In reviewing the current API 5C3 process for designing downhole tubulars against collapse under external pressure, it has been found that the API equations do not reflect a uniform 0.5% probability of failure as is commonly believed. By applying the API equations to an extensive database of collapse test results collected for this study, the API approach is demonstrated to result in designs that are unconservative for thicker tubulars and over-conservative for thinner tubulars.
After a careful review of the fundamental factors affecting tubular collapse, a risk-based collapse design equation is proposed and shown to provide a more uniform margin of safety. This equation contains an adjustable design factor which allows the engineer to prescribe the probability of failure to be reflected by the equation for a given application. For situations with very grave consequences of failure, a lower probability of failure can be used, while for cases involving less severe failure consequences, a larger probability of failure may be tolerable. The relationship between the design factor and the probability of failure is established through a rigorous reliability analysis based on statistics derived from the collapse test database and from regular measurements taken by steel mills. This same reliability analysis is also applied to the current API 5C3 criteria to demonstrate the non-uniform probability of failure that results from the use of these equations.