Cychc pre-stress increases subsequent fatigue crack growth rate in 2024-T3Sl ahrminum ahoy. This increase in growth rate, caused by the pre-stress, and the increased rate, caused by temper embrittlement as observed by Ritchie and Knott, cannot be explained by the crack tip blunting model alone. Each fatigue crack increment consists of two components, a brittle and a ductile component. They are respectively controlled by the ductility of the material and its cyclic yield strength.
THE EFFECTS of various metallurgical factors such as, grain size, alloying element, quenching, aging, and annealing, on fatigue crack growth have been studied. The effects of chemical environment and testing temperature had also been investigated [l-4]. In this study, plate specimens made of 2024-T351 aluminum alloy were pre-stressed cyclically. Subsequent fatigue crack growth rate were measured. The validity of the crack tip blunting model and the cumulative damage model for fatigue crack growth was analyzed in terms of the increased crack growth rate caused by the pre-stress cycles.
2. EXPERIMENTAL RESULTS
The effects of pre-stress cycling on fatigue crack propagation had been studied on 2024-'I'351 aluminum alloy. The static and cyclic stress-strain curves of the material are given in Fig. I. To construct the cyclic stress-strain curve, the total stress range of a stabilized, hysteresis loop is plotted against the total strain range. The static yield strength and the cyclic yield strength range, or(c), are respectively, 52000 and 134,000 psi, Each specimen was cyclically pre-stressed at certain stress range for a given number of cycles. After the pre-stress cycling, a slot was machined into the specimen, and the f&&e crack propagation rates were measured. The specimen geometry is shown in Fig. 2, Due to the load limit of our fatigue machine (20,000 lb), a reduced test section was necessary in order to apply a meaningful pre-stress range, ho,,. Two thicknesses, O*lOand O-08 in. were tested. Because of the difficulty in applying a compressive load to a thin plate, the pre-stresses were tension-tension type, and they were stress-controIled. Four pre-stress ranges, two above and two below the static yield strength of the material were tested in this investigation The fatigue machine was a deflection controlled type. In order to maintain the desired levels, the maximum stress levels and the stress ranges had to be reset frequently during a test. This was especially true for the pre-stress tests above the static yield strength. Figure 3 shows the recording of the load strain curve for pre-stress at 35,750 zf: 29,250 psi. The first ten loading cycles were conducted in an Instron machine. The specimen started to yield and strain hardened at the first cycle. During the next 18 cycles, the recording revealed very small, if any, plastic deformation. The specimen was then put into the fatigue machine and pre-stressed up to 3300 cycles. At the two lower