PSI - Issue 42

S. Lindqvist et al. / Procedia Structural Integrity 42 (2022) 42–49 Author name / Structural Integrity Procedia 00 (2019) 000–000

47

6

cooling thermal transient. However, during the cooling transient the splitting behavior is observed close to but below 100 °C [5]. During isothermal testing at 100 °C no splitting is observed. Thus, the thermal transient can promote the splitting phenomena due to the thermal gradient in the specimen and the resulting thermal stresses.

300

50

250

40

200

Temperature [  C]

30

25  C 100  C 200  C 300  C Thermal transient Temperature

150

Load [kN]

20

100

10

50

0

0

0

2

4

6

8

COD [mm]

Figure 5. Cooling started at load = 33 kN and CMOD = 0.6 mm.

3.4. The effect of thermal transient on the J-R curve For each specimen fractured during the cooling thermal transient, the basic method was used for determining the final J-integral and the crack length. The resulting J-R curve shows that the curve is intermediate to the curve at 200 °C with lowest toughness and 300 °C with highest toughness. This is what is expected as the specimen is fractured during a cooling transient. As the specimen is cooling, at 200 °C the crack growth is easier than when the specimen cools further and reaches room temperature. The result indicates that in structural integrity analysis by applying the lowest toughness properties in a specific temperature range a conservative estimate is obtained of the upper shelf toughness. However, using the higher toughness values for analyzing the toughness during a thermal transient can result in an unconservative estimate. Thus, if no other information is available one should use the toughness at the temperature giving the lowest value. Yet, application of the lower toughness values can result in an over conservative estimate of the fracture toughness. The results indicate that as long as the J-integral stays in the vicinity of the initiation stage the effect of the transient is insignificant. Also the resulting J-R curve for the transient condition is higher than the J-R curve for 200 °C. Furthermore, the behaviour of the J-R curve during cooling was investigated with the normalization method for two specimens fulfilling the requirements of the normalization method. The various methods (normalization, basic, compliance) appear to give similar J-integral values during the thermal transient, as was the case for the isothermal results presented in section 2.3, Figures 20 and 21. As the specimen cools through the low toughness region, the J-R curve is flatter, but as the specimen warms and approaches the temperature where the material is tougher, the J-R curve becomes steeper. Figures 22 and 23show that the J-R curve is less steep when the specimen is cooling through the area 250 to 130. Figures 24 and 25 show the crack growth relative to the isothermal J-R curves.