Existing single-string analysis methods may be inadequate for more difficult casing design problems, such as annular fluid heat-up and platform wellhead thermal growth, which require a multistring (or global) analysis of the whole well system. This paper presents a method for such an analysis and describes a finite-element formulation developed to implement it. The formulation is fully general and is applicable to a wide range of casing-/tubing-design problems.
IntroductionFluid heat-up pressures in trapped annuli have long been a concern of the petroleum industry and have been the subject of several recent studies.1-5 A general analysis method was first developed for BP Exploration, as part of a study into trapped annular stresses in subsea production wells. 1 ,2 It had a far wider scope than just solution of heat-up problems and proved to be a significant development for four main reasons.1. It is a multi string method and therefore permits analysis of problems like annular fluid heat-up, which was not previously possible with existing single-string techniques.2. Multistring analysis also allows quantitative risk analysis (QRA) of the whole well system by use of structural reliability methods. This opens up a whole new area, probabilistic analysis, which is proving to be of key importance to future developments in casing/tubing design. 6 ,7 3. The solution method uses a finite-element formulation for the axial response; therefore, it does not suffer from the applicability problems of the closed-form approach (discussed later).4. It reduces the whole analysis to the two fundamental equations that govern the behavior of the well system. It can be shown that this general approach, using a finite-element implementation, permits analysis of completely general design problems, even for the most complex aspects of the system response. Furthermore, any future theoretical developments can probably be included in this general finite-element treatment.It may be helpful to discuss the last two points briefly before going further. Currently, many equations exist for casing and tubing design, which makes the subject look more difficult than it really is. The reason for this (apparent) complexity lies not in the theory itself, but in our implementation of it. As an industry, we have largely chosen to implement the theory analytically as closed-form solutions for simple single-string cases. For example, we have closed-form solutions for helical buckling, with and without friction; bending resulting from doglegs; and pressure loads on combination strings. Unfortunately, these all are separate cases. In real well systems, we often have frictional buckling plus multiple doglegs plus mUltiple section changes, a case far too complex to solve in closed form. For real design, therefore, closed-form solutions are often inadequate and we need a more general approach that can incorporate all possible effects, even for complex well models.This paper shows that if we use a finite-element implementation for the axial response, the struct...