Halliburton officials defend Macondo well's cement job

Oct. 4, 2010
Cementing of BP PLC's ill-fated Macondo well in the Gulf of Mexico took place after extensive tests, followed the operator's specifications, and showed no signs of problems in the hours before the well blew out on Apr. 20 and its semisubmersible exploded, killing 11 workers, Halliburton Co. officials told a federal panel on Sept. 26.

Cementing of BP PLC's ill-fated Macondo well in the Gulf of Mexico took place after extensive tests, followed the operator's specifications, and showed no signs of problems in the hours before the well blew out on Apr. 20 and its semisubmersible exploded, killing 11 workers, Halliburton Co. officials told a federal panel on Sept. 26.

They said BP's internal investigation has provided the most information so far, but disagreed with its conclusion, announced on Sept. 8, that problems began with a bad cement job. The event caused a massive oil leak into the gulf which has taken months to stop and clean up.

"There were a number of red flags which should have given pause to BP, as the well operator. Notwithstanding multiple signals of problems, it did not adjust its operations," Thomas Roth, Halliburton's vice-president of cementing, told the National Academy of Engineering's committee that is examining the accident and spill and preparing a report for US Interior Sec. Ken Salazar.

Roth said Halliburton, as one of several contractors working for BP at the well site, did not have access to all the information about what was going on there. "We knew the mechanics of the well and its depth, but not specific reservoir information. We were provided specific details of the tubulars," Roth said. Halliburton also knew that BP was suspending drilling before it hit the well's planned 19,000 ft depth, he said.

Problems with the cement could have been traced to its possible contamination, incomplete laboratory testing beforehand, or use of an unstable foam slurry which would have resulted in nitrogen breakout, none of which apparently occurred, he explained. Roth said Halliburton supplied the cement based on BP's specifications, tested it in its own laboratory, and recommended a formulation based on that information. Tests took more than 400 hr and indicated that the foam system was stable on delivery, he said.

Series of tests

"When cement is prepared on location, we go through the actions of foaming, starting with the base slurry," Roth said. "We measure the density to determine the cement-water ratios are appropriate for the design and the foam system is consistent over the cement. We also monitor the nitrogen pressure and nitrogen rate to provide assure that the job is executed according to the design. The material balance is identified ahead and after the job to identify what's used in the job. This analysis shows that the materials used followed the plan at the Macondo well."

Halliburton performed three tests on the cement which would be used in the Macondo well and discussed the results with BP, Roth said. The foam stability test used a sample of the cement slurry, which was mixed in a high-speed Waring blender to compare it to test specifications, then cured in a water bath for 48 hr, where the cement hydrated. A specimen was gathered and examined to see if there is any separation, settling, or breakup. "On the identified specimen, there was good stability," Roth said.

BP said in the report of its investigation that it had to simulate the cement's condition because Halliburton would not provide samples. Roth said that the independent laboratory's conclusions were inaccurate because of this, and because the lab used a 3-blade blending system instead of one with 5 blades as specified in American Petroleum Institute standards.

Indications of channeling from using only six centralizers did not cause alarm at the Macondo well, he said. "We end up with channeled cement jobs in many instances, which are not inherently unsafe in the Gulf of Mexico," said Roth. "They commonly exist on work provided in the field. It's not seen as a red flag going forward."

John Gisclair, in-site support coordinator for Halliburton's Sperry Drilling division that also worked on the Macondo well, said that while various contractors are not required to share information, they often do "because we want to get the job done well." He said, "Typically, they share critical data. The Horizon had a closed-circuit TV system with a number of cameras throughout the rig so it was possible to see Sperry's mud-logging displays on one channel and Transocean's data on another."

Sensors bypassed

Multiple activities aboard a rig during a planned shutdown can keep some on board from knowing what's going on everywhere, Gisclair continued. "Because multiple rig activities affect the same data, it can become impossible to determine which activities are affecting the data and which are affecting the hole condition," he said. Once BP's crew determined there was a problem, it shut down the well's pumps, determined that the fluid inside the annulus could acceptable be dumped, and began that process, which bypassed several sensors, Gisclair said.

Sperry's mud loggers could no longer see what was happening because another crew was displacing sea mud from a pit which is normally not monitored, Gisclair said, adding, "Pit volume changes are typically the primary indicator of fluid movements. This is critical information to help mud loggers identify problems which may be occurring on the rig. They also didn't know about other activities, such as moving fluids from a boat onto the rig."

Gisclair said in one 30-min period the day of the accident, various contractors were dealing with dumping of the sand tanks, trip tank gains and losses, crane operations which are obscuring flow-out sensors, balancing operations moving light seawater from one part of the rig to another, and taking on heavy mud at another area. "All of this can make it impossible to determine what flow-out was coming from the hole, what flow-out was coming from the trip tank, and what was coming from the sand trap," he said.

Roth noted that on deepwater wells, float chutes are used to provide a redundant device independent of the float collar. This was not part of the Macondo well's design, he noted, which actually prevented the meaningful placement of cement bond logs to determine whether good zonal isolation was achieved.

"BP did not elect to run cement bond logs, with the understanding that it had equipment on board to handle that," Roth said. "It would have been difficult but not impossible."

More Oil & Gas Journal Current Issue Articles
More Oil & Gas Journal Archives Issue Articles
View Oil and Gas Articles on PennEnergy.com