Harsh Environment Wells

New technology cured severe mud losses in challenging HPHT wells

Enhancing wellbore pressure containment solved a downhole well control situation in a severe mud loss zone in a Mediterranean Sea well with complex, treacherous sand zones.

Ron Sweatman,  Hong Wang and Bob Engelman, Halliburton; Medhat Sanad and Carl Butler, BP

As described here, an innovative technology for enhancing wellbore pressure containment (WPC) allowed an operator in the Mediterranean Sea to control lost circulation and crossflows in a single treatment during an underground well control situation. In addition to stopping lost circulation, the new process widened the mud weight window by increasing the WPC over 2.0 lb ppge above the natural frac gradient.

The DrillAhead* Services (DAS) technology allowed the operator to quickly seal a severe loss zone and increase its pressure containment, bring the well safely under control and set a temporary plug. When the well was re-entered seven months later, the seal was found to be still competent. The wellbore was conditioned to run a 7-in. liner without losses or further remediation, which saved over US$1 million. The technology's software-aided analysis capacity is engineered to provide treatment designs based on rock mechanical properties and post-job analysis.

Drilling in the offshore Nile Delta of the Mediterranean Sea (Fig. 1) constantly challenges existing technology. The targets are mainly in the Pliocene and pre-Pliocene sands. Exploration successes largely hinge on understanding regional pressure trends and the complex structural elements and seismic quality challenges.^1,2 Pressure regimes are treacherous; geopressured, shale-bound sands may exhibit pressure regression, leading to loss or gain situations within an ECD margin of 0.03 ppge.

Fig. 1. Nile Delta, offshore Egypt in the Mediterranean Sea.10,11

The formations also show a tendency to “breathe” during trips, especially after a zone has been losing minor amounts of mud. The formation may pack off around the drill pipe and make tripping out impossible. “Pumping out of hole” is the normal method to facilitate drill pipe removal.

The WPC treatment technique recommended for this well had been used successfully in many locations worldwide to stop drilling/ completion fluid losses.^3–5 Additionally, versions of it had helped: increase or restore the WPC integrity of weak formations, extend casing shoe points, and eliminate pipe strings such as drilling liners.^6–9 Field studies indicated that the WPC treatment appeared to remedy losses without impairing production rates in cased-hole completions.^{3, 7–9}

The operator requested a post-well study^10,11 of the lost-circulation/ flow event that developed in the final hole section at 15,741 ft (4,798 m) with BHTs up to 300°F and pressures over 11,000 psi. Using a new numerical model, WPC specialists were able to accurately model the conditions and demonstrate why conventional lost circulation materials and crosslinked pills were unsuccessful.

PREDICTION/JOB DESIGN SOFTWARE – DrillAhead* PLATFORM

With the prediction capabilities offered by DAS, a well design can be optimized with consideration of the widened mud weight window.⁹ Two pieces of critical information are needed to predict the achievable WPC: formation properties and fluid/ sealant properties. Formation properties can provide information for defining the weak zone and the fracture geometry. Formation properties such as in-situ stress, Young's modulus and Poisson's ratio, can be derived from logs and normalized with available direct measurements. Fluid/ sealant properties can be used to define treatment details for a given required increase in WPC. Its properties come from direct lab measurements. After formation properties and fluids/ sealants are defined, the software can be used to:

• Simulate defined drilling conditions, such as equivalent circulating density (ECD) to identify weak zones and report their locations.
• Report for each weak zone's fracture geometry, such as height, width and length.
• Predict the loss rate at a certain ECD.
• Select an appropriate sealant product.
• Predict an achievable increase in WPC by treatments.
• Simulate the sealant placement process.

KICK, LOSS AND CRITICAL TIMING

The 9-5/8-in. liner was set at 14,022 ft (4,274 m) MD to isolate high-pressure A-60 and A-70 Pliocene sands from lower-pressured Serravalian objective sands. A 17.8-ppge leakoff test (LOT) and a change in the C3:C4 ratio of gas readings before and after running pipe, indicated that the previous interval had been isolated.

