Blunt and sharp notches: Revisiting the limit notch radius via the averaged SED method and validating it against a wide fatigue strength reduction database

When designing mechanical components, their functional requirements often lead to geometrical discontinuities with severe stress concentrations and gradients. These discontinuities, known as notches, can markedly reduce the structural reliability and fatigue strength of components. Depending on the notch severity, conventional point-based approaches may significantly overestimate their detrimental effects on fatigue behaviour. Notches are generally classified as blunt or sharp with the fatigue behaviour of the sharp ones not accurately captured by point-based approaches. Numerous studies have attempted to define the transition between these two behaviour and to develop design methodologies capable of consistently addressing both. Among these, the averaged Strain Energy Density (SED) method has demonstrated high accuracy and robustness for both blunt and sharp notches. In this work, the SED method is employed to identify a limiting condition, expressed through a limit notch radius, rho limit, that distinguishes between blunt and sharp notches. This condition is investigated through numerical simulations and validated against an extensive fatigue database from the literature. Defining the limit condition as a notch radius simplifies components design and may also serve as guideline for determining the required notches tolerances. Finally, a methodology is proposed for fatigue-oriented material selection, coupling bulk material properties, component geometry and notch sensitivity. Indeed, in fatigue design, the highest-performing component is not necessarily obtained using the material with the highest intrinsic fatigue strength. For sharp notches, materials with lower intrinsic fatigue strength, but reduced notch sensitivity, can indeed yield superior fatigue performance. The methodology can be readily extended to lightweight design applications.