PSI - Issue 13

Wei Song et al. / Procedia Structural Integrity 13 (2018) 2227–2232 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

2229

3

transducers were attached to the loading plates in the both side of weld toes (up and down). The gauge locations away from the weld toes are about 10m. Therefore, the strain-controlled set-up were not influenced by the stress concentration from the weld toe. Finally, the low cycle fatigue tests were performed under different strain amplitudes with a 0.1 strain ratio. On the other hand, high cycle fatigue experiments of LCWJs were conducted on a 250KN electro-hydraulic servo testing system MTS 809 with a loading-control condition. The illustration of LCWJ geometry details is shown in Fig. 2. It includes the plate thickness, attached plate thickness, penetration length, weld length and notch radius in Fig. 2(a). The different penetration lengths at weld root are also cut by the wire electrode machine after forming a 0.7mm hole in the middle of plate by spark-erosion drilling machine. The variation of penetration length are shown in Fig. 2(b).

(a)

(b)

ρ

θ Weld toe

L

L

h

L

2a

p/2

L

t

t

t

2α=135º

t

L

L

L

ρ Weld root

h

h

h

Fig. 2. Geometry illustration of load-carrying cruciform joint of LCF tests, (a) geometry details; (b) different penetration ratio.

3. Experimental results 3.1. material mechanical properties

Monotonic stress-strain curves of 10CrNi3Mov high strength steel and corresponding welds were shown in Fig. 4. The mechanical properties of these materials under monotonic tension loading were summarized in Table 1. As the metallurgical composition of undermatched weldments is similar with the base metal, the monotonic and fatigue mechanical properties of these metals are also consistence. For undermatched weldments, the mismatched ratio is 0.7 according to these monotonic tension stress and strain curves. Similarly, the cyclic Ramberg-Osgood curves of base metal and the undermatched weldments at the middle of fatigue life for different strain amplitudes are compared in Fig. 4. The cycle fatigue strengths of these metals, which are obtained from the cyclic R-O curves, are demonstrated in Table 1. The cycle Ramberg-Osgood curves of half life are all lower than the monotonic curves due to the material soften behavior. The cycle fatigue strength mismatched ratio is about 0.8.

Table 1. Mechanical properties of test materials.

Young’s modulus (GPa)

Poisson’s ratio

Steel

Kv (J) -20 ° C

Yield strength (MPa)

Cyclic Yield strength (MPa)

Tensile strength (MPa)

10CrNi3MoV

693

521

741

205

0.3

280

U-Welds

498

421

559

195

0.3

260

3.2. Low cycle fatigue failure locations of CLWJ Table 1. An example of a table. Fig. 5 shows the failure locations under low cycle fatigue considering penetration ratio and material mismatch effects. For the large penetration ratio, most of LCF tests specimens initiates from weld