When returns from the shale drilled below the shoe carried 700 units of gas and mud was observed bubbling over the bell nipple, mud weight was raised to 16.0 lbm/gal from 15.7 lbm/gal. ECD was 17.1 ppge at a flow rate of 425 gal/min. No losses were observed drilling Sand 1 but, as Sand 2 was penetrated, the loss rate reached 20 bbl/hr.

The mud weight was reduced to 15.8 lbm/gal and, by “pumping out,” the operator was able to make several trips and a coring run in Sand 1 without incident. Conventional lost circulation material (LCM) pills, along with controlled drilling and flowrate reductions, were used to manage loss rates up to 120 bbl/hr. Drilling stopped at 15,741 ft (4,798 m) MD, 14,852 ft (4,527 m) TVD, to log and evaluate the wellbore. A total of 1,300 bbl of oil-based mud had been lost.

Mud weight was increased to 15.9 lbm/gal and several short trips into the 9-5/8-in. liner were made to help ensure that the well was stable. Some wellbore “breathing” was observed, so the operator circulated bottoms-up and monitored flow with pumps off to determine the nature of the flow. When convinced that the well was not flowing, the operator continued pumping the drillstring out the hole in stages, checking frequently to confirm that a well-control situation was not developing.

During the third logging run, the well began flowing and was shut-in at the surface with 120 psi. To reduce casing pressure so that drill pipe could be stripped in the hole, 17.0 lbm/gal mud was bullheaded into the well. The well was taking fluid at 16.3 ppge. After stripping drill pipe to 5,612 ft (1,711 m) MD, 18.0 lbm/gal mud was circulated to kill the well. Drill pipe was run in the hole to the 9-5/8-in. shoe, where the mud was conditioned with substantial losses.

Resistivity trends implied that losses were occurring at numerous points in the 8-1/2-in. openhole below the top of Sand 2. Several LCM pills containing 40 to 60 ppb of mixed, fibrous particulate and platelet-type additives were pumped with little success. Following this effort, two 40-bbl crosslinked polymer pills were pumped. After waiting the prescribed time for the pills to set, loss rates remained unchanged.

The WPC process was selected as the next step to enable tripping and subsequent liner installation operations while minimizing the risk of losing the openhole section. Design criteria for this successful application were taken from experience gained in the cases listed in Tables 1 and 2. Before the WPC treatment was applied, the well had lost about 4,000 bbl of oil-based mud over a period of eight days of failed pill attempts. Timing increased the urgency of finding a reliable remedy. Plans to install a jacket had to be executed during the stable weather window occurring at that time of year. The opportunity would be lost if the rig did not get off location immediately.

IMPROVED WPC STABILIZES THE WELL

If the well could be brought under control, the next options were either running a liner in the openhole section, or setting temporary abandonment plugs and returning later to complete the well. Before running the 7-in. liner, it was necessary to define the pressure gradient in Sand 1 (no pressure point data was available from offset wells), get a water sample and confirm reservoir continuity in Sand 2 and column height in the new reservoirs using formation testers.

To carry out either option, the well had to have a sufficient mud weight window. For example, to run the 7-in. liner, an ECD of 17.0 ppge was required. This meant an increase of about 0.6 ppge in WPC, often called the “apparent frac gradient (FG),” to sustain these pressures and avoid the likelihood of a liner top squeeze and remedial squeeze job. A solution was urgently needed to avoid undesirable cost overruns.

The selected treatment. WPC treatments react with drilling mud to create a pressure barrier at the face of the loss zone or over the exit zones in a crossflow. After reacting with the mud, a moldable sealant develops within 30 sec. The material forms a moldable, ductile, non-brittle pressure seal inside the fracture of the loss zone near the wellbore, which can seal by conforming to fracture faces as they change in width.^6–9 On the subject well, the main treatment consisted of 25 bbl of sealant material (9.2 lbm/gal) pumped down the drill pipe.

The plug was slowly displaced to the suspected loss zone with a final squeeze pressure of about 600 psi. After this displacement, the well was circulated out. The hole was slowly washed out with the bit pumping at a rate of 1 bbl/min. When the bit was run to 14,524 ft (4,427 m), the expected top of the sealant placement, the well was static and under control.

