Offshore Petroleum Operations Minimizing surface hole washout helped prevent conductor pipe failure

Brian Kelly Talkington
Phillips Petroleum International Corp. Asia
Shekou, China

Roger L. Thomas
Phillips Petroleum Co.
Bartlesville, Okla.

Earl H. Doyle
Shell Offshore Inc.
Houston

The failure of a driven conductor in the Xijiang 24-3 field in the South China Sea led to changes in the drilling and cementing procedures for the surface holes in subsequent wells.

The conductor failure problem was solved by minimizing washout during drilling of the 171/2-in. hole and by ensuring that the top of the cement for the surface casing was inside the 24-in. 3 133/8-in. annulus.

On Aug. 12, 1994, during drilling of the 121/4-in. hole on well A1, the first development well from the Xijiang 24-3 platform in the South China Sea, the well's predriven 24-in. conductor casing fell to such a depth that the top of the pipe was below the mud line. This failure presented a potential well control problem and created a significant risk to the future of the Xijiang development.

The conductor failure problem was eventually overcome without using costly options, such as driving additional casing. Control of Well A1, although in jeopardy for a period of time, was never lost. There has not been any additional conductor slippage since then.

The 24-3 platform, installed in April and May 1994, was the first of two platforms planned for the Xijiang development. The second platform was set in April and May 1995 to exploit the 30-2 field. Techniques used on the 24-3 platform to overcome the conductor failure problem will be used on the second platform where drilling is already under way.

Xijiang development

The Ppica Group consists of Phillips Petroleum International Corp. Asia and its coventurers, Conhe and Pecten Orient. Ppica is currently developing the Xijiang 24-3 and 30-2 fields. They are 8 miles apart and approximately 80 miles southeast of Hong Kong in the South China Sea (Fig. 1 [33681 bytes]).

The water depth at the platforms is 325 ft. Production from the two fields is piped to a floating production storage and offtake (FPSO) vessel positioned between the platforms (Fig. 2 [33777 bytes]).

Ppica has been exploring for hydrocarbons in the South China Sea since 1983. The Group drilled 16 exploration/appraisal wells before determining that a commercial development existed. Of these wells, three were drilled in the 24-3 field and two in the 30-2 field.

The predevelopment wells were drilled with semisubmersible rigs. The general casing program for 15 of these wells included 30-in., 20-in., 133/8-in., and 95/8-in. casing. For the 16th well, the 20-in. casing was eliminated. The 30-in. structural/conductor casing that was run in each of these wells was set in a 36-in. hole and cemented back to the mud line. The hole for the surface casing in each well was drilled with returns taken at the seabed. (That is, a pin connector/riser was never used.)

Fig. 3 [99965 bytes] shows the casing program for the development wells. Note that the 135/8-in. wellhead, which makes up to the top joint of the 133/8-in. casing, lands on top of the 24-in. conductor casing. The 95/8-in. casing and 41/2-in. tubing hangers land in the wellhead. Therefore, upon landing of the 41/2-in. tubing hanger, the load distributed to the conductor casing included the weight of all the casings, tubing, wellhead, riser, and blowout preventers (BOPs).

Site-specific soils

The site-specific soils investigation for Xijiang 24-3 was conducted during the summer of 1988. Two borings were taken at the platform location. B-1 consisted of continuous cone penetration testing, and B-2 consisted of sampling the subsurface soils with a 3-in. diameter sampling tube pushed into the soil.

Samples were taken continuously from the mud line to 40 ft penetration, at 5-ft intervals from 40 to 60 ft, and at 10-ft intervals or at changes in soil profile to 417 ft of penetration. No unusual or abnormal soils were noted by the contractor in its final report.

Based on field and laboratory testing, the contractor developed the recommended design conditions (Fig. 4 [122661 bytes]). Using these design values and following the American Petroleum Institute (API) Recommended Practice 2A, "Planning, Design, Construction, and Installation of Fixed Offshore Platforms," the unit skin friction in kips per square foot of pile was calculated.

Conductor design

The first step in design of the conductors was to determine the maximum axial load that would be placed on the conductor. The maximum top load, approximately 763 kips, that would be applied to the conductors would be the cumulative weight of all the casing strings, BOPs, wellhead, and riser.

A safety factor of two was applied, resulting in a design vertical loading of 1,526 kips.

Using the recommended skin friction, the axial capacity of the 24-in. OD conductors vs. depth was calculated. Because the conductors were to be drilled out, no tip loading was included.

