DRILLING BRAZILIAN SALT: CENPES uses FEA to design casing for Santos basin

June 12, 2006
Planning, modeling, collecting data, and maintaining flexible response options are necessary to drill thick, highly mobile salt sections successfully.

Planning, modeling, collecting data, and maintaining flexible response options are necessary to drill thick, highly mobile salt sections successfully. A continuous circulation system to monitor and control fluid density closely, and seismic-while-drilling sensors are valuable tools; high-collapse-strength casing is vital.

Petrobras is faced with drilling through thick sections of salt offshore in the Campos and Santos basins. This study, conducted at the Petrobras Research and Development Center (CENPES) in Rio de Janiero, involves developing the drilling and casing plans for the first of four deep exploratory wells in the Santos basin. The first well site is in 2,140 m water, planned to 6,000 m TVD, including nearly 2,000 m of various salt rocks, including halite, carnallite, and tachyhydrite.

Part 1 reviewed salt behavior, experimental test results, numerical simulations and models of salt creep and well closure, leading to a preliminary drilling strategy. Some of the experimental work was carried out at the Rock Mechanics and Rock Hydraulics Laboratory at the Institute for Technological Research of the State of São Paulo.

This concluding article focuses on the finite-element modeling in numerical simulation used for the casing design, mud and cement for the well, planned through highly mobile layers of carnallite and tachyhydrite salts.

FE model of casing design

The casing was designed after several cementation failure scenarios were tested. These ranged from 5%-20% uncemented in the annulus casing and borehole through the salt layer. The aim was to determine nonuniform loading and the timing of salt loading on well casing causing deformation.1

For pre- and postprocessing and the numerical simulations of the finite-element model, we used the same software as in the well closure simulation.

The plane strain model, perpendicular to the longitudinal axis of the well was built with 14,506 quadratic isoparametric elements (with eight nodes) and 43,639 nodal points in the finite-element model (Fig. 1). We analyzed the deepest tachyhydrite layer (4,301 m).

The plane strain model, perpendicular to the longitudinal axis of the well, was built with 14,506 quadratic isoparametric elements (with eight nodes) and 43,639 nodal points in the finite-element model (Fig. 1).
Click here to enlarge image

We used 100 m diameter for the model, sufficient to avoid boundary effects (borehole diameter 17½-in.) and anchored ends. The modeling consisted of two steps: We applied the mesh rezone (“excavation”) in a circular borehole with symmetrical closure; then, after a specific time, we introduced the casing and cement, using a process developed with Petrobras’s ANVEC code.

Simulation of casing design

We used numerical simulations to predict the loading on well casing induced by the salt creep. The casing was designed to be capable of supporting the high creep rate of tachyhydrite.

The results of numerical modeling with a 100% cemented annulus, after 500 hr in a concentric wellbore, showed that the stress in the high-collapse casing (14-in. x 0.722-in.; P110; 9,500 psi; 1.5% maximum ovalization) is just 0.23 of the material’s specified minimum yield strength (SMYS), due to the uniform loading.

In deep-set casing through the salt layers, it is probable that the salt-casing annulus will have to remain uncemented. We ran simulations to evaluate 5%, 10%, 15%, and 20% of cement channeling after 500 hr (Figs. 2a-d, respectively). The results show that with 15% cement channeling or greater, the nonuniform loading induces casing collapse. To avoid creating loading points in the casing caused by the cement channeling, at least 90% cementation is necessary. Securing this requires a short job cementation. This makes it possible to set deep casing.

Simulations evaluated 5%, 10%, 15%, and 20% of cement channeling after 500 hr (Figs. 2a-d, respectively). The results show that with 15% cement channeling or greater, the nonuniform loading induces casing collapse.
Click here to enlarge image

The tachyhydrite layers will first impinge upon the casing, however, and a nonuniform loading condition could initially develop by the differential stress along the casing (squeezing). To minimize this possibility, we simulated a dense mud (16.6 ppg) placed between the top of the cement shoe and the casing head. This denser fluid redistributes the loads uniformly. The pressure increase in the shoe, created by the closure of the hole, was 818 psi in 490 hr; the radial casing deformation was 0.007 m, and the stress in the casing was 0.57 of SMYS.

