Verification you can measure. On public data.
The framework is only as good as the evidence its requirements produce. These are measured results from applying SCL verification methods to a public dataset, NASA C-MAPSS turbofan degradation. Every number is reproducible from the open repository. The methods that failed their own subject are kept in view, because a verification bench that cannot fail is not a verification bench. Methods and thresholds are those of framework v3.7.
The failures these
requirements target.
Each requirement area exists because a class of AI failure has already happened in the field. These are documented public cases, each mapped to the area written to catch that failure mode. The mapping names the target. It does not claim any outcome would have changed.
The same classes of failure, run as pass and fail measurements on public data. NASA C-MAPSS FD001, methods and thresholds of framework v3.7.
Splitting the data by row instead of by engine lets the same asset appear in both training and test. Measured accuracy rises from 80.33% to 83.05%. The gain is not skill. It is memorization the evaluation failed to prevent. A pass threshold set on the leaky number would certify a system that the honest split does not support.
Naive drift designs fired false alarms on healthy data, population indices past 7.8 where the real signal was noise. Only after declaring regime, per-asset baseline, and window and bin design did the monitor become trustworthy: it alerts at 0.58σ of injected bias and inhibits at 1.17σ. Left unmonitored, the classifier's false-fault rate climbs to 0.70 under full bias.
The popular softmax-confidence detector scores out-of-distribution inputs worse than a coin flip. A Mahalanobis detector recovers gross regime shift perfectly (AUROC 1.0) but catches the subtle near-boundary fan fault only 44% of the time at a 2% false-positive budget. No single OOD method passed every regime. That result is reported as a denial, not smoothed over.
Out of the box the model's stated confidence was off by 0.16 expected calibration error, a clear fail. A single temperature parameter fitted on validation data brought it to 0.030, a pass. The requirement is not that the model is born calibrated. It is that calibration is measured, corrected, and verified before the confidence is trusted.
A bench that cannot fail its subject is not evidence.
Two of the four results above are failures. The softmax detector performs worse than chance. The out-of-the-box calibration does not pass. They are kept because they are the point. A verification method that certifies whatever it is pointed at proves nothing. These are held to the same pass and fail criteria a real assessment would apply, and they are allowed to fail.
Open pipeline. Open data.
The full pipeline, configuration, seeds, and result files are public. Every figure on this page traces to a JSON artifact in the repository, produced under a pre-declared configuration. The framework these methods verify against is openly published under a citable DOI.
This is a demonstration of SCL verification methods on a public dataset. It is not an issued certification, and it makes no safety claim about any real turbofan engine. It exists to show that the framework's requirements resolve to measurements that can pass, and can fail.