No abstract

International audienceManufacturing industries perform mechanical surface treatments like shot peening at the end of the manufacturing chain to protect important working parts. This treatment modifies the near surface of the treated part with the introduction of compressive residual stresses due to the repeated impacts of the shot. Then, the treated part exhibits, not only a longer life, but also a better fretting behavior, an improved resistance to corrosion… The objective of the present paper is first to study the relation between the process parameters and the material state (residual stress and plastic variables…) for a complex geometry. Next, a numerical tool is proposed, able to predict this material state in a time frame that is consistent with industrial constraints. The originality of the proposed approach thus consists in the chaining of the different steps. The first step is to choose the process parameters for the shot peening process considering conventional or ultrasonic shot peening and model the shot dynamics for a complex geometry. Once the impact velocity field is known, the objective is to compute the local incompatible plastic deformation field due to the repeated impacts using analytical methods. Then, a finite element model is used to compute the residual and deformation fields in the considered mechanical part. The complete method has been performed on the model of a gear, a mechanical part that is most often shot peened and exhibits a complex geometry

In this study, a frequency-domain approach based on the harmonic balance method coupled to a predictor-corrector continuation algorithm is implemented for the qualitative analysis of blade-tip/casing contacts in aircraft engines. Unilateral contact and dry friction are taken into account through a regularized penalty law. To enhance the robustness of the methodology, particular attention is paid to the mitigation of the Gibbs phenomenon. To this end, the employed Alternating Frequency/Time scheme features a Lanczos σ-approximation so that spurious oscillations of the computed nonlinear contact forces become negligible. This approach is applied in combination with a model reduction technique on an industrial compressor blade: NASA rotor 37. In order to assess the influence of both the contact law regularization and the Lanczos σ-approximation, obtained results are thoroughly compared to an existing time integration-based numerical strategy relying on a Lagrange multiplier-based approach for contact treatment and that was previously confronted to experimental results. Presented results underline the very good agreement between the proposed methodology and the reference time integration numerical strategy. The proposed developments thus complement existing results on blade-tip/casing contact adding a much needed qualitative understanding of the interaction and an accurate assessment of the contact stiffening phenomenon.

The study presented in this paper focuses on the analysis of rubbing interactions ---including unilateral contact and dry friction--- between a rotating fan blade and a rigid casing by frequency domain methods. Two previously published Harmonic Balance Method-based methodologies are assessed: 1) an approach relying on augmented lagrangians and 2) a second method using a regularized penalty law combined with a Lanczos $\sigma$-approximation filter. As a reference point for this comparison, a time-domain numerical strategy relying on a Lagrange multiplier-based contact treatment is considered. All computations are run with the NASA rotor 67 fan blade, an open industrial blade geometry. As it undergoes structural contacts, this blade features an intricate dynamics response, thus making it a challenging case study for nonlinear iterative solvers. The contact scenario is chosen to be an ovalization of the casing with no external forcing. This scenario induces highly nonlinear reponses of the blade and complex phenomena such as isolated frequency response curves. The results presented underline a very good agreement of the different strategies with the reference time marching approach. An in-depth comparison is made with an emphasis on nonlinear frequency response curves and time signals. Finally, a physical analysis of these isolas is presented through an in-depth investigation of the main modal contributions for each computed solution. In the end, this paper provides new qualitative elements allowing for a better understanding of rubbing interactions that may not be efficiently obtained with time marching procedures.