Issue 35

L. Songsong et alii, Frattura ed Integrità Strutturale, 35 (2016) 74-81; DOI: 10.3221/IGF-ESIS.35.09 76 Tests type II were accomplished with MTS880 servo hydraulic fatigue test system and an observation system consisting of a digital microscope, a servo motor and a raster ruler. Constant amplitude (CA) loading cycles with stress ratio R = 0.06 were employed. Two specimens, labelled as L-T-2-1 and L-T-2-2, were tested at the maximum stress  max = 140 MPa and 150 MPa, respectively. All the tests were run at 20Hz in laboratory air. Test type III were carried out with SHIMADZU SS-550 SEM. The loading facility in the SEM is a microscopic static and dynamic testing machine produced by SHIMADZU. Constant amplitude loading cycles of R = 0.06 and  max = 150 MPa was employed in the in-situ SEM test. The test was run at 10Hz in vacuum. The specimen tested in this case is marked with L-T-3-1. All the specimen was pre-cracked under constant amplitude load of R = 0.06 and  max = 120 MPa, which resulted in an initial crack whose length is about 0.3 mm from the notch tip. C RACK BRANCHING PROCESS AND MECHANISM ig. 2(a), (b) and (c) show the crack morphologies observed in test type I, II and III, respectively. Two types of crack branching were observed in the crack paths corresponding to all the three test types. The ones marked by circle in Fig. 2 are the familiar type of crack branching which were captured once they appeared. The intersection angles are usually less than 90 degree. However, the branches marked by square in Fig. 2 are obviously different with those marked by circles. The following on discussion will centre on this type of crack branching. (a) (b) (c) Figure 2 : Crack morphologies observed in (a) M(T) specimen under spectrum loading [7], (b) single edge notch specimen (L-T-2-2) tested under CA loading, and (c) L-T-3-1 specimen tested by in-situ SEM under CA loading It can be seen from Fig. 2, for the less familiar type of crack branching (marked by squares), the turning angles are almost 90 degree. The procedure of the appearance of such branches is also different to the common ones. The growth of the leading crack slowed down firstly, and a surface secondary crack was observed away from the leading crack tip, both of the leading crack and the secondary crack kept growing for several loading cycles and finally they linked up to each other. Fig. 3 and Fig. 4 give the observed crack branching procedures in test type I and III, respectively, which indicate exactly the same manner. F