No abstract

This paper presents our perspective of the shallow-water flow (SWF) problemin the Deepwater Gulf of Mexico (GOM). The nature of the problem, includingareal extent and overpressuring mechanisms, is discussed. Methods for sandprediction and shallow sediment and flow characterization are reviewed. Theseinclude seismic techniques, the use of geotechnical wells, regional trends, various MWD methods, and cements and settable spots. Finally, examples of flowincidents with pertinent drilling issues, including well failures andabandonment, are described. Total trouble costs due to shallow-water flow for all GOM operators probablyruns into the several hundred million dollars. Though the problem remains aconcern, advances in our knowledge and understanding make it a problem that ismanageable and not the "show stopper" once feared. Introduction SWF may occur while drilling shallow over-pressured formations at deepwatersites. It is a high profile problem in the GOM, though it does occur elsewhere(Ref. 1) and will likely be encountered in other deepwater regions (Fig. 1). Drilling shallow over-pressured sands may cause large and long lastinguncontrolled flows, well damage and foundation failure, formation compaction, damaged casing, and re-entry and control problems. Most spectacularly, eruptions from over-pressured sands may result in seafloor craters, mounds andcracks (Figs. 2 and 3). Eaton2, 3 has described the significant problems causedby SWF in the Ursa area. A recent inspection of 106 wells (Ref. 4) indicatedthat $175 MM has been spent on SWF and prevention and remediation on thosewells. Total industry costs due to SWF likely exceed several $100 MM. The problem is compounded by the difficulty in seismically imaging thesesands (Refs. 5 and 6). This stems from the relatively low sand/shale contrastin acoustic impedance. The impact of this problem on well site selection andwell design is significant, and makes drilling in SWF areas particularlychallenging.

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractAs part of the effort to understand the problem of shallow water flow (SWF) in the Gulf of Mexico (GOM), and to obtain data and develop design criteria for offshore production structures, geotechnical wells were drilled in several different prospects. These prospects cover a 200+ mile swath across the central deepwater GOM where SWF control problems have been experienced. In these wells, pore pressure measurements, as well as core and other in-situ measurements, were taken in the deepwater shallow sediments.Key facts and observations from these measurements include the following:1. A majority of the pore pressure measurements were made in low permeability clay rich material, predominant in the shallow sediments. Some measurements were proximal to more permeable sand and silt zones. A few measurements were made in sandy/silty intervals.2. Where significant over-pressures were present they were found to begin at or very near the mudline, and to increase more-or-less linearly with depth below mudline (BML).3. Over-pressures in the shallow sediments of the central deepwater GOM are due primarily to rapid sedimentation rates generated by the Mississippi River depocenter. This is supported by the fact that measured over-pressures exhibit regional trends generally consistent with sedimentation rate. 4. Regional trends in over-pressures are also correlative with drilling experience and the incidence and severity of shallow water flow (SWF) occurrences.5. The degree of over-pressuring is consistent with sediment porosity based on core measurements, i.e., higher over-pressures are associated with higher porosities.6. There are exceptions to these general regional trends. Most notable is that over-pressures in shallow permeable sand or silt zones may not be in equilibrium with their bounding shales. Both pressure depletion and hyper-pressuring have been observed in the geotechnical-well pore pressure measurements and in drilling operations. Thus well location (up dip, down dip), faulting, fracturing and other factors need to be considered in predicting pore pressures in these permeable zones.

Pile to soil shear transfer and load distribution behavior for a 3D-inch diameter test pile driven from 190 to 263 feet of penetration in stiff to hard, moderately overconsolidated (OCR = 2-3) silty clay are presented. Measured shear transfer vs. displacement (t-z) curves are compared to t~ose predicted using the methods of Vij~yvergiya (1977) and Kraft, Ray and Kagawa (1981).Effects of degradation during cyclic one-way tension and two-way tension-compression on both load distribution and shear transfer behavior are presented. Degradation effects were minimal for the one-way loading and substantial for the two-way loading. Rate of loading effects on shear transfer are also presented.The measured static and cyclic pile behavior was modeled using the PSAS pile3soil interaction algorithm in INTRA (Bea et al, 1984). The pile was modeled as a series of beam elements. Non-linear hysteretic springs and dashpots were used to model the soil.When elasto-plastic soil support curves were used in conjunction with the Unconsolidated-Undrained Triaxial (UU) Strengths from driven sample~and the recommended design method of API-RP2A (1986), INTRA underpredicted the long term static pile capacity. Using the measured t-z curves, the measured pile-top load deflection behavior under both static and cyclic loading was successfully modeled.

The design methods for axially loaded piles contained in the API RP 2A, Recommended Practice for Planning. Designing and Constructing Fixed Offshore Platiorms are continually evaluated and updated. To support this activity, API has fimded research over the past 10-12 years which has resulted in the establishment of an internationally recognized data base of axial pile load tests. Subsequent interpretation of these data led to sigrdllcant changes in the design methods for axially loaded piles for both sand and clay. The sand revisions, first incorporated into the 15th Edition of RP 2A provide improved statistical predictions (bias and coeftlcient of variation) of the sand data base than did the design methods for sand contained in earlier editions. The clay revisitms, first incorporated into the 17th Edition of RP 2A provide a statistical description of the clay data base that is similar to descriptions provided by earlier clay methods. Both the revised clay and sand methods provide a more technically sound framework than the methods contained in earlier editions. The revised methods are contained in the current, i.e., the 20th Edition of RP 2A (1993). These revisions have initiated considerable debate, and have prompted the authors to present their thoughts herein. A discussion of the data base, review of the philosophy and evolution of the design method, and a summary and analysis of the key criticisms are provided. The data have been exceedingly useful in formulating design practice, but in our opinion they are too limited to conclusively resolve many of the important issues. While some moditlcations to the present practice seem to be indicated, the References and figures at end of paper 253 authors believe that when combined with a careful site investigation, attention to installation detail and good engineering judgment, RP 2A continues to provide a sound basis for design of axially loaded piles.