A groove-position dependent probabilistic model for collapse resistance of worn casing tubulars: applications to well reuse

This paper presents a comprehensive probabilistic analysis of the collapse resistance of worn casing tubulars for reuse applications, such as Carbon Capture, Utilization, and Storage (CCUS) projects. The study focuses on failure due to collapse, a critical loading condition, particularly in depressurization scenarios. A collapse prediction model, based on data obtained from Finite Element Analysis (FEA) and considering a single wear groove, is developed and applied to probabilistic analyses. The casing tubulars are installed in highly complex environments and often exposed to multiple loading and wear conditions. Evaluating the collapse strength of tubulars subject to mechanical wear requires consideration of key model variables, e.g., wear depth, wall thickness, outer and tool joint diameters, and yield stress. The position of the wear groove also has a significant influence on the residual strength of tubulars, as it may occur at any position of the cross-section, but is not usually considered in prediction models. The proposed model also accounts for manufacturing imperfections, such as cross-section ovality and eccentricity. The methodology was divided into three main stages: a deterministic analysis introducing a new collapse pressure prediction model; a parametric probabilistic analysis of casing tubulars; and a case study of a well in the Campos Basin, eastern Brazil, hypothetically assessed for reuse in CCUS. In this context, the first stage proposes a new prediction model that considers the main variables affecting collapse failure, including the position of the wear groove. This model is validated against an experimental dataset and compared to: (i) other models from the literature; and (ii) nonlinear plane strain analysis using the finite element method. The proposed equation showed low model uncertainty and standard deviation, with values of 1.005 and 0.060, respectively. The improvement of the proposed model is also shown in tests against a FEA dataset created with different wear groove positions and compared with other models. It is observed that the proposed formulation achieves high accuracy and low dispersion (2.93%), whereas models from the literature may yield collapse pressures up to 21.17% lower than those from the FEA. In the second stage, the probabilistic analysis focuses on a parametric study of unworn and worn tubulars, evaluating the failure probabilities for different slenderness and wear to thickness ratios. Since the tool joint’s passage position is uncontrolled relative to the ovality axis, wear can occur at any position on the tube’s internal wall. As a result, the wear groove position is treated as a random variable ( X ) with a uniformly distributed probability, which is one of the main advantages of the proposed model. The probabilistic analysis indicated that wear groove position has a significant impact on failure probabilities, as reflected in the importance factors ( IF ) of the X considered. Finally, in the third stage, the proposed model was applied to a real-world well, assumed to be converted for reuse in a CCUS injection scenario. The failure probability was determined along the depth for an intermediate casing string for several prediction models. For the critical depth, the proposed model yielded the lowest failure probability among them, with a value around 1.36%. It is important to highlight that, at the critical position, the deterministic analysis yielded a safety factor of 0.965, indicating the unsuitability of the casing tubular for the collapse scenario evaluated.