We ran the same analysis for non-concentric casing. The stress in the non-concentric casing reaches 0.61 of SMYS, just 4% higher than the concentric case.

Well Plan A

Based on evaluation of salt movement and no additional problems, the basic case well planning (Fig. 3) is to drill a 36-in. hole, 70 m below mud line, with seawater and set the 30-in. conductor; drill a 26-in. hole up to 60 m inside the salt with seawater and change to salt-saturated water-base mud after reaching the salt, and set the 20-in. surface casing.

Click here to enlarge image

After a change to synthetic base mud with 14 ppg, a density necessary to control the salt creep according to the creep simulations, we then drilled the hole 14¾-in. x 17½-in. with a rhino reamer from Sii Smith Services up to 200 m above the base of the salt to set the 14-in. OD intermediate casing.

Based on our experience in drilling salt in other Campos basin wells, we decided to use reamers instead of bicenter bits. Reamers are known to produce a more concentric borehole and allow backreaming, even though the bit can potentially get stuck in the pilot hole between the bit and the under reamer, since there is no hole enlargement in this interval.

We reduced the mud weight to 10 ppg to drill the 12¼-in. x 14¾-in. hole through the remaining salt and the rubble zone that is expected just below the salt, up to the top of the first target, then ran the 10¾-in. second intermediate casing.

We planned to drill the last interval of the well through the objectives and down to TD using a 9-in. bit. If necessary, we could run 7 5/8-in. production liner. This is an optimistic well design, assuming that everything is going right.

Uncertainties in Plan A

There are some uncertainties associated with the project. The first is related to the leak-off test (LOT) at the 20-in. casing, about 50 m into the salt. We know that the pressures recorded during a LOT in salt should be greater than in any other sedimentary rock at the same depth. We do not know, however, if it will reach the minimum necessary value of 14.5 ppg (based on 14 ppg mud weight plus 0.5 ppg ECD + safety factor), at a depth less than 1,000 m below the mud line. There are no LOT data available in salt in nearby wells.

Assuming that Eaton’s method is a valid way to evaluate the fracture gradient (Fig. 4); the result is 11.7 ppg, not high enough for what was planned. However, we know that the method is not defined for this situation.

Click here to enlarge image

Based on published data2-4 and scout information from the Gulf of Mexico, we determined that the result from the LOT at the 20-in. shoe should be greater than the overburden gradient (between 10% and 20%); that’s closer to the required value of 14.5 ppg. Since the mud operational window for this situation is so critical, we are considering using Varco’s continuous circulation system (CCS), in order to keep the equivalent circulating density (ECD) continuously higher than the wall of the borehole, even while making or breaking connections.

The second possible source of problems is related to the relative amounts and thickness of layers of carnallite and tachyhydrite (with high creep rate). The expected salt section was determined by a geophysicist analyzing seismic data. If we find more layers during drilling than were modeled in the creep-rate simulation, it might be that we could not keep the hole open long enough fully to penetrate the salt section and run the intermediate casing.

If any of the two situations occur, it will be necessary to run an intermediate casing string in the middle of the salt section to cover to soluble salt layers. Since this is a high possibility, we included contingency plans, explained in the next section.

Plan B with contingencies

A second well plan B involved contingencies (Fig. 3). Similar to Plan A, the crew would drill a 36-in. diameter hole to 60 m below mud line, set the 30-in. conductor; drill a 26-in. hole as far as 60 m into the salt using salt-saturated, water-based mud, and set the 20-in. surface casing.

After changing to synthetic base mud with 14 ppg, as in Plan A, the crew would drill the hole 14¾-in. x 17½-in. with a rhino reamer as far as half way through the salt column, where we expect to reach the base of the soluble salt (carnallite and tachyhydrite), and then set the 14-in. intermediate casing.

The next salt interval would be drilled with 12¼-in. x 14¾-in. with rhino reamer as far as 30 m above the base of the salt, using the same mud weight (14 ppg). This would deliberately avoid tagging the rubble zone expected just below the base of salt.

In order to detect the base of the salt during drilling, we are planning to use seismic-while-drilling equipment. The 10¾-in.-diameter, second intermediate casing string should be set 40 m above the top of the salt.

After the setting and cementing of the 10¾-in. casing, the mud weight will be reduced to 10 ppg to avoid lost circulation in the rubble zone. Then, we will finish drilling with a 9 in. bit to the top of the first target and set the 7 5/8-in. liner.

