NESC Develops Method for Estimating Risk When Reducing NDE

Why It Matters

It gives NASA programs a data‑driven way to cut inspection costs and schedule delays while preserving safety, influencing future hardware certification practices.

Key Takeaways

• Probabilistic model quantifies risk of NDE descoping
• Historical data gave 0.04% upper bound flaw probability
• Minimum 5,000 inspections required for acceptable risk
• Method yields 3.4‑5 nines reliability depending on sample size
• Assumes time‑invariant process control and part similitude

Pulse Analysis

Nondestructive evaluation (NDE) is a cornerstone of NASA’s fracture‑control strategy, but its implementation can add significant cost and schedule pressure to spaceflight programs. NASA‑STD‑5019A mandates a combined approach of design, analysis, testing, and NDE to ensure that fracture‑critical parts meet stringent damage‑tolerance requirements. While the standard treats flaw detection deterministically, program managers have long questioned whether selective reduction—or descoping—of inspections could preserve safety while freeing resources. The challenge has been the lack of a quantitative framework that translates historical inspection outcomes into a clear risk metric for decision‑makers.

The NASA Engineering and Safety Center (NESC) answered that gap by developing a probabilistic risk‑assessment method that leverages existing NDE databases. Using 33,630 bolt‑hole inspections collected over three years, the team identified six crack‑like findings and calculated a 95 % confidence upper bound of 0.04 % for the probability that a detectable flaw exists. Assuming any flaw larger than the critical initial flaw size would cause failure, the resulting structural reliability is about 0.9996, or 3.4 “nines.” Sensitivity studies showed that expanding the dataset to 100,000 inspections modestly improves reliability to 3.5 nines, while a minimum of 5,000 inspections is needed to keep risk within acceptable limits.

The NESC methodology offers NASA programs a data‑driven lever to justify NDE descoping without eroding the safety envelope, potentially reducing hardware costs and accelerating schedules. Its reliance on time‑invariant process control and part‑similitude ensures that historical rates remain predictive, but it also demands rigorous monitoring and justification when aggregating data across different components. If adopted broadly, the approach could influence commercial aerospace and defense sectors that face similar trade‑offs between inspection intensity and program cadence. Future work will focus on integrating the model with the NASA Fracture Control Board’s broader certification workflow and expanding the database to cover diverse material families.