Application Fuzzy Logic and Expert Elicitation for Quantitative Offshore Well Integrity Data Collection

Abstract The risk of uncontrolled hydrocarbon releases in oil & gas wells is mitigated through well barrier elements (such as casings, Xmas trees and packers), which are physical components capable of holding the undesired flow of reservoir fluids. Along with operational and management measures, the well barrier elements ensure the well integrity. Typically, these elements compose well barrier sets and, at least, two tested well barrier sets should be in place during the entire life cycle. Currently, limited quantitative failure data regarding well barrier elements exists, and this fact limits a migration from the qualitative approach currently in place to a quantitative one. There are several possible sources of quantitative failure data, such as analysis of historical data, manufacturer’s data, use of commercial databases, and expert elicitation. This latter approach is based on the operational experience of experts and is adequate when there are few existing data from the other sources and limited resources (including time) to collect the data empirically. This paper presents the application of an expert elicitation methodology for collection of well integrity failure data. This methodology is based on the application of fuzzy set theory and allows aggregating the opinions of different experts within a group in an attempt to reach consensus. The results include a sample of the obtained data and comparison with the results from commercial databases. The collected data serves as input for reliability analysis models, which allow predicting the probability of failure/availability of well barriers along the life cycle.