Residual stresses induced by laser shock peening in orthopaedic Ti-6Al-7Nb alloy. (November 2020)

• Record Type: Journal Article
• Authors: Shen, Xiaojun; Shukla, Pratik; Subramaniyan, Anand Kumar; Zammit, Ann; Swanson, Philip; Lawrence, Jonathan; Fitzpatrick, Michael E.
• Is Part Of: Optics & laser technology. Volume 131(2020)
• Publication Date: 2020-11
• Subjects: Laser shock peeing -- Incremental hole drilling -- Artificial neural network; Optics -- Periodicals; Lasers -- Periodicals; Electronic journals; 621.366
• DOI: 10.1016/j.optlastec.2020.106446
• Languages: English
• ISSNs: 0030-3992

Abstract

Highlights: The LSPned residual stress filed of Ti-6Al-7Nb alloy was characterized by IHD method. Laser energy & overlap are proportional to the magnitude of the near-surface residual stress. Laser energy & overlap are in proportion to the affecting depth. The LSPned residual stress was predicted by ANN and the results are in good agreement with IHD data. Abstract: Prolonged usage can lead to fatigue failures in prosthetic implants. This is an emerging concern for the prosthesis designer and as well as for end-users. Laser shock peening (LSP) can be used to improve the life of prostheses by introducing a compressive surface stress that delays or arrests the initiation of cracks. We report measurements of residual stresses induced by LSP in a medical-grade Ti-6Al-7Nb alloy. In this paper, the LSP introduced residual stress field was characterized by using Incremental Hole Drilling (IHD) with predictive modelling using an Artificial Neural Network (ANN). The Ti-6Al-7Nb samples were processed by a pulsed Nd: YAG laser with laser energy of 3 J, 5 J and 7 J with overlaps of 33%, 50% and 67%. The IHD results show that a range of −42 MPa to −516 MPa compressive residual stresses are formed at the near surface. The residual stress value and affecting layer depth are proportional to laser energy and overlap. What is more, a gradient descent learning algorithm was employed in ANN prediction. The microhardness, laser energy, overlap and depth are set as input parameters while the residual stresses are output. The predicted results have a good agreement with the experimental data with accuracy of 96.16% and 95.16%. It proves that ANN is applicable in LSP residual stress simulation when the experimental data is limited. The work in this paper enables to have subsequent predictions of fatigue life of implants via residual stress fields for future applications.