Edit Article

API 1163 Level 2 Validation is a statistical assessment process used to evaluate the sizing performance of an Inline Inspection (ILI) tool by comparing reported anomaly dimensions against field verification measurements obtained through excavation and NDT activities.

The objective of the validation is not only to determine whether the ILI tool meets its expected performance but also to understand the accuracy, consistency, and sizing tendencies of the inspection system.

Overall Statistical Assessment

The overall statistical assessment evaluates the performance of the complete validation dataset rather than individual anomalies.

Key statistical parameters include:

Bias

Bias represents the average difference between the ILI measurements and the verification measurements.

For example, a Bias of -2.0 mm means that the tool reports corrosion depths approximately 2 mm shallower than the actual depths measured during field verification.

Deviation

Standard Deviation represents the spread of the sizing errors around the Bias.

Observed Accuracy

Observed Accuracy represents the actual sizing performance demonstrated by the validation dataset and can be compared against the tool's specified performance.

Individual Anomaly Assessment

While the overall statistical assessment evaluates the entire population of anomalies, each anomaly should also be assessed individually.

Combined Accuracy

Combined Accuracy accounts for uncertainties from both the ILI tool and the verification measurements.

Combined Accuracy is typically calculated as:

Combined Accuracy = √(ILI Accuracy² + NDT Accuracy²)

Interpreting Validation Results

A robust validation assessment should consider both acceptance criteria together:

Criterion 1 – Overall Statistical Performance

The overall validation dataset should demonstrate acceptable statistical performance based on:

• Bias
• Standard Deviation
• Observed Accuracy
• Distribution of sizing errors

These parameters indicate whether the tool performs as expected across the entire population.

Criterion 2 – Individual Anomaly Acceptance

Each anomaly should be assessed against its calculated Combined Accuracy.

The percentage of anomalies falling within tolerance provides an additional measure of performance.

A commonly used interpretation is:

Percentage Within Tolerance

> 80% Good - acceptable

70–80% Acceptable with review

60–70% Marginal performance

< 60% Poor agreement

Conclusion

API 1163 Level 2 Validation should not be viewed as a single pass/fail calculation. A meaningful evaluation requires consideration of both the overall statistical behavior of the inspection tool and the acceptance of individual anomalies.