Probability-Based Sentencing Criteria for Volumetric In-Line Inspection Data

ASME B31.8S, Figure 7.2.1-1 (referred to as Figure 4 in earlier editions of the Standard) is used by many operators of natural gas transmission pipelines to schedule the remediation of corrosion features found via in-line inspection (ILI). The underlying philosophy of this approach is that wall loss features should be repaired before the calculated failure pressure falls below 110% of the maximum allowable operating pressure (MAOP). ASME B31.8S Figure 7.2.1-1 provides a basis for establishing maximum response times as a function of pipeline operating stress level, based in part on assumed corrosion growth rates. The corrosion rates assumed in the derivation of ASME B31.8S Figure 7.2.1-1 depend on the wall thickness of the pipe and the operating stress level as a percent of SMYS. As documented in PHMSA’s March 17, 2016 Notice of Proposed Rulemaking, the 1.1xMAOP repair criterion that forms the basis of Figure 7.2.2-1 has a demonstrated successful history of use in response management for wall loss ILI data. Despite this successful record, some potential exists for the underlying corrosion growth rate assumptions that are incorporated within that criterion to be non-conservative. Under some circumstances, the underlying corrosion growth rate assumption that is incorporated in Figure 7.2.1-1 can be significantly less than that which is provided in the guidance provided in NACE SP0502 (referenced in Appendix B of ASME B31.8S). Therefore, operators should ideally take measures to verify that the growth rate assumptions incorporated within Figure 7.2.1-1 are appropriate for their circumstances before adopting the scheduled response criteria from that Figure. On the other hand, for the majority of circumstances, it could be demonstrated that the Figure 7.2.1-1 criteria may represent overly-conservative response times, particularly where feature-specific information related to corrosion rates are available, and/or can be inferred from ILI data. A desirable solution would be to employ a response time threshold that utilizes the 1.1xMAOP repair criterion that has been demonstrated to be successful through industry’s widespread adoption of the Figure 7.2.2-1 criteria, along with some basis for incorporating feature-specific corrosion growth rates (from ILI data), and additionally, some basis for accounting for tool measurement error. Techniques for estimating the relative probability of failure (Pf) exist that employ ILI data and account for tool measurement error, model error, and tolerances in pipe dimensions and material properties. The problem to date is that probability targets have not been available for use in conjunction with a Pf analysis. Building on previous work done by Kiefner and Kolovich, this paper derives an approach for expressing Pf targets in terms of the 1.1xMAOP repair criterion adopted by ASME B31.8S, Figure 7.2.1-1. The Pf targets are derived using stochastic modeling, and incorporate probability density functions on tool error for feature depth and length, wall thickness, yield strength, and model error. Using a wide range of pipeline material and design parameters, a relationship for establishing lower-bound Pf targets is developed for broad application.