PSI - Issue 19

Shinji Hashimura et al. / Procedia Structural Integrity 19 (2019) 204–213 Author name / Structural Integrity Procedia 00 (2019) 000–000

212

9

significantly larger than A5056 except for Q m =2 kN. Fig. 10 shows relationships between  1 -mean and Q m at zero cycle and ten cycles. In Fig. 10, the ordinate is  1 -mean and the abscissa is Q m . The white circle symbol shows  1 -mean at the root of the incomplete thread for A5056 at 0 cycle. The black circle symbol shows  1 -mean at the root of the incomplete thread for A5056 at 10 cycle. The white square symbol shows  1 -mean at the root of the first thread for A6056 at 0 cycle. The black square symbol shows  1 -mean at the root of the first thread for A6056 at 10 cycle. It can be seen in Fig. 10 that the difference of  1 -mean between Q m =4 kN and 6 kN for A5056 was almost the same in N =0 cycle and N =10 cycles. However the difference of  1 -mean between N =0 cycle and 10 cycles becomes quite larger at Q m  4 kN and  1 -mean at Q m  4 kN become almost the same. The results indicate that the increase of the mean stress  1 -mean at Q m  4 kN gets disappear and the influence of the mean stress clearly decreases. Incidentally the reason that the fatigue strengths decreased until Q m =6 kN despite the mean stress was released over Q m =4 kN is considered the effect of residual stress at root of the first thread because the bolts were made by thread

Fig. 9 Behaviours of the local mean principal stress  1 -mean at the root of the first thread for A6056 bolts.

Fig. 10 The mean stresses of principal stress  1 -mean for A5056 bolts and A6056 bolts at N =0 and 10.