The objective horizons will be drilled with 6½-in. bit down to TVD at 6,014 m.

Uncertainties in Plan B

Another critical point for this well design is that in the event we have to use Plan B, that is, the creep rate is too high for drilling and casing the whole salt section, it will be necessary to run the 14-in.-diameter intermediate casing. According to results from the simulation that we ran to determine the magnitude and timing of salt loading on well casing, this casing does not have enough collapse resistance to keep the hole open. Even the strongest 14-in. casing available will suffer collapse due to accelerated salt creep during the subsequent drilling stage (12¼-in. x 14¾-in.). It is not advisable to cement the entire annulus, since it is possible to create some cement channeling. This would be worse in generating point, or nonuniform, loads on the casing, resulting in casing collapse.

To avoid this situation we are planning to put a slug of very dense mud (16.6 ppg) in front of the cement slurry. That mud will be placed between the top of the cement and the casing head. This fluid will redistribute the loads created by the closure of the hole.

It is important to point out that the 14-in. casing is a temporary solution to sustain the salt creep, that is, just to work during the drilling of the next phase. The definitive solution will be reached after setting the 10¾-in. diameter, 133 lb/ft (high collapse strength) casing that is designed to resist the salt movement permanently.

Learnings

This project allowed Petrobras R&D to develop a methodology to design a mud program and casing, as well as define a strategy to drill subsalt prospects. We used computer modeling to evaluate the creep behavior of salt rocks under high differential stress, high temperature, and applied casing loads to replicate salt creep. Computer code developed in-house, based on the finite-element method, was used for numerical simulations.

We discovered that the:

  • Ideal mud weight used through the salt layers should be 14.0 ppg and 10.0 ppg, respectively, in two drilling stages.
  • Maximum allowed failure of cementation in the annular space between the wellbore and 14-in.casing is 10%.
  • Minimum drilling fluid weight used between the casing and wellbore, placed between the top of the cement shoe and the casing head, should be 16.6 ppg.
  • Leak-off test should run up to fracturing. Until now, the LOT in salt has been run with limited pressure to support the combined pressure (MW + ECD) planned for the salt section. In order to drill through thick salt layers, thorough knowledge of the fracture gradient in salt is necessary.

It would be useful to use a continuous circulation system (CSS) for this prospect, at least on the first well, while we calibrate some of the parameters, such as the critical temperature, and get a more reliable creep rate. With this equipment, the mud weight could be continuously monitored and the results of the LOT at the 20-in. casing shoe would become less critical.

We also believe it would be useful to use seismic-while-drilling to determine the base of the salt and then set the casing just above the expected rubble zone. It may help avoid loss of circulation.

The lessons learned on this exploratory well will allow Petrobras to optimize the well designs for the next three wells proposed for this area.

Acknowledgments

The authors thank Petrobras SA for permission to publish this work and all the people involved in this project. Part of the work was performed at the Rock Mechanics and Rock Hydraulics Laboratory at the Institute for Technological Research of the State of São Paulo, funded by Petrobras.

References

  1. Willson, S.M., Fossum, A.F., Fredrich, J.T., “Assessment of Salt Loading on Well Casings,” paper No. 74562, SPE Annual Drilling Conference, Dallas, Feb. 26-28, 2002.
  2. Barker, John W., and Meeks, W.R., “Estimating Fracture Gradients in Gulf of Mexico Deepwater, Shallow, Massive Salt Sections,” paper No. 84552, SPE Annual Technical Conference and Exhibition, Denver, Oct. 5-8, 2003.
  3. Falcao, Jose L., “Drilling in High-Temperature Areas in Brazil: A Wellbore Stability Approach,” paper No. 69516, SPE Latin American and Caribbean Petroleum Engineering Conference, Buenos Aires, Mar. 25-28, 2001.
  4. Barker, John W., Feland, K.W., and Tsao, Y.H., “Drilling Long Salt Sections Along the U.S. Gulf Coast,” paper No. 24605, SPE Annual Technical Conference and Exhibition, Washington, Oct. 4-7, 1992.

Based on SPE paper 99161, presented at the IADC/SPE Drilling Conference, Miami, Feb. 21-23, 2006.