Calculating the pile side shear determined a pile penetration of 161 ft was necessary for the design vertical load of 1,572 kips. This design load was not intended to depend on support from the 133/8-in. casing cement (that is, grouting inside the 24-in. 3 133/8-in. annulus). The final depth of penetration of 175 ft was selected.

24-3 conductor study

A pile driveability study determined if the 24-in. conductors could be installed to the desired penetration. The aim was to predict the expected blow count vs. depth and to determine the stresses induced into the conductor by the pile driving operation.

A commercially available computer program (Grlweap) performed the calculations. Default program values were used for the analyses to describe the hammer, soil quake, and soil damping. The soil resistance to driving was modeled by assuming the skin friction acted only along the outside of the conductor and was equal to 50% of the unit skin friction in clay and 70% in the sand.

The conductor was also assumed to drive unplugged. Thus, soil end bearing effects were applied only to the exposed steel area at the conductor tip.

Fig. 5 [39024 bytes] shows the predicted drivability for the Menck MRBS-1800 steam hammer used in the installation and the actual driving record for the conductor that failed. The predicted blow counts were in reasonable agreement with the measured values except that below about 150 ft, the measured values were higher than predicted.

These higher blow counts could have been due to an unexpected dense sand layer that thickens from one corner of the platform to the other. Cone penetrator tests (CPT) were performed in another borehole some 250 ft from the sample boring taken from the platform location. The CPT borehole showed a dense sand layer from 147 ft to about 165 ft below the mud line.

Although the high-resolution geophysical data showed this layer dipped sharply away from the center of the structure, it is possible that some of the conductors might have been driven through some part of this dense sand layer. Several factors other than soil stratigraphy could also explain the higher blow counts. Reduced steam pressure and deteriorating hammer cushion could have reduced hammer efficiency, or the conductor could have become plugged with soil.

Common practice in exploration wells is to drive the conductor pipe until refusal. Some operators will then drill out the plug inside the casing and then drive to refusal again. It is assumed that at refusal, the maximum load capacity of the conductor has been attained. The widely accepted method of determining pile capacity is by calculation of the pile side shear, however. Driveability studies are only performed to determine if the pile can be installed to the required depth and to ensure that the steel will not be overstressed during driving.

Based on standard soil mechanics and offshore engineering practices, the conductor penetration of 175 ft was correct for the 24-3 field. Using the same principles, it was determined that the conductor penetration for the 30-2 field should be 205 ft. This depth of penetration was not changed as a result of the failure of the Well A1 conductor.

Conductor driving

The 24-3 platform has 24 well slots arranged in a 4 3 6 pattern with 7-ft centers (Fig. 6 [63360 bytes]). The original intention was to drill 16 production wells in this field. Sixteen 24-in. OD 3 1-in. wall thickness conductor strings were batch driven using the derrick barge during the installation of the platform, prior to setting of the topsides and the drilling rig. All 16 conductors were driven to the desired depth of penetration in 3 days. Upon installation of the topsides, the tops of the conductors were each one foot above the production deck (91-ft elevation).

The absence of shallow gas in the Xijiang fields, as confirmed by drilling the exploration/appraisal wells, allowed the use of subsea circulating ports in the conductors. On 24-3, these ports were positioned at 117 ft below sealevel (289 ft from the rotary kelly bushing, or RKB). Returns while drilling the 171/2-in. hole and cementing the 133/8-in. casing were taken through these ports. Using the subsea ports alleviated potential hydrostatic problems and eliminated the need for rigging up risers, diverters, and flow lines; however, it was impossible to monitor returns.

Well A1

Well A1 was spudded on Aug. 8, 1994. A 171/2-in. surface hole was drilled from 672 ft (24-in. shoe) to 1,538 ft RKB using seawater with frequent high-viscosity sweeps to clean the hole. The 133/8-in. casing was then run, with the shoe at 1,515 ft. A 135/8-in. thread-on-type compact wellhead housing had been made up to the top joint of this string, and the weight of the string landed on top of the 24-in. conductor.

The casing was cemented with 171 bbl of lead and 136 bbl of tail cement. No verification of the top of cement was attempted. Subsequent to the failure of the conductor, it was assumed that the 24-in. 3 133/8-in. annulus was void of cement.

