Advanced modeling, real-time model updating improve extended-reach drilling success

Jan. 27, 2003
By applying advanced predrill modeling combined with real-time model updates, ChevronTexaco Corp., San Ramon, Calif., successfully drilled three extended-reach wells to 19,000-ft horizontal displacement in the Petronius field of Gulf of Mexico's Viosca Knoll Block 786 (Fig. 1).
Click here to enlarge image

By applying advanced predrill modeling combined with real-time model updates, ChevronTexaco Corp., San Ramon, Calif., successfully drilled three extended-reach wells to 19,000-ft horizontal displacement in the Petronius field of Gulf of Mexico's Viosca Knoll Block 786 (Fig. 1).

Compared to conventional vertical wells, complications from formation pore pressure, borehole stability, hole cleaning, drillstring torque and drag, and other parameters create serious challenges for extended reach and horizontal well drilling in deep water.

In most deepwater drilling environments, high formation overpressures combined with low fracture pressures place narrow limits on the stable mud-weight window.

Although the pore pressures in the Petronius field were normal, a wellbore-stability problem reduced the safe mud-weight window to as little as 0.5 ppg.

The uncertainty of initial pore pressure and fracture-gradient predictions carry many different drilling risks, especially in a difficult deepwater environment.

By applying new technology to deepwater extended-reach drilling (ERD), the company reduced drilling uncertainty and either avoided or was prepared to deal with drilling risks to successfully meet the well's objectives.

Field development

With the Petronius field under development since 2000, ChevronTexaco has drilled one vertical well and several deviated wells in various directions from the platform before drilling the three ERD wells (Fig. 1).

As horizontal displacement increased, at the same TVD, the wells had become increasingly difficult to drill and the need for wellbore-stability modeling and management became apparent.

To accomplish the task, the company applied a service Schlumberger Ltd. calls "No Drilling Surprises" (NDS) that provides predrill modeling, real-time model updating, and drilling-parameter change implementation at the rig site to lower and manage risk and reduce nonproductive time.1 3

The success of the process required flexibility, efficient communication, and close collaboration between the operator and service company.

By integrating the many disciplines and with open collaboration in one process, the companies successfully drilled the ERD wells. The operator avoided potentially stuck pipe and lost returns and saved up to 30% of drilling time.

The approach relied on integrated solutions that reduced uncertainty in pore pressure, fracture-gradient, and various drilling-parameter predictions to avoid problems. The approach also allowed workers to apply fast-growing knowledge to subsequent wells.

Objectives, well design

Engineers gathered the available field data and identified key concerns to determine options available to identify and solve problems.

Hole cleaning, excessive circulation time, tight hole, pack offs, tools lost-in-hole, and sidetracks were among the field's typical drilling problems.

Due to a narrowing safe-mud-weight window, the problems worsened with increasing hole inclination. Formation instability made the operational mud-weight window close to or less than the difference between equivalent static density (ESD) and equivalent circulating density (ECD).

The pressure fluctuations from the mud-circulating pumps on-and-off condition were greater than the pressure difference for the hole to stay intact, prevent flow, and prevent drilling fluid loss.

The difference between ESD and ECD became close to or greater than the difference between minimum safe pressure, or the density required to control formation pore pressure, and the fracture pressure or minimum horizontal stress.

The Petronius platform is one of the world's deepest fixed structures, positioned in more than 1,750-ft water depth, and is at the frontier of the Gulf of Mexico shelf and deep water.

Click here to enlarge image

There are directions in which the seabed depth rapidly changes, becoming shallower toward the north and deeper south of the platform (Fig. 2). Reservoirs south of the platform are under almost 3,200 ft of water but under only 700 ft of water to the north.

Such fluctuation in seabed depth creates significant environmental effects along the wellbore trajectory. Well designers must take these changes into account, extensively studying and modeling the stresses. The problem required a novel approach to modeling.

The new ERD wells would penetrate a previously unexplored block. The company planned one of the boreholes with inclination of 70° into the deep water of up to 3,200 ft water depth.

Operations had to drill through some low-pressure sands, as well as some unknown formations and dipping beds, before reaching the target location and reservoir that the company had not previously penetrated.

The main ERD well objectives were to:

  • Avoid high over pulls, stuck pipes, lost-in-hole tools or BHAs, or lost circulations.
  • Set the 95/8-in. casing past the unstable zone and avoid drilling into weak zones with excessively high mud weight.
  • Monitor and maintain in real time the ECD and ESD within the set limits.
  • Monitor and maintain the hole condition to keep drilling parameters within the rig limitations.