Results: Long-term effects of improved WPC. After the successful WPC treatment, a formation tester log run was made to get pressure points and wellbore fluid samples. The well was temporarily abandoned and not re-entered for seven months, making this the first case where a WPC treatment had been left in place for an extended time period with hole conditions of 11,000 psi and up to 300°F BHT.

After re-entry of the openhole section, a successful FIT of 17.0 ppge was achieved. No repeat treatment had to be implemented, and a 7-in. liner was run and cemented with no losses during the entire cement job. It was estimated that an ECD of 17 ppge was acting at the liner shoe depth during liner cementing. A liner packer was set on top to isolate the liner lap.

Pressure containment integrity of the near-wellbore region improved to 17 ppge from 16.4 ppge (an increment of 0.6 ppge), which was sufficient to proceed with the planned completion activity. When the mud weight window measurement is based on the lowest far field fracture gradient, the value exceeds a 2.0 ppge increase in WPC.

POST-WELL STUDY AND GEO-MECHANICAL ANALYSIS

In the offset well (Well 1), casing was set just above Sand 2 after mud losses were experienced when the sand was entered with 16.2 lbm/gal mud. The mud weight was reduced to 14.5 lbm/gal and the section was drilled without additional losses. The pressure in Sand 2 was measured with wireline MDTs to 14.1 ppge.

Sand 1 was not seen in Well 1, but seismic data suggested that it might be better developed in Well 2. Regional modeling indicated that Sand 1 was probably on the same pressure gradient as Sands 2 and 3 found in Well 1. Therefore, the 9-5/8-in. casing point was above Sand 1 in Well 2 to isolate the higher pressured Pliocene sands from the lower-pressured Serravalian gas sands, Fig. 2.

Fig. 2. Well schematic, final hole section. 10,11

As a result of the post-well study, a pore pressure (PP) and FG analysis was generated to compare PP and FG predictions vs. measured values. Pore pressure was measured in sands with a wireline formation pressure tester. A detailed PP/FG profile is shown in Fig. 3, which also contains plots of wireline formation PP test results from both Wells 1 and 2. Well 1 formation pressure data is plotted as red circles. Well 2 data is plotted as green circles. Because no pressure measurements were recorded over the bottom 300 m of the well, offset data was used to construct the final PP curve. The final PP curve presented in Fig. 3 was generated by combining predicted values through the shale intervals with measured values through sand intervals.

Fig. 3. Detailed pore and frac pressures.11

Note that the shale FG curve lays over the LOT at the base of the 9-5/8-in. shoe.

      For interpretation, the well took a kick during TD logging operations before losses occurred. The suspected interval for kick initiation was 13,517 to 13,648 ft (4,120 to 4,160 m), where PP exceeds hydrostatic pressure. Wireline pressure test results indicate the presence of a sand containing 16.17-ppge PP at 13,621 ft (4,151.8 m) TVD. The well was drilled with 16.0-lbm/gal mud through this section and was logged with 15.7-lbm/gal mud.

The effective mud weight was increased to about 17.0 ppge after the well kicked. Increasing the mud weight induced lost circulation and tensile fracture initiation, most likely over the bottom portion of the well where the predicted far-field FG was the lowest. As noted above, the well held a 17-ppge FIT after the WPC treatment. Red shading in Fig. 3 indicates where the computed far-field FG was less than 17.0 ppge. This shading highlights the intervals where tensile failure and lost circulation could occur with 17.0-ppge mud. This increase in wellbore containment pressure is due to the creation of a “stress cage” in the near-wellbore region as described by Alberty and McLean, 12 and Aston, et al.¹³

POST-WELL STUDY AND FDA MODELING

Planning the next well required a better understanding of the lost circulation conditions experienced in Well 2 and how the WPC treatment increased the fracture re-initiation pressure or the near-wellbore fracture gradient. A new Finite Difference Analysis (FDA) software model (DAP) was developed to help provide this understanding.

