Frac-gradient enhancement controls lost circulation for HPHT well in East Mediterranean Sea

In its drilling/production operations in the offshore Nile Delta of the Mediterranean Sea, BP-GUPCO has employed a fracture gradient enhancement squeeze system to halt loss of circulation caused by a pressured crossflow.

The formations also show a tendency to balloon during tripping out of hole, especially after mud has been lost to a Sand 2 zone. In many cases, the formation starts to pack off around the drill pipe and makes tripping out difficult. Swabbing in lost wellbore fluids, or swabbing in formation fluids, can complicate the assessment of well control while tripping.

Drilling problems

In the offset well (Well 1), the drilling team set casing just above Sand 2 after mud was lost when entering the sand with 16.2 lb/gal mud. Subsequently, they reduced MW to 14.5 lb/gal and the team drilled the regressive section without additional losses. They measured the pressure in Sand 2 with wireline modular formation dynamics testers at 14.1 lb/gal. They did not see Sand 1 in Well No. 1. During the planning phase, seismic imaging suggested that Sand 1 might be better developed at the Well 2 location. Regional modeling indicated that Sand 1 was probably on the same pressure gradient as Sands 2 and 3 found in Well No. 1. Based on this interpretation, the team set casing (9 5/8 in.) above Sand 1 in Well No. 2 to isolate the higher pressured Pliocene sands from the lower pressured Serravalian gas sands.

In Well 2, the team landed the 9 5/8 in. liner at 13,445 ft TVD to isolate the high-pressure A-60 and A-70 Pliocene sands from the lower pressured Serravalian objective sands. A wet shoe track led to concerns about the effectiveness of the cement job, but a 17.8-lb/gal leakoff test at the shoe and a change in the C3:C4 ratio of gas readings before and after running pipe confirmed that the previous interval had been isolated.

The original plan was to reduce MW from 16.3 to 14.5 lb/gal to drill Sands 1, 2, and 3. The drilling team cut the MW to 15.7 lb/gal during drillout; however, returns from the shale drilled below the 9 5/8-in. shoe contained 700 units of gas with mud bubbling over the bell nipple. As a result, the team increased the MW from 15.7 to 16 lb/gal and drilling started with a max ECD of 17.1 lb/gal at a flow rate of 425 gal/min.

The team observed no losses drilling Sand 1, but as they penetrated Sand 2, mud losses began at 10-20 bbl/hr. The team reduced MW to 15.7 lb/gal to continue drilling for ECD control and raise it back to 15.9 lb/gal before tripping. Although pumping out was required, they made several trips without incident.

The drilling team used conventional lost-circulation material (LCM) pills along with control drilling and stepped reductions in the flow rate to manage loss rates up to 120 bbl/hr as they continued drilling through Sand 2. However, they stopped drilling to log and evaluate the wellbore at 14,852 ft TVD (Sand 3) where they experienced complete loss of circulation. At this point, the drilling team reduced the flow rate from 320 to 168 gal/min, which they deemed too low to effectively clean the 40° wellbore. They lost 1,300 bbl of oil mud to this point.

Before pulling out of the hole to log, the team increased MW to 15.9 lb/gal and made several short trips into the 9 5/8 in. liner to ensure that the well was stable. The swabbing action of the drill string when pulling pipe exacerbated the ballooning tendency. They spent significant time circulating bottoms-up and monitoring flow at the trip tank with pumps off to determine the nature of the flow. When convinced that ballooning was occurring and that the well was not flowing, the team pumped the drill string out of the hole in stages, checking frequently to confirm that a well-control situation was not developing.

Monitoring the well on the trip tank, the first two logging runs proceeded uneventfully. During the third logging run, the well began flowing and the drilling team shut it in at the surface with 120 psi. To reduce casing pressure so that drillpipe could be stripped in the hole, they bullheaded 17 lb/gal mud into the well. Injection pressures confirmed earlier ECD calculations that the well was taking fluid at 16.3 ppge. After stripping drillpipe to 5,612 ft MD, the team circulated 18 lb/gal mud to kill the well. The well had complete losses. They then ran drillpipe in the hole to the 9 5/8 in. shoe where the mud was conditioned to the original density of 16.2 lbm/gal with substantial losses.

Resistivity trends from multiple passes with the logging-while-drilling tool implied that losses were occurring at numerous points in the 8 1/2-in. open hole below the top of Sand 2. At this point, almost 4,000 bbl of mud had been lost in this interval. The drilling team pumped several LCM pills containing 40-60 lbm/bbl of mixed, fibrous particulate and platelet-type additives with little success. Following this effort, they pumped two 40 bbl crosslinked (CL) polymer pills. After waiting the prescribed time for the pills to set, loss rates remained unchanged. The team selected FGESS to enable tripping and subsequent liner installation operations while minimizing the risk of losing the openhole section.

