PSI - Issue 14

Vamsi Krishna Rentala et al. / Procedia Structural Integrity 14 (2019) 597–604 Vamsi Krishna Rentala et al. / Structural Integrity Procedia 00 (2018) 000 – 000

600 4

BB W P S N Q

  

   W f a

K

. 3/ 2



(7)

Q

where

   

   

2

W a

W a

3/ 2 W a

W a

  

1    

  

  

1.99

2.15 3.93 

2.7    

   

W a

   W f a

   

3

.

(8)

W a

W a

   

  

   

  

2 1 2 1

2. Experimental Procedure

In order to perform the maximum flaw size (Type 1) and sum of flaw sizes (Type 2) approaches for POD analysis of multiple cracks at a site, experimental NDT inspection data generated by the current authors in their previous study was utilized (Vamsi et al. (2017)). Fluorescent penetrant and eddy current inspection data in the form of HIT/MISS was obtained from 22 samples. These samples provided 40 inspection sites with a total of 69 fatigue cracks. This data was segregated according to the number of flaws at an inspection site and the corresponding NDT inspection response at that site for applying the POD approaches. Further, for a site containing multiple cracks, the NDT outcome was identified and the new NDT response was generated for each individual crack at that site according to the 2 POD approaches. For example, in maximum flaw size approach, if the NDT response is HIT for a site containing multiple cracks, the maximum flaw size present at that site is considered to be HIT and the remaining flaws are marked as MISS. If the NDT response was a MISS indication at a site containing multiple cracks, all the individual cracks at that site are also considered to be MISS. In sum of flaw sizes approach, if the NDT response is HIT for a site containing multiple cracks, the cumulative sum of flaws is considered to be HIT whereas for a MISS indication, the cumulative sum of flaws are considered to be missed. In the current study, the FPI inspection data was processed with both types of POD approaches whereas for ECI data only Type 2 was applied. Along with Type 1, Type 2 was also applied for FPI data because it was assumed that the FPI indication after a prolonged dwell time for developer indication will actually results in the cumulative sum of penetrant from all the cracks at that site. Further, the ECI inspection data was only processed with Type 2 approach as the ECI response is always the cumulative effect of all the cracks at a site and is difficult to resolve the response for each individual flaw at a site containing multiple cracks close to each other. In addition, the maximum flaw size approach was slightly modified and applied on the FPI inspection data for POD analysis. In the modified Type 1 approach, if the FPI outcome is HIT for a site containing multiple cracks, the maximum flaw at a site containing multiple cracks was marked as HIT and the remaining flaws are not considered in POD analysis. Similarly for the MISS indication at a site containing multiple cracks, the maximum flaw at a site containing multiple cracks was marked as MISS and the remaining flaws are not considered in POD analysis. All the FPI and ECI HIT/MISS data generated from all the three types i.e., Type 1, Modified Type 1 and Type 2 of POD approaches are processed using the standard log-odds distribution model for obtaining the POD curve. In addition, the 95 % lower confidence curves were also plotted in order to identify the a 90/95 value of all these techniques with different POD approaches. Further, the obtained a 90/95 values of FPI and ECI techniques from the three types of POD approaches were substituted for the initial crack size values in the remnant life calculations of DFM approaches. As a thumb rule, the critical crack size was considered as 50% cross sectional thickness of the specimen used for fatigue testing. Moreover, the material constants C and n in the remnant life equation were obtained from the fatigue crack growth rate test of a compact tension (CT) specimen (ASTM E647-15e1 (2015)). C and n are the intercept and slope of the Paris curve. The specimen geometric constant f(a/w) was calculated for the 3-point bend specimen for an initial crack size equal to a 90/95 using Equation 8 (ASTM E399-12e1 (2012)). After obtaining the values of all variables and