E-Archive


Vol. 5
March Issue
Year 2004
BACK

Science Update


in Vol. 5 - March Issue - Year 2004
A Computer Modelling Capability for Shot Peen Forming



Figure 1: Shot peen forming is used to produce the lower wing skins of many civil airliners such as the Airbus A340


Figure 2: Equivalent plastic layer used to model the residual stresses


Figure 3: 5mm thick plate shot peen formed at Sheffield University


Figure 4: Contour plot showing the predicted vertical deflection of a peen formed plate modelled using ABAQUS

P.L. Blackwell*, M.D. Griffiths*, T. Ward* & S. Gardiner+

Shot peening is a metal working process in which the metal surface is impacted by many small projectiles termed ‘shot’. The geometry, hardness and size of the shot can be varied according to the application, but often the shot takes the form of steel spheres up to a few millimetres in diameter. Many within the metal finishing industry will be familiar with the use of shot peening to enhance fatigue properties. This effect is the result of the compressive residual stress field that can be generated near the surface of the component. The presence of the compressive stress inhibits the early growth of fatigue cracks and can also have benefits in terms of corrosion properties [1]. A secondary consequence of the shot peening process is that the residual stress distribution developed in the material may induce distortion in the component. This effect is used in the shot peen forming (SPF) process to straighten or shape thin flexible metallic structures, and has been recognised as a useful manufacturing process for various aircraft components since the 1960s. For example, the process is commonly used to form the lower wing skin panels for many civil airliners (Figure 1). These wing skins are very long (up to around 30m) but relatively narrow (2m) and of varying thickness (between 2mm and 32mm). The size range of components that can be manufactured, and the fact that the process is dieless, demonstrate the potential versatility of the technique, but also make its analysis more complicated. When hard, usually spherical, shot is fired at a metal target and the impact velocity is sufficiently high, the material underneath each impact undergoes plastic deformation. This causes compression and elongation of the surface (Figure 2). Following peening, the surface compressive stresses are balanced by elastic tensile stresses deeper within the material. The thickness of the plastically deformed surface region is a function of a number of parameters including the material hardness, shot type and shot velocity. The plastically deformed layer causes the metal panel to undergo several geometric changes in order to reach stress equilibrium. As noted, the upper layer elongates in order to alleviate the compressive stress, thus creating a growth differential across the panel. This differential causes the metal part to curve in the direction of the peened surface. Any numerical method used to model the SPF process must therefore be able to predict both these bending and stretching effects. Presently, much of the success of SPF lies with the skill of the operator and often for new parts a degree of trial and error is involved before a process route is established. It was in this context that an initial study was carried out to try to better quantify the SPF process through a systematic investigation aimed at producing a computer modelling capability. By understanding the effects of the various process parameters on the residual stress pattern, the model could be used for process optimisation. The model may also function as a useful design tool, allowing the designer to make full use of the capabilities of the forming process. The work carried out at QinetiQ; part funded by Airbus UK and the Department of Trade and Industry (DTI) was directed at modelling the peen forming of lower wing skin panels in AA2024. Some initial model development had been carried out at Cambridge University [2,3]. The basic modelling concept was founded on an original idea by Levers et al [4] who suggested that, given that SPF is a cold forming process, it should be viable for modelling purposes to apply a through-thickness temperature gradient to mimic the surface residual stresses in the material. Within the model, the temperature gradient generates material expansion similar to that induced by the plastic deformation as a result of the impact of the shot (Figure 2). In order to reduce the required processing time in the ABAQUS finite element (FE) package, shell elements are used rather than 3D solid elements to model the peen forming process. The temperature gradient is applied through the thickness of the panel in several layers; the magnitude of the temperature applied to each layer and the depth of each layer were related to the peening conditions used. A relatively high temperature is applied to the top layer compared to the bottom layer, thus the resulting difference in thermal expansion generates plastic deformation and the plate bends permanently upwards when any external loads are removed. A series of trials were carried out at the University of Sheffield in which the process parameters (e.g. shot size, air pressure, coverage, the angle of incidence between the incoming shot and the sheet surface) were varied and measurements made of the deformed panels produced. The initial trials on smaller samples were used to calibrate the FE model. An example of one of the panels is shown in Figure 3. The effect of the shot impact is to cause the sheet to become bowed towards the nozzle or wheel from which the shot is projected. The initial investigation was confined to relatively small panels, e.g. 200mm x 200mm x 5mm; at a later stage, however, panels up to 1m x 1.5m x 6.35mm were used. These were again measured and the results used to progressively refine the FE model. Figure 4 shows one of the modelling predictions in the form of a contour plot indicating the vertical displacement of the panel. Under most peening conditions investigated a reasonable agreement was obtained between the predicted and actual results. The model, though still in need of further refinement, represents an important step forward for the shot peening process. However, it is worth noting that no model for shot peen forming will be entirely accurate. The process contains a certain degree of unpredictability – even if two sheets of metal from the same batch of material are peened under nominally the same conditions there will be small differences in the geometry of the final product. This is because the distribution of shot impacts naturally varies slightly from part to part. Differences in the initial residual stress patterns in various batches of material prior to processing may also represent a significant challenge to the shot peen former. Any fully automated peen forming process will need to take this into account through closed loop feedback systems involving in-process measurement and adjustment of the peening conditions. Looking to the future, the laser shock peening process has many advantages over the traditional method involving metal shot. It is inherently more controllable, it produces significantly less surface damage, but can generate a much deeper layer of compressive stress. At present, the use of laser shock peening is confined to relatively small components e.g. aero-engine blades for the improvement of fatigue properties and resistance to foreign object damage. In the future, however, as the technology progresses, it may be possible to apply this technique to large panels. The modelling technique outlined in this article would apply equally well to this new process. Acknowledgement The authors are pleased to acknowledge sponsorship from the DTI Aeronautics Research Programme which part-funded this project. References [1] Metal Improvement Company, Inc., Shot Peening Applications, 8th Edition, 2001. [2] Gardiner, D.S. – A model for Shot Peen Forming with Prestress Conditions. University of Cambridge DPhil dissertation, May 2001. [3] Wang, T. Numerical Simulation and Optimisation for Shot Peen Forming Processes, University of Cambridge DPhil dissertation, July 2002. [4] Levers, A & Prior, A. - Finite element analysis of shot peening. Journal of Materials Processing Technology (1998) Vol.80-81 p304-308.


Author: P.L. Blackwell, M.D. Griffiths, T. Ward & S. Gardiner
* Authors based at QinetiQ Ltd, Farnborough, GU14 0LX + Author based at AIRBUS UK For Information: E-Mail: plblackwell@qinetiq.com