Fatigue crack growth prediction considering crack-tip plastic phenomena: limitations of current approaches and a novel methodology
- DIRECTORS: Miguel Muñiz and Haritz Zabala
- UNIVERSITY: Universidad de Oviedo
From a damage tolerance perspective, the structural integrity of flawed components is assessed following fitness-for-service principles, so the fatigue crack growth (FCG) observed in a critical component subjected to oscillating loads is acceptable if the resulting failure does not occur during a scheduled period. In this framework, the accurate prediction of FCG is indispensable for making the right run/repair/replace decisions and to design reliable inspection plans. Such accurate prediction implies considering the influence of crack-tip plastic phenomena inherent to the ductile metals widely employed in structural applications. In general, FCG is predicted by an iterative scheme based on the integration of a crack growth law, which relates the crack propagation velocity with a governing fracture parameter postulated to represent the crack driving force. The effective stress intensity factor range (ΔKeff) is the most popular governing fracture parameter. ΔKeff is estimated based on the maximum stress intensity factor (Kmax), which accounts for the action of external applied loads, and the crack opening load (Pop/Pmax), which includes the effect of the plasticity-induced crack closure (PICC) that occurs at the crack wake. Both variables that intervene in ΔKeff can be determined based on a wide variety of approaches, ranging from analytical approximations based on closed-form expressions to numerical methods based on two parallel finite element models.
In this thesis, the limitations of ΔKeff-based analytical approaches are demonstrated by considering a realistic case representative of a cracked wind turbine bearing. The use of closed-form expressions in the estimation of Kmax is found to result in excessively conservative predictions of the remaining useful life and critical crack length, what contradicts the damage tolerance philosophy, so the need for finite elements is concluded. Regarding the estimation of Pop/Pmax, the widely extended closed-recipe use of the popular Newman's crack opening equation is dissuaded even after eliminating the common but arbitrary assignation of the constraint factor value. Alternatively, a novel elastic-plastic fracture parameter-based methodology is proposed for the three-dimensional simulation of FCG considering the interaction between PICC and crack shape evolution, which consists in iteratively solving a single elastic-plastic finite element model for which load history is maintained by remeshing and mapping and crack advance is performed by node releasing. Two advantages are obtained with respect to state-of-the-art numerical approaches based on ΔKeff: 1) lower numerical effort due to the need for a single elastic-plastic finite element model and 2) wider applicability provided by the possibility of analyzing large-scale yielding scenarios. The methodology is applied to a practical case, obtaining results that highly agree with literature trends and experimental measurements, which demonstrates the validity of the methodology and the suitability of the selected crack-tip opening displacement-based fracture parameters.