13th International Conference on Fracture June 16–21, 2013, Beijing, China -4- to suppress the tendency for Mode I crack branching, the pre-tightening load S of a bolt was applied to the ends of the cantilevers and vertical compressive load Q was applied to the crack faces using the loading jigs and tightening jigs. The ceramic cylinder diameter (8 mm) was larger than that used in the former setup (1 mm) in order to reduce the plastic deformation of the specimen on the contact face. Moreover, a set of universal joints was used to cancel out the unexpected moment applied to the specimen. The specimen and jigs were designed considering their compatibility with the fatigue crack growth experiment equipment using a “CT (Compact Tension) specimen.” Therefore, the experimental setup shown in Fig. 3 could be used with environmental experiments equipment that is designed for a “CT specimen.” 2.5. Crack length measuring method The crack length during the fatigue crack growth experiment was measured by using the AC potential method [3]. Two electrodes were connected to the ends of the cantilevers. As the crack between the two cantilevers grew, the electrical resistance between these two electrode points increased. Then, the resistance was measured and converted into the crack length. The ratio of the increase in the electric potential (ΔE) caused by the crack growth to the electric potential at the beginning of the experiment (E0) correlated with the crack length. Therefore, the crack length could be measured without interrupting the experiment. The specimen was insulated from the jig using the ceramic cylinders. 3. Derivation of friction between crack faces In order to measure the friction between the fatigue crack faces, two strain gauges (Kyowa Electronic Instruments Co., Ltd., KFG-2N-120-C1) were placed on the specimen surface. Figure 4 shows the strain measurement positions on the specimen. The friction force on the crack face was derived from the load-strain curve over one cycle using the outputs of the strain gauges. Figure 5 shows the basic model for the derivation of the friction, which represents a mass spring model with a rough ground and a mass subjected to a cyclic load P. The load-displacement curve over a cycle is shown in Fig. 5(b). Because the direction of the friction changes over a cycle, the friction can be derived from the hysteresis in the load-displacement curve. However, for the real experiment, the curve became more complicated. When the loading began, the portion of the crack faces at the notch began to slide, and all of the other parts of the crack faces soon followed. As a result, the linear part of the curve marked (i) became nonlinear. Thus, because of the plastic deformation that occurred around the crack tip, the portion of the linear part of the curve marked (ii) also became nonlinear. The same phenomena also occurred at the portions marked (iii) and (iv). Therefore, the load-strain curve over one cycle was estimated to be that is shown in Fig. 6. The friction could still be derived from the load-strain curve. The vertical dashed line shown in Fig. 6 starts from point A and intersects at point B with the extended line of the elastic portion of the curve at the unloading. The length of dashed line AB is related to the friction, where the relation is determined using a finite element method (FEM). Figure 4. Strain gauges placed on specimen
RkJQdWJsaXNoZXIy MjM0NDE=