Section 2.4.1. Initial Flaw Distribution
A measure of initial quality in a component of service hardware
is given by the distribution of initial crack sizes as illustrated in Figure 2.4.1.
For predictions of safety limits, the initial cracks larger than the
nondestructive inspection (NDI) detectability limit are of principal
concern. Current specifications detail
NDI limits and require verification/certification of contractor capability to detect
cracks smaller than the specified NDI limits.
Normally, such certification is demonstrated with curves of the type
shown in Figure 2.4.2. The program of certification for a contractor’s quality control
inspector/inspection techniques allows the USAF to assess the probability and
confidence limits associated with detecting a given crack. Section 3 will present a state-of-the-art
summary of the technology and equipment that supports the establishment of
initial flaws via nondestructive tools.
Figure 2.4.1. Distribution of Initial Crack Size for a Given Type of Crack
(e.g., Radial Cracks Growing from Fastener Holes)

Figure 2.4.2. Certification of NDI Capability
Results generated by the F-4 Independent Review Team (IRT)
provided a method of characterizing the initial flaw population (apparent
initial quality) based on full-scale fatigue test-induced cracking behavior
[Lozano, et al., 1974]. Given the
measurable flaw distribution in a structure at some time subsequent to test
startup, the initial flaw population can be backtracked by analysis. The “back” extrapolation of the flaw
population is conducted using the damage integration package. The process is schematically illustrated in Figure 2.4.3.
Subsequently, the initial flaw distribution established as illustrated
in Figure 2.4.3 can be used to estimate influence of
load factors, mission profiles, and usage changes on the life of service
hardware. The F-4 IRT study also
provided an evaluation of statistical methods for describing the large crack
length extremes for initial flaw distributions established in this manner. The resulting distribution of F-4 initial
cracks is shown in Figure 2.4.4 [Lozano, et al.,
1974; Pinchert, 1976].

Figure 2.4.3. Determining Initial Quality by Back Calculation
Figure 2.4.4. Initial Flaw Distribution for F-4 Based on Back Calculation