Understanding the problem

Before losing circulation, the well took a kick during total depth logging operations. The highest pore pressure (PP) over the bottom open-hole interval exists from casing shoe to approximately 14,173 ft. The most likely interval for kick initiation was 13,517-13,648 ft, where PP exceeds hydrostatic pressure. Wireline pressure test results indicate the presence of sand containing 16.17 lb/gal PP at 13,621 ft TVD. This was the highest measured and suspected PP over the entire open-hole interval. The team drilled the well with 16-lb/gal mud through this section and logged it with 15.7 lb/gal mud.

Experience with a recently completed offset well (Well 1) helped operators anticipate lost-circulation troubles in the current well (Well 2). The pay zones are labeled Sand 1 and Sand 2.

They increased the effective MW to approximately 17 lb/gal after the well kicked. Increasing the MW induced lost-circulation and tensile fracture initiation, most likely over the bottom portion of the well. The well held a 17-lb/gal formation integrity test (FIT) after treatment. Observed increase in wellbore pressure containment or FRIP/NWFG most likely occurred in the near-wellbore region as a result of the FGESS treatment. The pressure trend in this well was consistent with the trend observed in the offset wells, where the lower sands were at a lower pressure than the upper sands and the shales. This PP regression is a key factor leading to the lowest, far-field fracture gradient (FG) calculation of 15 lbm/gal equivalent found in the interval 14,698 ft (4,480 m) to TD.

A closer look

First it was essential to get the well under control. The next step would either be to run a liner in the open-hole section or set temporary abandonment plugs and return at a later date to complete the well. Before making the decision on running the 7-in. liner, the drilling team had to define the pressure gradient in Sand 1, get a water sample, and confirm reservoir continuity in Sand 2 and column height in the new reservoirs using formation testers. Application of FGESS was recommended for this well.

FGESS treatments react with drilling mud to create a pressure barrier at the face of the lost-circulation zone or the exit zones in a cross flowing hole by short propagations of sealants within fractures and faults. After reacting with the mud, the FGESS typically develops into a sealant with a moldable consistency within 30 seconds. The consistency allows the material to form 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.

Before a squeeze placement into leak-off flow paths in open-hole formations, the FGESS slurry co-mingles with the well fluid below the bit and above the formation leak-off flow paths. The slurry may also intersect and mix with formation fluids flowing from an influx zone into loss or exit zones. Commonly found ions in the well and/or flowing formation fluid trigger chemical reactions. These change the slurry and mud or formation fluid mixture from easily pumped viscosities into extremely high viscosities that finger through the well and/or formation fluid before entering the loss or exit formation(s). The fingers aggregate to form strings of semi-solid agglutinates. The agglutinates are often larger than the openings into formations such as fractures, faults, and small vugs. The agglutinates cannot easily flow into the formation openings and are extruded under pressure to propagate a short distance while simultaneously agglomerating into a flexible seal.

Sequential sealing of openings then occurs from the initial flow path of least resistance to each subsequent flow path due to the agglutinates’ self-diverting properties. Moldable and high cohesive properties cause the agglutinates to form-fit in openings and maintain a seal under designed pressure differentials, controlled swab-surge loads, and HTHP conditions. Pay zones sealed during drilling do not impair production in cased-hole completions because the agglutinates have almost zero fluid loss and less than 1/8 in. penetration in Darcy permeability core tests.

The drilling team placed the bit at the 9 5/8 in. casing shoe to have access to all potential loss zones. The suspected lost-circulation (LC) points were approximately 14,928 ft MD (close to Sand 2) where first losses occurred. However, deeper zones could not be ruled out. The wellbore was imbalanced before the FGESS process; various MWs with possibly water-cut or gas-cut mud were present in the system.

Because a surface sample of the influx fluid was not available, it was hard to establish what the influx fluid was. Total losses would occur if the mud pumps were turned on, whereas pressure buildup on the drillpipe and casing would occur if they were left alone. The volume of the FGESS treatment was based on the 8-1/2 in. hole washed out to 9 in. The main treatment consisted of 25 bbl of FGESS the team pumped down the drillpipe. They made plans to pump mud and FGESS at a 1:1 ratio to form a plug of 50 bbl. The team pumped 1,000 ft of drillpipe length of compatible spacer (18 bbl) ahead and behind the FGESS to keep it separate from the mud.

