Observation and modelling of the short crack growth behaviour in aluminium alloys for the designing of aircraft structures

Kurzfassung

The fatigue crack growth behaviour of long and short cracks in two aluminium alloys used in the aircraft industry was characterised in this work. For the prediction of the life time of aircraft structures and especially for defining the intervals between maintenance checks of the structures it is important to have reliable data of the behaviour of cracks emanating from initial flaws under cycling loading. The fact that aircrafts are in operation up to their designed service goal and beyond, brings new problems that are usually summarised as “ageing aircraft problems. Short cracks, which can not be detected by non-destructive testing methods, start to grow from different rivet holes creating critical scenarios defined as Multiple-Site-Damage or Widespread-Fatigue Damage. These scenarios can end in a catastrophic failure of the structure and subsequently in the loss of the aircraft. Therefore it is necessary to know the crack growth rates of short cracks under cyclic loading conditions and to be able to predict these growth rates, so that a safe and economic designing of aircraft structures is possible. Especially about the short crack growth behaviour in the alloys Al2524-T351 and Al 6013-T6 little is known. The extensive fatigue crack growth tests on long cracks showed that the alloy Al2524-T351 is more damage tolerant than Al 6013-T6. Crack growth rates at the same stress intensity factor range and the same stress ratio are slightly smaller in Al2524-T351. The threshold values for the propagation of long cracks are approximately 10% higher in Al 2524-T351 compared to Al 6013-T6, but this difference is decreasing with increasing R-values. A new test method was developed, which allows to determine fatigue crack growth curves (da/dN – ΔK) for every stress ratio desired with only one single specimen. In a comparison with da/dN – ΔK curves determined with the multiple specimen method proposed by the standard E647-00 of the American Society for Testing and Materials, it could be shown that this method delivers equally reliable data. Even the sigmoidal shape of the fatigue crack growth curves, typical for aluminium alloys, is depicted. Additionally one gets threshold values for the crack growth at different stress ratios. This new method offers an enormous cost reduction to characterise the long fatigue crack growth behaviour of materials. For the short corner crack growth experiments three different methods to initiate a precrack in the material were investigated. The completely new method to cut very small notches (10 – 50μm) into the specimen with a Focussed Ion Beam gave the best results. Precracks were initiated under compressive cyclic loading, which were less than 100μm long (notch + crack). It was also found that when using notches with less than 30μm length, cracks often initiated at other sites like precipitation particles or scratches in the surface of the specimen. Although the method of creating precracks under compressive cyclic loads was actually developed for through thickness notches, it could be proven that it also works very well for notches at the edge of a hole. The fatigue crack growth tests on specimens with a short corner crack showed for both aluminium alloys the typical short crack behaviour. The cracks grow faster compared to long cracks at the same stress intensity factor range and at the same stress ratio. They also grow below the threshold of the stress intensity factor range for the propagation of long cracks. Below the threshold values determined with high, constant Kmax-tests short cracks grew but stopped after a extension of a few μm. The threshold values determined for short cracks with the stepwise increasing load method are in the range of the effective threshold of the stress intensity factor range for long cracks mentioned in the literature. For both aluminium alloys this effective threshold is around 1MPam1/2. Different models to predict the fatigue crack growth rates in materials were compared and their suitability to describe the short crack growth phenomenon was analysed. Although some of them give good approximations, none was found to be satisfying. Some gave non-conservative predictions; others are not very handy tools for an engineer in designing structures. The applicability of the stress intensity factor to describe the short crack growth behaviour was also examined. Although short cracks violate the rules of linear elastic fracture mechanics, it was shown that with corrections derived from elastic plastic fracture mechanics concepts, the typical short crack growth behaviour “disappears”. On the basis of the fatigue crack growth data derived from constant Kmax-tests a model was developed with which it is possible to predict the short crack growth behaviour at different stress ratios. The comparison with the experimental short crack growth data showed very good agreement. This model is a handy tool for engineers in the designing of structures and components and needs only little experimental expenditure.