The company designed the ERD wells with the sections, as follows:

  • Previously set 20-in. conductor pipe.
  • Build-up section with 171/2-in. hole and 133/8-in. casing.
  • Tangent section with 121/4-in. hole and 95/8-in. casing.
  • Reservoir section with 81/2-in. hole and 7-in. liner.

Discipline integration

A geophysicist might explain that a pressure ramp is expected at a seismic time of 4 ms, which would be difficult for the driller and others involved in the drilling process to understand.

Predrill modeling requires considerable interaction and integration across the various disciplines. Integration and internal consistency is crucial in the whole process from planning until the well is drilled.

Even with the most accurate representation of the earth, there still would be some uncertainty during drilling. Mud engineers design the mud weight and rheology. Together with many other drilling engineering parameters and drilling practices, it determines the effective mud pressures in the hole.

Depending on these variables, the ESD can often be different than the mud weight. ECD fluctuations depend on hole size, BHA and drillstring configuration, pipe movements, tripping speed, pumping rates and pressures, drilling penetration rate, etc.

Without integrating all of this knowledge, to yield an understanding of ERD wells and wells with small mud-weight tolerances, the wells simply become undrillable.

Modeling

To resolve the drilling problems, provide reliable results, and account for all environmental changes, the No Drilling Surprises process used a 3D modeling approach and integrated the knowledge across many disciplines.

Engineers built a full 3D mechanical earth model (MEM) using 3D seismic log and test information.4 The model also incorporated the drilling experience from all wells previously drilled in the area.

Click here to enlarge image

The dipmeter, FMI (Fullbore Formation MicroImager log), and cuttings returns that indicate formation failure breakout revealed the stress direction. Breakout is the failure of the formation rock due to stress causing compression of the borehole wall (Fig. 3).

A good understanding of rock properties, such as formation density above the well path, is very important to estimate the stress in extended reach wells.

Since much of these essential data are usually missing, modeling provides the answers. Simple vertical integration of formation density does not account for dipping seabed or lateral anisotropy effects. Workers estimated the vertical or overburden stress by modeling quality-controlled density data from the offset wells as the function of depth below mud line.

Density information was not available through the complete depth interval. Density extrapolation from the surface to the top of the density data provided the needed information.

Engineers created a 3D density cube from a 3D-seismic velocity survey, with sonic data providing quality control. The workers integrated overburden stress along the well path, extracting the stress along the trajectory.

Overburden stress

Fig. 4 shows a comparison of the overburden stress estimated without correcting for the water depth change and overburden stress calculated with water depth change taken into account.

Click here to enlarge image

Applying the uncorrected vertical stress profile would give incorrect outputs of pore pressure, the minimum-horizontal stress, and therefore incorrect fracture gradient. These are the main inputs to the wellbore-stability model computation.

It is very dangerous in terms of wellbore stability and well control if engineers under or overestimate any of these gradients, which could be up to 1.5 ppg in this case, and could lead to costly lost circulation, borehole collapse, and lost BHA's.

Over estimating gradients could cost time and money and jeopardizing the safety of the rig and its personnel.

Wellbore stability

The modeling process used log data together with the seismic velocities for pore-pressure modeling, with corrections applied for the changing seabed slope.

The model calibrated minimum horizontal stress for shale, using the formation breakdown tests. The breakdown tests provided limits for maximum horizontal stress. The model also used the mud-loss information that was available.

Engineers performed a complete petrophysical analysis to establish the formations' mineralogical composition and derive the rock properties.

The left panel of Fig. 5 shows a petrophysical analysis, with colors representing the volume percent of the rock mineral composition.

Click here to enlarge image

The gray color corresponds to the volume of illite clay, yellow represents quartz, black is the interstitial or bound water, blue represents water or total porosity, with green and brown representing oil (Fig. 5).

The available calibration data from MDT (Modular Formation Dynamics Tester), RFT (Repeat Formation Tester), and leak-off tests provided calibrations for all of the profiles.

Engineers created the wellbore stability model along the specified trajectory of the well.2 Stability analysis indicated that an unstable zone existed within the well path toward the reservoir section that required drilling with the higher mud weight (Fig. 5).

This forced a stable mud weight window of less than 1 ppg, which was narrower than the ESD-ECD difference.

Due to the very narrow stable mud-weight window, indicated as the envelope between the MW0 and Sh curves of Fig. 5, drilling operations had to establish new safe margins.

For a given hole size and drillstring design, workers modeled the maximum magnitude of failure that the rig hydraulics could handle with minimum probability of losses.

Engineers found that a borehole failure of up to 60° of the borehole circumference or maximum breakout angle would be acceptable without affecting hole cleaning and integrity (Fig. 3). This implied, however, that drilling operations would require cautious monitoring.

