PSI- Issue 9

Yu. Matvienko et al. / Procedia Structural Integrity 9 (2018) 16–21 Author name / Structural Integrity Procedia 00 (2018) 000–000

19

4

length increments. Thus, a total number of interferograms is equal to 2×3×7=42. Symmetrical configuration of all interferograms with respect to the crack line clearly demonstrates the validity of mode I conditions. This means that SIF and T-stress values can be reliably determined from the relationships developed for modified version of the crack compliance method. Initial experimental data allow constructing the dependencies of CMOD, SIF and T-stress values from total crack length for different loading cycles. This information provides the dependencies of fracture mechanics parameters for cracks of the fixed length as a function of the loading cycle number, which are shown in Figure 3. In these figures the loading cycle number is presented as a life-time percentage (see Table 1).

b

a

Fig. 3. CMOD (a) and SIF (b) values as a function of life-time percentage. Graphical information in Figures 3a and 3b reveals indicators, which can be used for fatigue life assessment. Indeed, CMOD 0   v related to the first crack length is equal to zero for N = 5000 and N = 6000 cycles. CMOD value 1   v is negative for N = 5000 and N = 6000 cycles. Positive SIF value 2  I K for N = 5000 cycles is less than positive SIF values for the second crack lengths at the other stages of cyclic loading. This means that previous measurement point N = 4000 cycles corresponds to reaching 63% of life-time. Negative SIF value 2  I K for N = 6000 cycles is related to 95% of life-time. It should be noted that no surface cracks were observed after applying N = 4000, 5000 and 6000 cycles. Lower plot in Figures 3a illustrates a process of residual stress relaxation near the hole due to cyclic loading. These curves prove that the influence of cold hole expansion is considerably displayed for all three crack lengths. This means that the area of residual stress influence on a crack growth rate is, as minimum, a circle of 0 3 2( )   r a ≈ 8.0 mm diameter for all stages of cyclic loading. On the other hand, we can conclude that the process of fatigue crack initiation and growth causes a marked residual stress relaxation for cracks both inside and outside the compressive residual stress region near the cold-expanded hole. Experimental SIF vs. the loading cycle number curves shown in Figure 3b exhibit 63% life-time indicator as a point of SIF fall beginning for C N = 4000 cycles. These points are reliably detected for both negative and positive SIF values as it is shown by vertical lines in Figure 3b. This fact means that a number of specimens essential for life time predictions might be considerably reduced. Plots that demonstrate experimental T-stress values vs. the loading cycle number for the first and the second crack length increments are also obtained. This information is of great importance for creating two-parametric fracture criteria. 3. Welded joints Experimental investigations are performed for 4 aluminium plates ( E = 74,000 MPa, μ = 0.33) of dimensions 180×80×4 mm 3 , each of which includes an electronically welded joint. All plates are manufactured by cutting initial welded plate by equal parts orthogonally to the weld. The first specimen is tested without influence of cyclic loading.