The growth and coalescence of multiple coplanar short fatigue crack in AA7050-T7451 aluminium alloy

2017-02-22T03:57:35Z (GMT) by Tan, James Tzeq Chiang
The effects of interaction and coalescence of multiple short cracks (sub mm) in AA7050-T7451 aluminium alloy and prediction of fatigue life have been studied in this thesis. Previous studies have largely been on the growth and coalescence of “long” cracks (> 1 mm) and existing predictive methods may not be applicable to the growth and coalescence of short cracks. Uncertainties persist in the effects of crack coalescence on crack growth rates and how the effects of crack interaction should be taken into account. Some studies have emphasized the importance of accounting for the interaction between coalescing cracks which may accelerate their growth rates, while others have reported contradictory findings. Quantitative fractographic methods were used to investigate how two or more coplanar short cracks of different initial configurations and dimensions interact and coalesce with each other, and to quantify their effects on the fatigue life of test specimens. The initial flaws were prepared on the surface of test specimens to study the effects of different parameters (such as separation distance between cracks, the relative crack sizes and the presence of multiple cracks) on their growth and coalescence. Relationships between these parameters and the rate of coalescence (i.e. determined by the crack growth rates throughout the process of coalescence) were established and used to develop new methods for modelling the growth and coalescence of multiple short cracks. The results from these new predictive methods were compared with existing methods (i.e. ASME [1] and BSI PD6493 [2] design codes, and the method by Iida [3]), and experimental data from this study and of other researchers. Major experimental and theoretical findings which include the development of new predictive methods are summarised in the following items: 1. For the growth of individual short cracks in AA7050-T7451 aluminium alloy test specimens, a variant of the original Hartman and Schijve model produced the most accurate fatigue life predictions, compared to the Generalized Frost and Dugdale model and the (ΔK+.Kmax)0.5 model. 2. Fatigue lives predicted using design codes (i.e. ASME, BSI PD6493) and the Iida method were found to be overly conservative when compared with the present experimental data and do not accurately describe the coalescence of short cracks. 3. The short cracks examined in this study showed very small increase in growth rates as the coplanar cracks grew toward each other. Therefore they can be assumed to grow independently of each other until their tips physically meet. 4. Quantitative fractography of the fatigue specimens tested in this study showed that after two approaching crack tips meet, the crack growth rates in the surface direction were retarded for a period, and only after the combined cracks have grown into a larger semielliptical shape, do they return to the growth rates of a single crack. A new method was developed to account for this behaviour in which the cracks would grow as independent cracks until they meet and if the crack length ratio, c2/c1 ≤ 0.25, the smaller crack (i.e. c2) can be neglected. Predictions using this method agreed well with experiments. 5. Quantitative fractography was used to study the evolving shapes and sizes of the fracture area of a newly coalesced crack. Results have shown that the size of the “remaining” area required for two coalesced cracks to form a single semi elliptical crack was associated with the period of crack growth “retardation”. Based on this relationship, a new fracture area method was formulated and shown to be superior to existing fracture area methods for predicting the coalescence of paired cracks. 6. The growth and coalescence of multiple (i.e. more than two) short cracks of varying sizes on the same and on different planes, under low constant amplitude cyclic loading were also studied. The experimental results corroborate previous findings that the growth of the ‘largest’ crack dominated the fatigue life of a specimen with very little influence from relatively shorter cracks.