PSI - Issue 18

Mehdi Mokhtarishirazabad et al. / Procedia Structural Integrity 18 (2019) 457–471 M. Mokhtarishirazabad / Structural Integrity Procedia 00 (2019) 000–000

466 10

Fig. 12. Effect of thickness (out-of-plane constraint) and crack length (in-plane constraint) on J Q for plane-sided samples. Unexpected trend in J Q can be attributed to erroneous crack extension measurement by UC method.

Fig. 13. Dependence of critical stress intensity factor on specimen thickness (Adams, Lai and Ferguson 1986).

3.4. Correction of Δ a UC Probably the easiest method for tackling under-prediction of Δ a UC is to use a linear scaling method. This can be applied by multiplying Δ a UC to the ratio of measured Δ a p divided by calculated Δ a UC (Dzugan 2003); J-integral is then calculated based on corrected crack lengths. This method was applied to SENB sample with thickness of 10 mm and a/W = 0.2 and 0.5, with the corresponding graph shown in Fig. 14. Although employing this method resulted in better estimation of final crack extension, it could not address the steeper blunting line of experimental results though. In addition, the value of J Q is still higher for higher constraint condition. The other way to correct Δ a UC is to analyse the effects occurring during the bending of the SENB sample. Fig. 15 shows the sources of errors arising during bending the SENB specimen for measuring the Δ a UC . These factors (Steenkamp 1988; Perez Ipiña and Santarelli 1989; Wallin 2014; Dzugan 2003) include:

Specimen deformation,



 Movement of and friction at the outer rollers,  Crack front curvature,  Roller indentation  Effective thickness and Young’s modulus.