The MW60 curve of Fig. 5 represents that limit. Note that this value is only theoretical. Once the well conditions initiate borehole wall failure, there is no predictive answer of how the breakout will behave. Experience has demonstrated that the failure has a time-dependent character.

Drilling operations, therefore, must place emphasis on ECD which is greater than breakout-initiation pressure when the circulation is on and there is a short connection time.

When no circulation is possible, the borehole is exposed only to ESD pressure.

The expected failure chart at the extreme right of Fig. 5 shows the predicted failure along the circumference of the borehole for the given mud weight.

Drilling mechanics

The model indicated a drilling mechanics response that called for replacing some of the drilling equipment.

Click here to enlarge image

An example is the improved hole cleaning from use of the rotary-steerable system called "Power Drive" (PD900) compared to sliding the drillstring when using the downhole positive-displacement motor (Fig. 6).

The PD900 allowed trajectory control while rotating and permitted better flow with less pressure drop within the tool.

Click here to enlarge image

The modeling process incorporated wellbore stability analysis, a cuttings and cavings corrected ECD, and annular velocities to optimize hole cleaning (Fig. 7). The model estimated the critical pump rate that the flow regime would change from laminar to turbulent.

The model also took into account the cuttings-corrected rate of penetration and the pressure loss over each of the drillstring elements. The process optimized hydraulic horsepower.

Click here to enlarge image

The modeling process analyzed torque and drag and generated theoretical profiles that were calibrated with the real-time pickup and slack-off weight data (Fig. 8).

To prevent drillstring failure, engineers performed a complete stress analysis evaluating bending stresses, sinusoidal buckling, effective axial load, total and inclination side forces, and torsional and tensile capacity.

To prevent or eliminate potential downtime, the modeling process considered the limitations of the most essential rig equipment.

In addition to a complete understanding of potential risks, hazards, and limitations found during the predrill preparations, the modeling process captured the lessons learned and best practices of actual drilling operations.

Engineers performed root-cause analysis for problem events and developed preventive and remedial actions for each.

Uncertainty

With the best-in-class techniques and processes, the tools and integration create the best representation of the parameters, qualities, and conditions. A problem, however, remains—the best available information along the actual borehole trajectory comes from modeling, estimates, and predictions.

The actual conditions begin to emerge only as the rig drills the well. Regardless of how detailed and how expertly engineered a predrill plan may be, it is obsolete as soon as new information becomes available. There is uncertainty in the prediction.

The process of integrating a multidisciplinary team to drill the wells, specifically focused on the uncertainty. This made it necessary to build the best possible internally consistent model that would minimize error.

In addition, the engineers built the model so that all new information acquired from the well would reduce the uncertainty, not only at the current depth but also ahead of the bit.

Real-time updating

As required by conditions, the predrill modeling allowed some failure of the hole to occur. To reduce the uncertainty and manage borehole integrity, drilling operations must closely monitor well conditions and update the wellbore stability model in real time.5

The engineers at the rig site continuously monitor the drilling parameters. The multidisciplinary team provides 24-hr support to the operations through connections between the office in town and the rig.

Various sources provide drilling information, which the engineers analyze appropriately and incorporate into the initial models.

The new measurements and new event information calibrates the model and provides new constraints to the various parameters.

Combining the drilling and logging information, the team in the office updates the wellbore stability model. The model updates the require information with use of the gamma ray, resistivity, sonic, and neutron density and porosity log measurements.

The process provides timely information and allows the team to give their recommendations for actions as necessary.

Borehole cleaning becomes a great challenge, to remove cuttings from the well and at the same time maintain stability in terms of both collapsing and fracturing the formation.

ECD was very sensitive to the hole condition and, in the case of the Petronius field wells, there was very little tolerance. Engineers closely monitor and manage the mud weight to within 0.1 ppg of the established, after formation test, baseline pressures.

Understanding the possible processes occurring in the borehole allowed real-time interpretation of the log and drilling parameters response.

Workers calibrated the pickup, slack off, and weight of the drillstring during rotation and compared the results with the actual measurements while drilling and during every trip (Fig. 8).

This provided the team with an overall understanding of the well conditions and, together with ECD and known responses, provided information that crews could act on.

Workers monitored the drilling-mechanics time logs and established several patterns to identify the borehole failure due to cavings generation and to differentiate hole loading with breakout material from drilling cuttings.

Even though not many cavings appeared at the surface, cavings appeared in the borehole.

Based on the observations, engineers could modify drilling, tripping, and circulating procedures.

