At IRDL, Théo Sévédé has developed a bimodal fatigue model to better understand material failure.
Théo Sévédé, chercheur·e au sein de IRDL (Institut National Polytechnique de Bretagne).
Thèse soutenue en 2025 à l'école doctorale École doctorale Sciences pour l'ingénieur et le numérique.
Référencée dans le réseau ABES/STAR, cette thèse répond aux critères de rigueur de l'enseignement supérieur français.
Material fatigue represents a critical issue in the field of mechanics, often responsible for the premature failure of components in service. This phenomenon is particularly concerning in sectors such as aerospace, automotive, and energy, where the reliability of components is paramount. The research conducted by Théo Sévédé at the Institut National Polytechnique de Bretagne highlights the crucial importance of considering fatigue in the design of systems. Fatigue tests, although relatively simple in appearance, are often time-consuming and require considerable resources. They allow for the generation of Wöhler curves, essential for assessing the resistance of materials to repeated stresses.
Historically, the endurance limit of a material was set at 10^7 cycles, a threshold established due to the constraints of conventional machines used for these tests. However, with the evolution of technologies and manufacturing methods, modern systems can be subjected to stresses of up to 10^10 cycles. This is particularly true in applications where low amplitude vibrations are ubiquitous, leading to unexpected and often catastrophic failures. For example, in the aerospace sector, components subjected to constant vibrations can crack well before the end of their expected lifespan, jeopardizing flight safety.
The emergence of ultrasonic fatigue machines, which operate at frequencies of 20 kHz, has opened new perspectives for exploring these extended lifespan ranges. These machines allow for the simulation of extreme fatigue conditions in a reduced time frame, thus providing a unique opportunity to better understand the mechanisms of fatigue at very high cycle numbers (VHCF). The main objective of this research is to examine the relevance of thermometric measurements to better understand fatigue behavior at VHCF. A one-dimensional (1D) data processing protocol has been proposed and validated, allowing for the calculation of thermal dissipation of the material in an ultrasonic configuration. This protocol has enabled the generation of a very high-frequency self-heating curve, providing valuable data to analyze the impact of frequency on the thermal response of the material.
A bimodal fatigue model has been developed, based on a distinction between the core and the skin of the specimens, as well as an energy criterion. This predictive model relies solely on the self-heating curve and on material-specific parameters, making it particularly suitable for practical applications. Very high-frequency fatigue tests have been conducted to validate this model on ML340 steel, which exhibited bimodal behavior up to 10^10 cycles. The results of this study underscore the importance of thermometry in evaluating material fatigue and open new perspectives for the design of components subjected to extreme stresses.
This research could transform the way engineers design and size mechanical parts. By integrating thermal data into their calculations, it becomes possible to better anticipate potential failures and optimize the design of components. This could also reduce costs associated with component failures, a major issue in sectors where production stoppages can lead to significant financial losses. Furthermore, improving the safety of systems in service is another direct implication of this research. By extending the lifespan of materials, we contribute to the sustainability of infrastructures and the reduction of waste, a crucial issue in the current context of ecological transition.
In summary, innovation in the field of material fatigue is not only a scientific advancement but also a significant economic opportunity. In a world where competitiveness is increasingly fierce, companies that can leverage this new knowledge will have an undeniable advantage in the market. By integrating these advancements into their design and manufacturing processes, they will not only improve the performance of their products but also meet the growing demands for safety and sustainability.
Accéder à l'étude complète
Sources et accès
Théo Sévédé. Étude et modélisation des propriétés en fatigue à très grand nombre de cycles à partir de mesures d’auto-échauffement sous sollicitation cyclique. Matériaux et structures en mécanique [physics.class-ph]. École Nationale Supérieure de Techniques Avancées, 2025. Français. ⟨NNT : 2025ESTA0001⟩. ⟨tel-05351748⟩