Reliability-Based Assessment of the Residual Collapse Strength for Damaged Tubes Scanned with Ultrasonic Log

Abstract

Abstract This work presents a casing integrity study that begins with data collected from an ultrasonic log inspection and progresses to the estimation of the residual collapse strength of damaged tubes, using a probabilistic-based Finite Element (FE) modeling. The goal is to provide a data-driven enhanced assessment of the structural integrity of tubular components in a high complexity and risk environment, as in the case of the Brazilian pre-salt region. Damage identification employs a technique that seeks an ellipse geometry resembling the probable intact shape of the casing's inner wall. Then, the residual collapse strength is estimated by using physically and geometrically nonlinear FE modeling. The probabilistic analysis is carried out using the First Order Reliability Method (FORM), considering the FE modeling limit state and random variables associated with material and geometry of the tubulars, as well as damage parameters such as maximum depth and position. Geometries of ovalized and eccentric cross sections, associated with multiple damages at various intensities and positions are evaluated. The nonlinear FE modeling allowed the capture of different collapse modes of the element, depending on its slenderness and damage configuration. Through the probabilistic approach, it became possible to account for inherent uncertainties associated with different design and damage parameters, enabling the calculation of the probability of element failure. In the conducted case study, the loads were set equal to the collapse pressures calculated with design equations from the literature. As the obtained probability of failure did not meet a pre-defined target, a finding procedure was implemented to achieve an adequate collapse pressure for design purposes. Although this procedure is computationally expensive and only two critical cross section were assessed, it mitigates many simplifications commonly observed in other works, and the results can significantly contribute to casing design and intervention plans.