Casing collapse capacity was identified by Equinor as a critical operational parameter on one of its fields in production. This led to re-evaluation and detailed studies of the overall well design, specifically the production casing’s collapse capacity, together with consequence and risk evaluations in case of a potential casing failure. As an important and useful input to the risk evaluations, the present paper presents a structural reliability analysis for casing collapse.
Initially, the casing collapse capacity was evaluated using API TR 5C3 / ISO 10400 , with insufficient capacity being documented. In order to investigate further, physical material testing and collapse testing were performed. Two kinds of collapse tests have been performed: i) tests of unsupported pipe and ii) test of pipes with external support from the cement and formation surrounding the pipe. While a paper from 2018 (OMAE2018-78767) considered casings without external support, the present paper pays attention towards supported pipes. Five collapse tests have been performed where test lengths of the 9 5/8” casing were installed inside a thick-walled pipe that simulates the support. A small gap leaves an annulus between the casing and the supporting pipe, allowing a controlled pressure to increase until collapse. The tests have been simulated by finite element analyses. Good correspondence was obtained, providing confidence that FE simulations can be used to predict the collapse capacity of supported pipes.
While the tests were only performed for an idealized case with support around the whole circumference, a large number of FE simulations have been carried out for different combinations of support conditions together with variations in pipe ovality and internal wear from drilling. Ideally, the space between the casing and the rock formation is filled by cement. However, in practice there may be channels where there is no cement, likely to occur if the casing is eccentric in the well bore during cementing. These results from these FE simulations have been used to generate a response surface. Subsequent structural reliability analyses have been performed, in which well specific uncertainty associated with the above parameters is considered. Measurements and logging are used to minimize the uncertainty in these inputs and thereby leading to a reduction in the calculated failure probability. The probability of casing collapse is calculated conditional on different magnitude of the differential pressure of the pipe.
By using SRA the potential over-conservatism in the conventional deterministic analysis is avoided. The SRA results were used to assist in the risk evaluation resulting in an allowance for continued production on existing wells.