to mathematically describe the stress–strain–time relationship and the corresponding age hardening behaviour during the uniaxial CA process. These equations were derived from dynamic ageing and power-law creep relations and are able to describe the stress relaxation, age hardening and their interactions in 7050. The equations were calibrated using the uniaxial CA experimental results and were prepared as a subroutine for the subsequent finite element (FE) simulation of the CA process. The effectiveness of the CA technique to reduce the RS and whilst achieving the target mechanical properties was examined by incorporating the CA technique into an improved manufacturing process and examining the RS evolution during the process. The manufacturing process consists of water quenching (WQ) followed by cold rolling (CR) and a subsequent CA. The whole manufacturing process was applied to a lab-sized T-section panel, which is the scaled-down version of the extra-large T-joint. RS measurements were made at each processing stage (i.e. after WQ, after CR and after CA) using both neutron diffraction (ND) and X-ray diffraction (XRD) methods. Hardness and displacement profile of the T-section panel were also measured during the process. The aim of the tests was to examine the effectiveness of the whole manufacturing process, especially that of the CA process, to reduce RS. Finally, a FE model of the scaled-down T-section panel, based on the calibrated constitutive equations, was built to simulate the RS and yield strength evolution during the CA process. To ensure correct initial conditions (i.e. residual stresses and prior-plastic deformations before CA), the predicted residual stress distributions after WQ and after CR were taken from previous work. The RS distributions at all processing stages were compared with the experimental results in this work. It has been concluded that the CA technique can reduce up to ~ 50% of the residual stresses and finally produce the T–section panel with RS ranging within ~ ±100 MPa and hardness ~ 159 HV10. In addition, higher pre-strain levels facilitate the RS reduction in CA. Therefore, it is suggested that before CA, relatively large plastic deformation is expected to maximise the RS reduction during CA. More importantly, an appropriate FE model of the CA process has been built. The model successfully predicted the RS and yield strength distributions, where the FE simulated RS and yield strength distributions agreed well with experimental results. From the FE predicted RS distribution during CA, it has been concluded that a shorter CA time (~ 13 h) is recommended to accommodate the sacrifice in components’ yield strength due to stress relaxation, whilst maintaining most of the RS reduction in CA. The comprehensive research work on the CA technique and its practical application in the improved manufacturing process provides a new solution for solving the RS problem of extra-long structures, such as T-joints in this work, where the quenching inducted RS (i.e. from ~ ±250 MPa) was reduced to within ±100 MPa.