PSI - Issue 19

Shota Hasunuma et al. / Procedia Structural Integrity 19 (2019) 194–203 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

198

5

Fig. 6(c) and (d) show grain reference orientation deviation(GROD) distributions. Fig. 6(d) is low magnification image of No.1 but observation area of Fig.6(c) and (d) are different. The GROD is a local misorientation parameter related to plastic strain; Sakakibara et al, (2010). GROD of No.1 specimen was larger than that of No.4 at the same distance from surface. Therefore, plastic strain of No.1 is larger than that of No.4. 3. Effect of the machined surface layer on the low cycle fatigue Low cycle fatigue tests were performed using the No.1 and 4 specimens. The testing method was similar to that used in our previous study; Hasunuma et al., (2014). Low cycle fatigue tests were performed by an electrohydraulic testing machine with a load capacity of 200 kN (Shimadzu, EHF-EB100kN) in laboratory air at 24  C. A displacement gage with G.L. = 11.5 mm was attached to measure the true strain,  . Experiments were performed under control of the true strain with an approximate strain rate of 0.2% per second. Fatigue tests were carried out under a strain range of  = 1 %. A replication technique was used to observe crack initiation and growth on the specimen surface. Cellulose acetate films were attached to the specimen surface under fully unloaded conditions with tensile and compressive residual strains. Comparing these films, the opening and closing behavior of cracks was apparent, and cracks could thus be distinguished from scratches. In addition, to separate the effects of variation in the surface shape and variation in the material property, fatigue tests were performed for specimens whose machined surface layers were removed by polishing. Two types of polishing were carried out for specimens machined for No.1 specimen. One of them was paper polishing that removed 60  m of the specimen; the polishing removed the variation in the surface shape, the fine grained layer and a part of the plastic deformed layer but retained the plastic deformed layer. This specimen is called the ground (GR) specimen in the present paper. The other one is paper polishing followed by electropolishing (EP) that removed all of the machined surface layer. This specimen is called the EP specimen in the present paper. Fig. 7 shows the true stress and true strain relationships obtained in the fatigue tests for the first cycle of EP specimen. This figure reveals that plastic strain occurred. It is thus suggested that residual stress was released. In addition, stress range at number of cycle N = 1000 was larger than that at N = 1; i.e., cyclic hardening occurred. Stress range  at N = 2000 was similar to that at N = 1000. Therefore, cyclic hardening was saturated. The saturated  was 602 MPa. Fig. 8 is a diagram of the fatigue life showing  , and the number of cycles to failure, N f . The relationship between  and N f of previous study is also shown in Fig.8; Ozeki et al, (2013). Fatigue tests by Ozeki et al. were performed for hourglass type specimen so their result was shorter than that of this study; Kamaya et al., (2010). The fatigue life of the GR specimen, which had a plastic deformed layer, is similar to that of the EP specimen, which had no machined surface layer. The fatigue life of the No.4-AM specimen was similar to that of the specimen without a machined surface layer. However, the fatigue life reduced by about 30% when fatigue test was performed for No.1 AM specimen.

2 3 4 5

400

EP  = 1 %

No.1-AM No.1-GR No.1-EP No.4-AM

10 3 Number of cycles to failure N f ,cycles Strain range  % Best fit curve(Ozeki et al.) Ozeki et al. 1 0.5 10 4

200

-200 True stress  , MPa 0

1 cycles 1000cycles 2000 cycles

-0.75 -0.5 -0.25 0 0.25 0.5 0.75 -400

True strain  , %

Fig.7 Hysteresis curve of No.1-EP under  = 1 %.

Fig.8 S-N curve of SUS316L.