Co-authors​ :

 Wenjie TU (NETHERLANDS), John-Alan PASCOE (NETHERLANDS), René ALDERLIESTEN (NETHERLANDS) 

Integration of advanced composite materials in highly loaded structures, such as in aerospace or large wind turbines, requires the ability to predict their fatigue behaviour. Predictive models for one-dimensional (1-D) fatigue delamination growth have been established relying on standardized test methods such as Double Cantilever Beam (DCB) and End Notched Flexure (ENF). But how much do these models teach us about delamination growth in full scale structures? These models have widespread application in the research literature, but fail to capture the multidirectional (2-D) nature of delamination growth in full-scale structures, rendering such models inadequate for accurate prediction of fatigue damage growth in real-world structures.
In this study, an experimental method is developed to investigate 2-D delamination behaviour of CFRP panels with embedded defects under mode II dominant fatigue loading. Detection and monitoring of delamination growth are accomplished through Digital Image Correlation (DIC) and ultrasound scanning (C-scan) techniques. A cyclic energy dissipation method is implemented to examine the energy release rate (ERR) across fatigue cycles.
Based on C-scan and fractographic observations, similar delamination patterns of two initial delamination interfaces, 0º//0º and 0º//90º, are depicted. A significant interaction between different damage modes is observed in 2-D fatigue delamination, driven by the varying local mode mixity at the delamination front. Fibre-matrix debonding and matrix cracking are dominant fracture mechanisms in regions where the growth direction deviates from the directing ply orientation. Delamination migration occurs in regions where the growth direction is perpendicular to the ply orientation. Pure mode II delamination emerges as the primary fracture mechanism in regions where the growth direction aligns with the ply orientation. Intriguingly, the migrated delamination growth follows the same trend observed at the original interface. The growth rate of delamination area, derived from DIC and C-scan analysis, exhibits an initial rise within a limited number of loading cycles, followed by a subsequent decrease as delamination growth continues. The deceleration of the growth rate is attributed to the development of a fracture process zone and fibre bridging. Nonetheless, a constant total ERR is deduced during the decelerating phase, indicating a slow and stable delamination growth.
Unfortunately, there is currently no prediction model capable of describing 2-D fatigue delamination growth. This is because the existing models depend on calculating the ERR along a 1-D delamination path, fitting coefficients from 1-D test data. None of these variables can be directly applied to 2-D delamination growth due to local mode mixity and interplay between damage modes. Consequently, the development of a more generalized prediction model is imperative to adequately describe the nature of planar delamination growth.