Laser Shock Peening allows the introduction of deep compressive residual stresses into metallic components. It is applicable to most metal alloys used for aerospace applications. The method is relatively expensive in application, and therefore development studies often rely heavily on Finite Element Modelling to simulate the entire process, with a high computational cost. A different approach has been used recently, the so-called eigenstrain approach. The present study looks at the feasibility of applying the eigenstrain method for prediction of the residual stress in a sample that contains curved surface features. The eigenstrain is determined from a simple geometry sample, and applied to the more complex geometry to predict the residual stress after Laser Shock Peening. In particular the prediction of residual stress at a curved edge, and for different values of material thickness, have been studied. The research has demonstrated that the eigenstrain approach gives promising results in predicting residual stresses when both the thickness and the geometry of the peened surface is altered.

Residual stresses play a fundamental role in mechanical engineering. They can be generated by manufacturing processes or introduced purposely by surface treatment technologies. One of the most recent technologies developed to introduce residual stresses is Laser Shock Peening. Since it is a relatively expensive technology, a fundamental role is played by the Finite Element Analysis approach to predict the final residual stress profile. The FEA approach consists of either direct simulation of the LSP process or the application of the eigenstrain approach. The application of the eigenstrain theory in predicting residual stresses after LSP treatment in curved edges is the subject of this research.

Laser shock peening (LSP) is a reliable, repeatable, and successful surface technique for introducing high magnitude, deep compressive residual stresses that can significantly increase the fatigue life of metallic components. However, depending upon how the LSP treatment is applied, the induced residual stresses can result in the undesirable deformation of the part. In this work, traditional shot peening has been applied over LSP as a means to optimize the stress distribution at the surface of a part while constraining deformation. A single edge notch test specimen of AA7075 was laser peened local to the notch region and then shot peened over the entire central region. The resulting residual stress distribution has been characterized using neutron diffraction to measure the stress distribution in the bulk, and it was compared with (1) incremental center hole drilling to measure the stress distribution up to depths of ∼1 mm and (2) near-surface stresses obtained in a previous X-ray diffraction (XRD) study on nominally identical specimens subjected to the same surface treatments. For regions where the two techniques overlap, the residual stresses are in good agreement (within uncertainty). Comparing the bulk stresses obtained from neutron diffraction in this study and XRD data published elsewhere, it can be shown that shot peening applied after LSP has a profound effect on near-surface stresses; however, these effects disappear at depths of ∼0.7 mm or more.

We present a comprehensive study of the effects of internal boundaries on the accuracy of residual stress values obtained from the eigenstrain method. In the experimental part of this effort, a composite specimen, consisting of an aluminum cylinder sandwiched between steel cylinders of the same diameter, was uniformly heated under axial displacement constraint.During the experiment, the sample temperature and the reaction stresses in the load frame in response to changes in sample temperature were monitored. In addition, the local (elastic) lattice strain distribution within the specimen was measured using neutron diffraction. The eigenstrain method, utilizing finite element modeling, was then used to predict the stress field existing within the sample in response to the constraint imposed by the load frame against axial thermal expansion. Our comparison of the computed and measured stress distributions showed that, while the eigenstrain method predicted acceptable stress values away from the cylinder interfaces, its predictions did not match experimentally measured values near them. These observations indicate that the eigenstrain method is not valid for sample geometries with this type of internal boundaries.