CAD based model of ultrasonic shot peening for complex industrial parts J. Badreddine a,b , S. Remy b,⇑ , M. Micoulaut c , E. Rouhaud b , V. Desfontaine a , P. Renaud d a SONATS, 2 rue de la Fonderie, 44475 Carquefou, France b ICD-LASMIS, Université de Technologie de Troyes, 12 rue Marie Curie, 10010 Troyes, France c LPTMC, Université Pierre et Marie Curie, 4 Place Jussieu, 75252 Paris, France d SNECMA Evry-Corbeil, route Henri Auguste Desbruères, 91003 Evry, France article info Article history: Received 2 October 2013 Received in revised form 29 May 2014 Accepted 29 May 2014 Keywords: Ultrasonic shot peening Simulation Shot dynamics Complex geometries Spur gear CAD abstract This paper presents a numerical model developed specifically for ultrasonic shot peening (USP). It allows simulating the shot dynamics (trajectories in the chamber and impacts on the peened sample) in industrial configurations. The model supports complex 3D geometries, rotating parts and employs efficient collision detection algorithms for short computation times. The aim is to improve peening cham- ber designs and the choice of process parameters. The algorithm and main assumptions are presented. Numerical studies are then conducted to determine the performances of the model, in terms of compu- tation time. Finally, a case study on a spur gear tests the model in an industrial configuration and shows a high correlation between the numerical results and experimental data. Ó 2014 Elsevier Ltd. All rights reserved. Introduction Ultrasonic shot peening (USP) is a mechanical surface treatment process, developed by SONATS (Stressonic Ò technology) [1], that enhances the mechanical strength [2,3], the fatigue life span [4,5] and the resistance to stress corrosion cracking [6] of the high- added value metallic components, such as bladed disks, compres- sors, gears and nuclear power plants pressure vessels. This is achieved by projecting spherical shot onto the surface of a compo- nent (part), at high velocities (up to 20 m/s), with the help of a sonotrode. The latter is part of an acoustic system that vibrates at ultrasonic frequencies (generally 20 kHz). In an industrial context, customized peening chambers are usually designed for each type of components. It allows holding the part in place and contains the bouncing shot, thus influencing its flow and dynamics. The parts that are shot peened with an ultrasonic process are usu- ally high added value components, like components of airplane engines, with very complex geometries. The measurements of shot velocities are difficult, although it is necessary that it should be well distributed to ensure an adequate residual stress field. The peening chambers are thus designed with trial and error processes to verify the impact density with, for example a coverage analysis. It is thus of interest to construct a predictive model of the shot dynamics for ultrasonic shot peening in any chamber geometry. It is important to specify that the induced residual stresses highly depend on the shot diameter [7] and total mass, the amplitude of the sonotrode [8], the peening time [9], as well as the shot impact velocities [7] and angles [10]. Although the commonly expected outcome of shot peening is subsurface compressive residual stres- ses, it might be just as important to optimize surface characteris- tics: hardening or grain size like in the SMAT process [8,11]. In other words, the main expected outcome of pre-stressing pro- cesses is an increase in fatigue life. In conventional shot peening (CSP), a continuous flow of shot (many kilograms per minute) is projected onto the peened part, making the measurements of shot velocities and angles relatively straight forward [12]. However, in ultrasonic shot peening the few grams of spheres, propelled by the sonotrode, bounce around in the peening chamber. This results in complex and repetitive interactions between the spheres and the rest of the peening setup, i.e. sonotrode, chamber and part. This particular feature of USP makes it difficult to measure experimentally the shot velocities and angles. As a result, the design of USP chambers and the choice of process parameters remain empirical, making it time consuming and partially optimized especially for complex parts. Modeling such processes, from peening parameters to fatigue optimization, can be reached with two types of approaches. The first type consists on chaining models (usually simple ones) and optimizing each step to reach the expected fatigue life. The second http://dx.doi.org/10.1016/j.advengsoft.2014.05.010 0965-9978/Ó 2014 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +33 325 718 556; fax: +33 325 715 675. E-mail address: sebastien.remy@utt.fr (S. Remy). Advances in Engineering Software 76 (2014) 31–42 Contents lists available at ScienceDirect Advances in Engineering Software journal homepage: www.elsevier.com/locate/advengsoft