Marlin Failure Analysis and Redesign: Part 1 - Description of Failure
- D.W. Bradford (BHP Petroleum) | D.G. Fritchie Jr. (BP America) | D.H. Gibson (BP America) | S.W. Gosch (BP America) | P.D. Pattillo (BP America) | J.W. Sharp (BP America) | C.E. Taylor (BP America)
- Document ID
- Society of Petroleum Engineers
- SPE Drilling & Completion
- Publication Date
- June 2004
- Document Type
- Journal Paper
- 104 - 111
- 2004. Society of Petroleum Engineers
- 1.11 Drilling Fluids and Materials, 1.3.2 Subsea Wellheads, 5.3.2 Multiphase Flow, 5.9.1 Gas Hydrates, 4.2 Pipelines, Flowlines and Risers, 1.10.1 Drill string components and drilling tools (tubulars, jars, subs, stabilisers, reamers, etc), 1.14 Casing and Cementing, 5.2.1 Phase Behavior and PVT Measurements, 1.10 Drilling Equipment, 2.7.1 Completion Fluids, 1.6 Drilling Operations, 3 Production and Well Operations, 1.1.2 Authority for expenditures (AFE), 1.7 Pressure Management, 4.1.5 Processing Equipment, 4.6 Natural Gas, 2 Well Completion, 4.3.1 Hydrates, 4.2.4 Risers, 5.1.1 Exploration, Development, Structural Geology, 6.1 HSSE & Social Responsibility Management, 4.5 Offshore Facilities and Subsea Systems
- 3 in the last 30 days
- 746 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 10.00|
|SPE Non-Member Price:||USD 30.00|
This is the first of three related papers that describe the actions of an incident-investigation team formed to evaluate the failure of Well A-2 on the Marlin tension-leg platform (TLP). This paper outlines several possible failure modes and narrows the field of candidate failure modes to a small subset by deduction from both analysis and physical evidence. The application of these results and analytical techniques, as incorporated into the redesign of the subsequent completions, is addressed in the second paper of the series. The third paper discusses real-time monitoring of the redesigned completions.
The Marlin field is located in the Gulf of Mexico, Viosca Knoll Blocks 871/915, and was originally intended to be produced from a TLP by means of five predrilled dry-tree penetrations. First production from Well A-2 began in November 1999. Shortly thereafter, a minor but persistent tubing leak occurred. Between 7 and 20 November, Well A-2 was alternately flowing and shut in, depending on the shakedown of surface equipment and pipeline availability. On 20 November, the casing pressure jumped to shut-in tubing pressure, signifying a major tubing failure.
This paper describes the actions of an incident-investigation team formed to evaluate the failure of Well A-2. Possible failure modes investigated by the team include:
1. Helical buckling of the production tubing, with or without the combined loss of a tubing centralizer.
2. Crushing of the tubing by ratcheting displacement from a failed centralizer.
3. Lateral deflection of the subsea wellhead.
4. Collapse of the production (and/or intermediate) casing onto the tubing because of one or a combination of the following.
- Hydrate formation outside the intermediate casing owing to gas migration from a shallow hydrocarbon zone with subsequent dissolution during initial production.
- Nonuniform loading of the production casing because of the geometry of the submudline packoff tubing hanger (POTH) slips.
- Leak in a production-casing connection.
- Annular fluid expansion (AFE).
- Formation of a heat pipe in the production tubing by production casing/riser annulus.
- Casing wear on the intermediate casing.
- Excessive initial pressure resulting from setting the casing hanger seal assembly in the subsea wellhead.
5. Inadequate performance of the tubing or casing caused by either an incorrectly run or an inadequately manufactured joint.
Deducing from both analysis and physical evidence, the current installment will narrow the field of candidate failure modes to a small subset. Applying these results and analytical tech- niques, as incorporated into the redesign of the subsequent completions, is addressed in the second paper of the series.1 The final paper describes the implementation of the redesign as an insulated-tubing completion.2
Fig. 1 is a wellbore sketch of Well A-2. The well is located in 3,230 ft of water and is tied back to the Marlin TLP by means of a single 10 3/4-in. production riser. The wellbore trajectory is vertical to a depth of 5,400 ft relative to the rotary kelly bushing (RKB), then builds at 2 to 3° /100 ft to a sail angle of 45°.
This paper focuses on well failure shortly after first production, as evidenced by severe ovalization of the production tubing and loss of pressure containment by this tubing. The discussion follows the line of the investigation, delineating various events that might culminate in tubing deformation. Some of these events may now be reasonably discarded because of observations following recovery of the tubing above the deformed section. These discarded candidate failure modes are still discussed, however, with the intent of prompting other project teams on design considerations that might otherwise be overlooked.
Throughout this paper, the submudline tubing hanger-packer is called a POTH.
The well annuli are designated in alphabetical order, proceeding outward from the production tubing/tieback (5 1/2-×10 3/4-in.) annulus, designated "A." Thus, the "B" annulus is outside the production tieback (e.g., tapered 10 3/4×8 5/8× tapered 13 3/8×10 3/4 in.), and the "C" annulus is outside the intermediate casing (e.g., tapered 13 3/8-×10 3/4-x16-in. liner hung inside 20-in. casing).
Events Before and Immediately After the Failure
In Fig. 1, note the locations of pressure/temperature (P&T) gauges near the mudline [3,392 ft measured depth (MD)] and also near the production packer (12,636 ft MD). Fig. 2 summarizes the values at these gauges during a portion of the early flow of Well A-2. (The well was actually flowed for a period on 5 November 1999, as indicated in Figs. 3 and 4. This initial flow was shut in because of an upset in the not-fully-commissioned surface facilities. The gradual cooldown from this initial flow period can be seen in the mudline annulus temperature in the left portion of Fig. 2.)
For the flow period beginning at 1500 hours on 6 November 1999, note the increase in mudline temperature, which approaches a steady-state value of 160°F after approximately 20 hours. Simultaneously, the annulus pressure of the nitrogen column, set at 50 psi at the surface, increases slightly to 70 psi. These values are associated with a flow rate of 30 MMscf/D of gas, 3,000 B/D of condensate, and a surface tubing pressure of 4,140 psi.
The steady downhole conditions cease at 1045 hours on 7 November 1999, as evidenced by a significantly increasing rate of pressure in the annulus (13.6 psi/hr) as well as a subtle (~2°F) cooling of this same annulus before well shut-in (and major cooling to return to ambient conditions). A minor leak developed in the flow conduit, which was later diagnosed to be above the surface-controlled subsurface safety valve (SCSSV) and the POTH. The well was shut in at 1500 hours on 7 November 1999 because of a facility upset unrelated to the minor leak. Shortly after the last recordings shown in Fig. 2 and while the well was shut in, the downhole gauges were functionally lost.
|File Size||2 MB||Number of Pages||8|