RELIABILITY-BASED ANALYSIS OF BURST STRENGTH OF CASING TUBES: THE INFLUENCE OF CORRELATION BETWEEN DESIGN VARIABLES

Resumo

In structural design normative codes, safety factors are used to reduce the value of the
resistance and the probability of failure of the structure due to project uncertainties. Regarding oil
exploration, there is a concern about the risk of oil well failure, considering the new challenges of this
sector, especially in offshore exploration. Even applying safety factors, these uncertainties leads to a
non-zero probability that the structural system does not meet the requirements for which it was
designed. In order to improve the decision-making process in structural safety analysis, the structural
reliability methods arise with the possibility of calculating the probability of failure of the structure,
taking into account the inherent uncertainties of the design variables. However, in order to assure the
success of the reliability analysis, it is necessary to have the statistical description of the design
variables and a mathematical model to determine the failure functions of interest. In the context of oil
well design, the API 5C3 [1] recommends the use of data with independent random variables (r.v.) in
the reliability analysis of well casing tubes, because this consideration simplifies the problem and is
beneficial for safety. However, the present work aims to analyze the influence of the use of correlation
between the r.v. in the results obtained through structural reliability methods. The limit state of ductile
rupture under internal pressure is defined, in which the tube strength is defined by the Klever-Stewart
[2] model, as determined by API 5C3 [1]. The First Order Reliability Method (FORM) and the Monte
Carlo simulation are applied for extreme correlation values of a real sample of casing tubes
manufactured in industry. Finally, the results obtained are discussed, in order to evaluate the
importance of the use of correlation between r.v. on the casing integrity analysis.