The following was modeled for Well 2:

• Lost circulation characteristics during drilling operations:
  • Loss-zone depths that took losses
  • Mud-loss volume and resulting fracture geometry
  • Effectiveness of different types of LCM pills, i.e., bridging particle pills and crosslinked gel pills
• WPC treatment design and results:
  • Sealant placement and propagation
  • Treatment pressures and rates
  • WPC treatment sealing effectiveness, and
  • Amount of sustained increase in wellbore FG.

The lost circulation model simulation¹¹ indicated the maximum fracture length and width created by the mud losses in different intervals, and most of the mud losses, were in the lowest depth loss zone at 14,700 to 14,728 ft (4,481 to 4,489 m) TVD.

Why did conventional treatments fail? The fracture width during the mud-loss period appears too large at 0.35 to 0.36 in. for effective near-wellbore bridging to occur by the 40 and 60-ppb LCM pills. Instead, the LCM particles continue to propagate the fracture for long distances, get diluted in the large lost-mud volume and fail to bridge.

The crosslinked gel pill (CLGP) modeling results, based on the time the CLGP enters and is completely placed in the fracture, show a fracture width too small to create the near-wellbore “wedge effect” needed to provide an effective pressure seal to halt the losses and increase near-wellbore fracture gradient.

The maximum created width of the CLGP was 0.23-in. in both dynamic and short-static conditions, less than the flowing mud loss created fracture width of 0.35 to 0.36 in. This may have allowed the mud flow to bypass the set CLGP inside the fracture. The apparent lack of any initial or sustained fracture width increase also helps explain why the CLGPs failed to allow circulation of the heavier mud needed to overbalance Sand 1.

Why did the WPC treatment work? The model simulation results in Figs. 4 and 5 are for the WPC treatment on Well 2. Fig. 4 presents WPC treatment parameters for squeeze pressures from 8:08 to 8:17 and “pack-and-bleed” section from 8:26 to 8:56. The latter depicts the increases in BHPs and fracture width vs. staged pumping and shutdown periods. When wellbore pressure was increased or decreased, fracture width also increased or decreased. This helps illustrate how the WPC sealing mechanism can sustain the increase in near-wellbore FG. The sealant form-fits to the changing fracture widths to continue sealing at different wellbore pressures that can be above the natural FG.

Fig. 4. WPC treatment squeeze placement and pack-bleed.11

The sealant shape may continue to change for an indefinite number of pressure cycles without leaking. WPC sealant placement inside the largest fracture loss-zone at 14,700 to 14,728 ft TVD is modeled in Fig. 5. Increasing fracture width near the wellbore is the sealant-induced “wedge effect,” indicating that a competent pressure seal was present that could contain wellbore pressures above the natural FG. The model also calculates the net pressure across each node in the 3-D grid with the node next to the wellbore predicting the amount of increased near-wellbore FG that may be provided, depending on formation/ sealant properties.

Fig. 5. WPC sealant's wedge placement creates a stress cage near the wellbore.11

Other characteristics and performance properties may help explain why the WPC treatment performed as planned, vs. poor results of other systems:

• Extremely low filtrate loss helps control filtrate losses and prevent skin damage.
• Fast sealing in wide cracks can provide immediate well control by plugging exit-flow zones.
• Limited penetration into fractures by agglutinate diameters larger than fracture widths.
• Extremely high extrusion pressures pack agglutinates to seal fractures near the wellbore.
• Self-diversion to other flow paths occurs after sealing the initial paths of least resistance.
• High resistance to swab/ surge pressures.
• Resistance to dilution by crossflows resulting from rapid reaction into high consistencies.

MORE HPHT APPLICATIONS

In a deep California HPHT gas exploration well, WPC treatments were used to help the operator avoid setting a drilling liner far above the desired depth.⁸ Setting the liner would have reduced the production hole diameter too much for a successful test and economic production of the newly discovered gas reservoir.

Applying the treatment increased wellbore pressure containment by more than 1.1 ppge above the natural FG and allowed an additional 960 ft to be drilled before a 7-in. liner was set and the well drilled to 19,724 ft. The operator was able to obtain full penetration through the objective sands, leading to subsequent discovery of a payzone. Had the operator decided not to apply the new WPC treatment to increase leakoff test results, the 7-in. liner would have been set early and the discovery zone may have been missed.