The 135/8-in. riser and BOPs were nippled up and tested prior to drilling out the 133/8-in. shoe. After the shoe track and 10 ft of new formation were drilled out, an actual leak-off test was taken to 11.5 ppg equivalent mud weight. During drilling at approximately 3,585 ft, the conductor casing failed under its own weight and fell below sea level and out of sight.

Failure of the conductor left the 133/8-in. surface casing still standing with the wellhead above the production deck. The string was exposed to environmental and buckling loads that it was not designed for, however. Drilling was halted immediately, and the well was secured. A cement plug was placed from 100 ft below the 133/8-in. casing shoe to 200 ft inside the casing. After the proper equipment was mobilized, the 133/8-in. casing was cut and retrieved from 529 ft (30 ft below the seabed). At this point Well A1 was considered permanently abandoned.

The consensus was that, during drilling of the 171/2-in. hole on Well A1, the hole below the conductor was severely washed out in the silty clay formation. This washout propagated around the shoe, disturbing the soil and lowering the side shear of the pile, causing the casing to fail. The washout did not immediately cause failure of the 24-in. conductor, but it left a void behind and below the casing into which the silty clay could fall during subsequent drilling. Thus, the conductor failed later than when the load was initially placed on it.

A washout of such magnitude was completely unexpected because the drilling parameters used while drilling the 171/2-in. hole section on Well A1 were not aggressive. The circulation rate was maintained between 500 and 1,000 gpm (the lower rate was due to start-up problems with the rig equipment during drilling of the upper section of the hole), and the rate of penetration never exceeded 100 ft/hr. The total time expended from spud until the 133/8-in. casing was landed and cemented was 48 hr.

A remotely operated vehicle (ROV) was mobilized some days later to assess any possible damage to the jacket and to determine how far the conductor had dropped. No damage to the jacket was detected. The ROV could not see the top of the conductor. This means that the conductor had fallen at least 419 ft such that its top was below the mud line.

Solution

Several options to continue were considered:

1. Drive replacement conductors in the remaining empty slots.

2. Drive the existing conductors deeper.

3. Add an additional 185/8-in. string to the casing program.

4. Drill the 171/2-in. hole on another slot then run and cement 133/8-in. casing.

The fourth was chosen. For this to be successful, it was believed mandatory that the 133/8-in. cement operation bring the top of cement inside the 24-in. 3 133/8-in. annulus. The cement would fill the void created by the washout, thereby halting further propagation of the soil around the conductor. In addition, the cement inside the annulus would lend structural support to the conductor.

As a contingency, driving equipment was mobilized from Singapore, and drive pipe was made ready to perform the first or second option in case this operation failed.

Several precautionary steps were taken before drilling ahead on the next well. First, it was decided that this attempt would be made in the slot farthest from Well A1. This slot was for Well A3 in the southwest corner of the platform. It was spudded on Aug. 15. A 171/2-in. hole was drilled from 672 to 772 ft using seawater and high-viscosity sweeps. A cement plug was then set in the open hole to give some support to the conductor string and protect the hole. The thinking was that when the cement was drilled out with a 171/2-in. bit, some of the cement would remain, protecting the hole from additional washout.

The height of the A3 conductor above the production deck was constantly monitored during the drilling operation. Prior to drilling, the top of the conductor was 125/8 in. above the production deck. After drilling but before the cement plug was set, the top was 145/8 in. above the production deck.

After the cement plug was set in Well A3, the competency of the conductors in each of the slots surrounding Well A1 was checked. The formations around the A1 well bore could have been washed out so severely as to affect the surrounding conductors.

Using a modified casing spear, 500-kip pull tests were performed on each of these five conductor casings. The load was applied in 50 kip increments with the full 500 kip load being held for 30 min.

Each string withstood the pull test, stretching between 1.25 and 1.55 in. (theoretical stretch was 1.1 in.). All strings returned to their original position when the load was removed. These tests proved that the axial capacities of these conductors were adequate and not affected by the failure on Well A1.

The rig was skidded back to Well A3 on Aug. 22. Before drilling operations commenced, four support gussets were welded to the conductor. These gussets were positioned about 2 in. above the platform's production deck. The gussets would land on the production deck if the conductor slipped, preventing it from falling farther. The production deck was not designed to withstand much more than the weight of a single conductor; therefore, the gussets had to be removed after the 133/8-in. casing was set and cemented.

