PSI - Issue 14

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

2

598 This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/) Selection and peer-review under responsibility of the SICE 2018 organizers. Vamsi Krishna Rentala et al. / Procedia Structural Integrity 14 (2019) 597–604

Keywords: a 90/95 ; multiple cracks; POD approaches; remnant life calculations

1. Introduction

Damage Tolerance (DT) design philosophy of aero-engines assumes that every component contains an initial crack of certain size which grows during the service. The component under operation will fail/fracture once the initial crack size reaches the critical crack size of the component (Dieter (1986)). The number of remaining fatigue cycles of the component can be calculated when the initial crack size and the critical crack size of the component are known (Koul et al. (1990)). In general, the critical crack size of a component is usually calculated using fracture mechanics approach. However, the initial crack size of the component can only be obtained using non-destructive testing (NDT) techniques as the serviceability of the component should not be jeopardized. In this case, the designer is completely dependent on the NDT inspection result. In general, the NDT outcome consists mainly of a defect being detected or undetected. In case of NDT technique resulting in a false call i.e., indicating a crack when there is no crack, early withdrawal of component results in wastage of remnant life and cost. Similarly, no NDT response for a component with a crack would result in huge loss of human and property cost. Hence, it is extremely essential to estimate the reliability of the NDT techniques used. Probability of detection (POD), being a standard metric for measuring the NDT reliability yields the a 90/95 (detecting a flaw with 90 % probability and 95 % confidence) value (Annis (2009)). This a 90/95 value can be obtained from different POD procedures depending on the type of NDT techniques used such as qualitative (visual inspection, penetrant inspection, Radiography, etc.) or quantitative (Eddy Current inspection, ultrasonic inspection, etc.). The two most common POD procedures are HIT (defect detected)/MISS (defect undetected) (Vamsi et al. (2017), Vamsi et al. (2018a)) and â (signal response) vs. a (flaw size) (Vamsi et al. (2018b)) for qualitative and quantitative NDT inspection techniques, respectively. Irrespective of the POD procedures adopted, the a 90/95 values obtained are considered as the initial crack sizes detected by the NDT techniques for use in remnant life calculations. Hence, it is essential to estimate the a 90/95 value to the closest accuracy. However, the NDT inspection data at a site containing multiple cracks results in ambiguity of POD approaches to be adopted for the estimation of a 90/95 values. Amongst several approaches attempted by researchers, the most common of all happens to be the maximum flaw size approach and the sum of flaw sizes approach (Spencer (2006)). In those cases, the a 90/95 values obtained will be different for different approach for the same ND T inspection data. To the best of the author’s knowledge, the physical significance of the a 90/95 value and the different HIT/MISS approaches on the remnant life calculations of aero-engine components was not discussed in the literature. Therefore, in the current study, physical manifestation of a 90/95 obtained from maximum flaw size and sum of flaw sizes approaches for inspection of natural fatigue cracks in a nickel based superalloy using fluorescent penetrant (FPI) and eddy current inspection (ECI) techniques on remnant life calculations was attempted.

Nomenclature a 90/95

detecting a flaw with 90% probability and 95% confidence

P i a i

probability of detection for defect i

size of defect i α & β constant parameters that define the regression curve m(a j ) Maximum flaw size at site "j" s(a j ) Sum of flaw sizes at site "j" a 1 the dysfunction crack size which is equal to 2/3 a d

initial crack size in the component which can be detected by the NDT technique

rd of the critical crack size of the component

C & n material constants K Q

stress intensity factor

P Q

load applied, N