283

Pipeline operators use the verification and validation requirements of API Standard 1163, In-Line Inspection Systems Qualifications, to ensure quality in-line inspections (ILIs). Verification confirms an ILI was conducted according to plan, procedures, and processes and that the inspection conditions are consistent with those used to establish the ILI Performance Specification. Validation evaluates the accuracy of the reported anomaly types and characteristics (depth, length, width, etc.) and provides assurance the ILI met its Performance Specification. 

Many of the guidelines (i.e., “should” statements) and requirements (i.e., “shall” statements) in API 1163 require the pipeline operator to establish acceptance limits based on pipeline industry experience. Often, API 1163 does not provide guidance on how to establish the acceptance limits. PRCI has developed a guidance document and spreadsheet for conducting the Level 2 and 3 validations and establishing acceptance limits in accordance with the 3rd Edition of API 1163. Both API 1163 and the PRCI documents are widely used in industry to verify and validate metal-loss ILIs, but questions have been raised about the ability to use them for crack-detection inspections. These questions primarily concern the ability to use field non-destructive evaluation (NDE) techniques, which have accuracies similar to or worse than stated ILI Performance Specifications for cracks, to validate the inspections. Simply put, how can an operator validate an inspection when field validation measurements are questionable?

This paper assesses the ability to use API 1163 for crack-detection ILIs including, but not limited to the following: 

  • API 1163 and the PRCI guidance report/spreadsheet provide a method for incorporating field NDE uncertainty into validation calculations. However, large field NDE uncertainties increase the risk of accepting an ILI’s accuracy when, in fact, the ILI did not meet its Performance Specification. This is called a Type 2 statistical error.
  • Monte Carlo simulations were used to evaluate the impact of typical field NDE uncertainties. The simulations demonstrate that using API 1163 and the PRCI spreadsheet for conducting Level 2 and Level 3 validations can be nonconservative if applied without correction factors.  
  • The Monte Carlo simulations provide a path forward for validating a crack-detection ILI. The implications of using such simulations are identified and discussed, and broad guidelines are given.

 Several case studies are included to demonstrate the impact of field NDE uncertainties and their impacts on API ١١٦٣ validations.

KEYWORDS: ILI Analysis, ILI Verification and Validation

Read/Download PaperGradeAvailable after the event