A 171/2-in. bit was used to drill out the cement plug to 771 ft. The height of the conductor above the production deck increased 3/4 in. during this drilling. After the bit was pulled out of the hole, a successful 500-kip pull test was performed on the conductor. The 171/2-in. hole was then drilled to casing point at 1,538 ft using seawater and frequent high-viscosity flushes. Throughout this section, the circulation rate was maintained at 550 gpm. The average penetration rate was 90 ft/hr. The hole was displaced with high-viscosity mud prior to pulling the bit out. Again, another successful 500-kip pull test was performed on the conductor.

Next, a caliper log (range 16-22 in.) was run through the open hole. The caliper arms were fully extended from a depth of 757 ft to the 24-in. shoe. A wiper trip was then made prior to running casing. The hole was displaced again with high-viscosity mud during this trip. Circulation was kept to a minimum to avoid further washing out of the hole. The 133/8-in. casing was then run with the shoe set at 1,499 ft. The wellhead was landed on top of the conductor. The casing was then cemented with 350 bbl of lead (including mica to prevent losses) and 70 bbl of tail cement (about 40% more total cement than was pumped on Well A1).

Results of a temperature log to determine the top of cement were inconclusive. The top was confirmed by running a 1-in. bolt tied to a rope down the 24-in. 3 133/8-in. annulus. The bolt tagged cement at 327 ft (about 38 ft below the subsea ports). The bolt, when retrieved, had "green" cement stuck to the threads, further confirming that the top of cement had been tagged. The "bolt on a rope" was used on all future 133/8-in. cement jobs to confirm the top of cement.

The rig was then skidded to Well A1A (the replacement well for A1) to set the surface casing. Based on the results obtained in Well A3, the procedure was modified to include drilling the first 100 ft of new hole with 9-ppg mud to limit washout. After 100 ft, drilling was continued with seawater and frequent high-viscosity sweeps. (It was impractical to drill a longer section of hole with mud because all returns were lost through the subsea ports.) In addition, the initial cement plug to stabilize the hole was eliminated. The rest of the procedure was the same as that used on Well A3.

The caliper log run after the first 100 ft of hole was drilled showed the hole to be washed out greater than 22 in. for only 5-10 ft below the conductor shoe. The washout then tapered back to 171/2 in. to the bottom of the hole. The second caliper (range 15-291/2 in.) run after reaching casing point showed the hole to be washed out greater than 291/2 in. from 1,276 ft to 951 ft. From 951 ft to the 24 in. shoe, the hole ranged from 25 in. to 291/2 in. The conclusion was that drilling the first 100 ft with mud did little good, as the hole was later washed out with the seawater.

The casing was run and then cemented with 630 bbl of lead and 70 bbl of tail cement. A temperature log was run; however, the top of cement was confirmed with the bolt-on-a-rope to be at 520 ft RKB (152 ft above the conductor shoe).

The rig was then skidded to Well A4. The procedure for this well was the same as that for Well A1A, except that the cement volume was reduced to 400 bbl of lead and 70 bbl of tail. The temperature log was not run. The bolt-on-a-rope confirmed the top of cement in the 24-in. 3 133/8-in. annulus at 410 ft RKB (262 ft above the conductor shoe).

Following the successful completion of the competency tests on the conductors and successfully setting these three surface casings, cementing them into the 24-in. 3 133/8-in. annulus, Ppica felt confident that it was safe to continue with the field development.

The rig was skidded back to Well A3 on Aug. 31, 17 days after the failure on A1, to continue drilling and completion operations. Before drilling commenced, a cement bond log was run on the 133/8-in. casing to confirm the top of cement. Operations continued on the well until it was completed for production on Oct. 12.

As more experience was gained, the procedure for setting the remaining surface casings was eventually modified to eliminate the pull tests, the caliper logs, and the support gussets. The main steps of the modified procedure remained slow drilling and circulation with frequent sweeps and pumping significant excess cement to ensure that the top of cement was inside the 24-in. 3 133/8-in. annulus.

By May 3, 1995, 12 wells had been drilled, and 10 wells completed for production. This ended the first phase of drilling and completion activities on this platform. Current plans are to complete the remaining two wells during the summer of 1996.

Results

The conductor failure problem was solved by taking steps to minimize washout during drilling of the 171/2-in. hole and by ensuring that the top of cement for the surface casing was inside the 24-in. 3 133/8-in. annulus.

The 24-3 conductor design, including the penetration of 175 ft, was correct.

By using careful step-by-step procedures, the problem was solved without resorting to more expensive options.

The Authors