In preparation for the FGESS job, the operators conducted on-site compatibility checks among the mud, the FGESS, and the spacer. As the FGESS started to mix downhole with the mud, the injection pressure began to rise and reached the 1,000-psi limitation set at the 9 5/8 in. casing shoe. In reality, the team pumped a ratio of 0.5:1 (mud: FGESS), which provided a smaller amount of FGESS material for increasing borehole pressure containment integrity. They slowly displaced the plug to the suspected lost-circulation zone with a final squeeze pressure of approximately 600 psi. After this displacement the well was circulated out. The mud returns were in the range of 12.1 lb/gal to 16.3 lb/gal and finally settled at 16.3 lb/gal (the pump-in MW). The team slowly washed out the hole with the bit pumping at a rate of 1 bbl/min. The bit ran to 14,524 ft (4 ft below the base of Sand 1), the suspected top of FGESS. The well was static and under control at this point.

Well planning

Planning future wells will require a better understanding of the LC conditions experienced in Well No. 2 and how the FGESS increased the FRIP or NWFG. A new finite difference analysis (FDA) software model will help provide this understanding. The FDA model is a grid-oriented, planar 3D hydraulic fracturing simulator designed to provide information and analysis.

In a post-job analysis, the FDA model’s LC simulation showed the maximum fracture length and width created by the mud losses. Another model calculation display indicated that the mud losses propagated several weak zones. The model’s calculations also indicated that most of the mud losses were in the lowest depth loss zone at 14,700-14,728 ft TVD. A timescale also showed the increases in bottomhole pressures and fracture width versus staged pumping and shutdown periods. When the wellbore pressure was increased or decreased, the fracture width also increased or decreased. This illustrated the FGESS sealing mechanism to sustain the increase in near-wellbore FG. The sealant can form-fit to the changing fracture widths to continue sealing at different wellbore pressures that can be above the natural FG.

These changes in sealant shape may continue for an indefinite number of pressure cycles without leaking. The increasing fracture width near the wellbore is the sealant-induced wedge effect. This indicates 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 grid with the node next to the wellbore predicting the amount of increased NWFG that may be provided, depending on formation and sealant properties.

Long-term effects of FGESS

After the successful FGESS treatment, the team made a formation tester log run to get pressure points and wellbore fluid samples. Following the log run, they temporarily abandoned the well for seven months for a re-entry and completion at a later date. It was not known before the well intervention whether the FGESS would stand up in an HPHT gas well with hole conditions of 11,000 psi and 270-285° F bottomhole temperature at the treated interval. It was not clear whether LC or an underground crossflow would occur again as was the case before the FGESS job in November 2001. However, prior lab tests performed with the FGESS material demonstrated favorable long-term results from the treatment.

On re-entry of the open-hole section, the drilling team carried out a formation integrity test (FIT) with 16.3-lb/gal mud. The open hole tested to 17 lb/ge. After seven months, the FGESS treatment was still holding up to significant wellbore pressures. The team did not have to repeat treatment; they made a conditioning trip to a point just above the treated zone depth and encountered no losses or flows. They then ran and cemented a 7 in. liner through Sand 1 with no losses during the entire cement job. They estimated that an ECD of 17 lb/gal was acting at the liner shoe depth. The team set a liner packer on top of the liner to isolate the liner lap.

The pressure containment integrity of the near-wellbore region conclusively improved from 16.4 lb/gal to 17 lb/gal (an increment of 0.6 lb/gal), which was sufficient for them to proceed with the planned completion activity.

Applying system to future wells

FGESS can effectively control lost circulation and crossflows in a single treatment in an underground blowout situation. In addition to stopping LC, this system helped save a well and allowed continued drilling operations without typical waiting times such as those for cement-squeeze or CL gel-pill setting times. In total, the treatment created a combined savings of over $1 million.

The widened MW window FGESS enabled can be sustained for at least seven months. This process can provide an effective way to deal with challenging HPHT wells where geo-pressure predictions may not be simple, and a better use of casing strings can be implemented.

A PP/FG prediction model and formation test indicated the likely zones with flow/loss potential, served as input to define log-based stresses, and allowed these formation stresses to be digitally read into the new LC and NWFG enhancement FDA model. This new FDA modeling software can help predict and post-well evaluate LC conditions, FGESS treatment results, and the effectiveness of other types of LC treatments.

Finally, modeling potential LC and FRIP/NWFG increases may help design wells with substantially reduced well construction costs to justify the use of FGESS technology.