Removing the larger cavings from the borehole required a mechanical action since circulation alone would not remove the material. Prior to coming out of the hole, crews increased the circulation times at the bottom, the casing shoe, and at the critical inclination angle.

Cavings or breakout material, indicating borehole failure, is easily distinguished at surface from normal drill cuttings (Fig. 9).
Click here to enlarge image

After several full circulations, when the shakers were clean of the normal PDC cuttings, caving formation material came to the surface for several hours (Fig. 9).

Although the modeling process had broadened the stable mud-weight window without jeopardizing borehole stability, the possibility of formation fracturing remained.

The first ERD well drilled encountered some ballooning and losses.

The team identified the fractures and promptly treated the losses with lost-circulation material and updated the real-time wellbore stability model to allow crews to lower the mud weight with confidence.

Click here to enlarge image

To identify the fractures, the team used a novel resistivity analysis of time- lapsed resistivity (Fig. 10). The engineers found the minimum horizontal stress to be 0.3 ppg less in the sands than in the shale.

The team developed a new model for minimum horizontal stress for different lithologies.

The information from the well and modeling power allowed the team to understand the processes going on in the well around the clock, continuously.

Results, conclusions

The company successfully drilled and competed the three ERD wells while applying real-time wellbore-stability estimation and management. The team integrated the process across the various disciplines.

Managing minor mud losses, the drilling operations successfully reached all targets and ran all casing strings to the planned depth. No stuck pipe incidences, lost-in-hole, or costly sidetracks occurred.

For the three wells, the company saved an average of 15% of the AFE days, with about 45% savings in planned drilling time.

The real-time updating of the earth model and the drilling events database allowed the operator to have the best possible representation of the earth at any point while drilling.

Effective knowledge capture of the best practices observed and lessons learned, as well as an updated earth model and drilling database, made an efficient knowledge transfer process to the next well and improved the learning curve.

Acknowledgments

The authors thank Gemma Keaney, Patrick Hooyman, and Charlotte Sodolak, of Schlumberger, Houston, for help in preparation of this article.

References

1. Bratton, T., Edwards, S., and Fuller, J., "Avoiding Drilling Problems," Oilfield Review, Summer 2001, p. 32.

2. Addis, T., Boulter, D., Last, N., and Plumb, R.A., "The Quest for Borehole Stability in the Cusiana Feld, Colombia," Oilfield Review, Summer 1999, p. 33.

3. Aldred, W., Plumb, R.A., and Cousins, L., "Managing Drilling Risk," Oilfield Review, Apr.-July 1993, p.2.

4. Plumb, R.A., Edwards, S., Pidcock, G., Lee, D., and Stacey, B., "The Mechanical Earth Model Concept and Its Application to High-Risk Well Construction Projects," paper No. SPE 59128, presented at the IADC/SPE Annual Drilling Conference, Feb. 23-25, 2000, New Orleans.

5. Bradford, I.D.R., Aldred, W.A., Cook, J.M., Elewaut, E.F.M., Fuller, J.A., Kristiansen, T.G., and Walsgrove, T.R., "When Rock Mechanics Met Drilling: Effective Implementation of Real-Time Wellbore Stability Control," paper No. SPE 59121, presented at the IADC/SPE Annual Drilling Conference, Feb. 23-25, 2000, New Orleans.

The authors

Click here to enlarge image

Nikolay Y. Smirnov works in the Holditch-Reservoir Technology Department of Schlumberger Inc. in Houston. He previous worked for Schlumberger Sedco-Forex as a drilling engineer and for Schlumberger's geomechanics group. Smirnov has a degree in geophysics from the Novosibirsk State University.

Click here to enlarge image

John C. Tomlinson is Schlumberger Inc.'s in-house sales and service engineer in drilling and measurements for ChevronTexaco Corp. He previous worked as a mud-logger and for the Anadrill product line of Schlumberger. He is an SPE and SPWLA member.

Click here to enlarge image

Samuel D. Brady is a drilling engineering coordinator for ChevronTexaco Corp. on its Agbami project, off West Africa. He previously worked as a deepwater development engineer for Texaco Inc. in the Gulf of Mexico. Brady has a BS in engineering of mines from West Virginia University and an MBA from the University of New Orleans. He is an SPE member and a registered professional petroleum engineer in Louisiana.

Click here to enlarge image

W.E. (Bill) Rau III is a senior project drilling superintendent with ChevronTexaco Corp.'s Gulf of Mexico deepwater business unit in New Orleans. His career has included various drilling assignments around the world, and currently he is with the Petronius project. Rau has a BS in mechanical engineering from Tulane University.

Based on a presentation to XIV Deep Offshore Technology Conference and Exhibition, New Orleans, Nov. 13-15, 2002.