The trouble zone causing the low WPC was probably depleted because of an adjacent well blowout that produced up to 100 million scf/d gas for a lengthy period. The treatment solution saved the value of the multimillion-dollar well and allowed early gas production totaling several million dollars.

The cases presented here demonstrate that the widened mud weight window provided by WPC treatments can be sustained for a significant period of time, even under HPHT conditions. The WPC modeling can help the operator optimize well and casing design based on an improved mud weight window. As a result, it may be possible to eliminate casing strings, and overcome narrow mud weight windows to economically produce marginal HPHT reservoirs. 

LITERATURE CITED

¹ Marten, R. F. and M. V. Shann, “Seismic challenges of developing the Pre-Messinian Akhen field Offshore Nile Delta,” Paper presented at the 2002 AAPG Conference, Cairo, Egypt, October 27 – 30.

² Alberty, M. W., et al., “Regional pressure study in the Deep Nile Delta,” paper presented at the 2002 AAPG Conference, Cairo, Egypt, October 27 – 30.

³ Sweatman, R. E., C. W. Kessler and J. M. Hillier, “New solutions to remedy lost circulation, crossflows, and underground blowouts,” Paper SPE 37671, presented at the 1997 SPE/IADC Drilling Conference, Amsterdam, March 4 – 6.

⁴ Sweatman, R., R. Faul and C. Ballew, “New solutions for subsalt-well lost circulation and optimized primary cementing,” Paper SPE 56499, presented at the 1999 Annual Technical Conference, Houston, October 3 – 6.

⁵ Rueda, F. and R. Bonifacio, “In-situ reactive system stops lost circulation and underground flow problems in several Southern Mexico wells,” Paper SPE 59059, presented at the 2000 International Petroleum Conference and Exhibition, Villahermosa, Mexico, February 1 – 3.

⁶ Kelley, S., R. Sweatman, and J. Heathman, “Treatments increase formation pressure integrity in HTHP wells,” Paper AADE 01- NC-HO-42 presented at the 2001 AADE National Drilling Technology Conference, Houston, March 27 – 29.

⁷ Sweatman, R., S. Kelley and J. Heathman, “Formation pressure integrity treatments optimize drilling and completion of HTHP production hole sections,” Paper SPE 68946 presented at the 2001 European Formation Damage Conference, The Hague, the Netherlands, May 21 – 22.

⁸ Webb, S., et al., “New treatments substantially increase LOT/FIT pressures to solve deep HTHP drilling challenges,” Paper SPE 71390 presented at the 2001 Technical Conference and Exhibition, New Orleans, Louisiana, September 30 – October 3.

⁹ Wang, H., R. Sweatman and W. Deeg, “The difference between FG and wellbore pressure containment and the effect on drilling beyond natural pressure limits,” Paper AADE-03-NTCE-26, presented at the 2003 AADE National Technology Conference and Exhibition, Houston, Texas, April 1 – 3. 

¹⁰ Sanad, M., et al., “A new treatment increases the frac gradient to cure lost circulation and control a flowing HPHT well in the East Mediterranean Sea,” Paper SPE 84318, presented at the 2003 SPE Annual Technical Conference and Exhibit in Denver, Colorado, October 5 – 8.

¹¹ Sanad, M., et al., “Numerical models help analyze lost-circulation/ flow events and control an HPHT well in the East Mediterranean Sea,” Paper IADC/SPE 87094 presented at the 2004 IADC/SPE Drilling Conference in Dallas, Texas, March 2 – 4.

¹² Alberty, M. W. and M. R. McLean, “Fracture gradients in depleted reservoirs – Drilling wells in late reservoir life,” Paper SPE/IADC 67740 presented at the 2001 Drilling Conference, Amsterdam, February 27 – March 1.

¹³ Aston, M. S., et al., “Drilling fluids for wellbore strengthening,” Paper IADC/SPE 87130 presented at the 2004 IADC/SPE Drilling Conference, Dallas, Texas, March 2 – 4.