PSI - Issue 28

Rhys Jones et al. / Procedia Structural Integrity 28 (2020) 26–38 Author name / Structural Integrity Procedia 00 (2019) 000–000

30

5

measurements being taken from the DCB test, the initial delamination in the specimens was firstly grown to give a relatively short pre-crack extension length of a p - a o to form a natural crack front, as is the commonly adopted procedure. (It should be noted that if a pre-crack (of extension length, a p - a o ) is not used in the DCB test but, instead, the test is started directly from the starter film crack of length a o , then optimistically high values of the fatigue resistance (and indeed also the quasi-static interlaminar fracture energy) will be measured. This is because the starter film crack, of length a o , possesses a relatively blunt crack tip compared to that of the sharp pre-crack which is naturally-grown ahead of the starter film crack prior to the commencement of the test.) However, a refinement in these studies was that, for any given DCB test, the test program now involved growing the crack under fatigue loading for a certain, relatively short, distance from this initial value of the pre-crack extension length, a p - a o , whilst taking readings of the number of cycles, crack length, load and displacement, in order to deduce the corresponding values of da/dN , ��� and ��� . Next, the fatigue test was halted and then repeated, but now with respect to the new, longer, pre-crack extension length, a p - a o , that was present in the DCB test specimen. Hence, a series of plots of log da/dN versus log ∆√ , via Equation (1), were obtained as a function of the value of pre-crack extension length, a p - a o . More recently, additional fatigue data on this ‘M30SC/DT120’ CFRP using multidirectional lay-ups have been reported by Yao (2018a). One multidirectional lay-up studied was of 32 plies and had a lay-up of [(±45/0 12 / ∓ 45)//(±45/0 12 / ∓ 45)], giving a nominal thickness of 5.0 mm, and the other was of 48 plies and had a lay-up of [(±45/0 20 / ∓ 45)//(±45/0 20 / ∓ 45)], giving a nominal thickness of 7.5 mm. The fatigue testing of these two multidirectional lay-ups was conducted using the DCB test specimen and the same type of test program was followed as described above, such that an essential feature was again that the fatigue tests were conducted with different lengths of the pre-crack extension, a p - a o , i.e. with different degrees of fibre-bridging initially present in the pre-crack extension used in the DCB test specimen. 3.2 Fatigue crack growth (FCG) rate results Fig. 1 shows the results from this latest work by Yao et al. (2018a) using multidirectional DCB tests where the log FCG rate, da / dN , is plotted against the log ∆√ , see Equation (1). This figure contains the results from the two duplicate sets of test program data that were obtained for the [(±45/0 12 / ∓ 45)//(±45/0 12 / ∓ 45)] 32 ply lay-up and these sets of measured log da/dN versus log ∆√ values are labelled ‘Yao Test 1’ and ‘Yao Test 2’. Also shown is the one set of data generated for the [(±45/0 20 / ∓ 45)//(±45/0 20 / ∓ 45)] 48 ply lay-up and this test program is labelled as ‘Yao Test 3’. All these results are given as a function of a p - a o , where this term represents the pre-crack (i.e. pre delamination) extension length initially present in the DCB test specimen prior to any cyclic-fatigue fracture measurements being taken. Several interesting points emerge from Fig. 1. Firstly, in general, as the pre-crack extension length, a p - a o , used for the fatigue test is increased, the plots of log da / dN versus log ∆√ move to the right. This implies that there is a retardation of the FCG rate as the value of a p - a o , is increased. Thus, as the value of a p - a o is increased then, for a given value of ∆√ , the corresponding value of da/dN is lower, i.e. at a given value of ∆√ the FCG rate slows as the value of a p - a o is increased. This retardation has been shown to arise from the development of fibre-bridging that occurs behind the crack tip across the faces of the DCB tests, which is the more extensive the more the initial pre-extension crack length, a p - a o , has been allowed to grow from the starter film crack, a o , prior to the start of measurements being taken from the DCB fatigue test. Secondly, there is inherent scatter in the results shown in Fig. 1, as would be expected when testing CFRPs under fatigue loading. For example, the data for the ‘Yao Test 1 9.9 mm’ mainly lies to the right of that for the ‘Yao Test 2 11.4 mm’, when the reverse situation would be expected. Also, the data for the ‘Yao Test 1 20.9 mm’ lies somewhat to the right of that for the ‘Yao Test 3 33.7 mm’, when again the reverse situation would be expected. It is unclear whether this inherent scatter only arises from the scatter in the extent of fibre-bridging that develops behind the crack tip in the test specimens, or arises from other causes such as (a) the reproducibility of the CFRP material itself and/or (b) the reproducibility in the manufacturing and testing of the DCB specimens. Notwithstanding, however it arises, then it must be taken into account when undertaking material selections, comparisons, design and service-lifing studies.