NASA’s Engineering and Safety Center (NESC) has developed a probabilistic analysis method to help estimate the risk associated with reducing or eliminating nondestructive evaluation (NDE) inspections on certain spaceflight hardware, addressing a question with direct relevance to fracture control and damage tolerance verification.
The work examines whether an equivalent risk posture could, in some cases, be achieved without relying fully on the traditional approach defined under NASA-STD-5019A, the agency’s technical standard for establishing fracture control plans for spaceflight systems. Under that framework, fracture control depends on design, analysis, testing, NDE and tracking of fracture-critical parts to verify damage tolerance and reduce the risk of catastrophic failure.
In the existing 5019A framework, damage smaller than NDE detection capability is assumed to exist, and the part must still be shown through analysis or testing to survive the required service life. In practical terms, NDE serves to screen out flaws that could otherwise lead to failure. NESC’s assessment explored whether, in some situations, descoping NDE from the damage tolerance verification process could still maintain the required level of safety.
The objective of the assessment was to develop a method that would allow NASA programs and projects to estimate the risk associated with descoping NDE requirements for single-wrought materials. The effort included the use of historical inspection data, sensitivity studies and identification of the minimum supporting data needed for approving a descoping request.
NESC framed damage tolerance as a probabilistic problem built around three factors: the probability that a flaw of a given size exists, the probability that the flaw is missed by NDE, and the probability that the flaw results in failure if it exists and is missed. These factors are combined into a joint failure probability model.
The analysis notes that conventional damage tolerance is generally treated as deterministic, with an NDE detection threshold and a binary pass-fail outcome. However, the probabilistic method focuses on estimating the likelihood of flaws above critical thresholds and how that risk changes if NDE is reduced.
To demonstrate the framework, NESC used a historical NDE database containing 33,630 bolt-hole inspections conducted over a three-year period. Six crack-like features were identified by NDE. After accounting for sample-size uncertainty, the analysis produced a 95% confidence upper bound of 0.04% for the probability of a detectable flaw existing in each hole.
Using a conservative assumption that any flaw exceeding the critical initial flaw size would lead to structural failure, NESC estimated that this probability corresponds to a structural reliability of about 0.9996, or approximately 3.4 “nines.”
The analysis also found that increasing the inspection dataset to 100,000 inspections would raise the reliability only marginally to 3.5 “nines.” At the observed NDE rejection rate, four nines of reliability would not be achievable even with infinite samples and zero uncertainty. Based on the case study, NESC indicated that a minimum sample size of 5,000 inspections would be needed at an NDE rejection rate of 0.04% to avoid a significant increase in risk relative to the baseline risk posture for NASA Human Spaceflight Programs.
The methodology is built on several assumptions. One is the requirement for time-invariant process control so that historical inspection data remain predictive of future probabilities after any NDE descope. Another is that combining data across multiple parts to increase sample size would require justification through qualitative and quantitative assessments of similitude.
NESC said the methodology is intended to serve as one component of a broader fracture control evaluation by the NASA Fracture Control Board and the responsible Technical Authority, offering a structured way to assess risk when considering reductions in NDE within damage tolerance verification.
Reference: https://www.nasa.gov/general/nesc-develops-method-for-estimating-risk-when-reducing-nde/
