IN THE UNITED STATES DISTRICT COURT FOR THE EASTERN DISTRICT OF LOUISIANA
IN RE: OIL SPILL BY THE OIL RIG MDL NO. 2179 “DEEPWATER HORIZON” IN THE
GULF OF MEXICO, ON APRIL 20, 2010
EVALUATION OF THE CEMENTING ON THE
9 7/8 x 7” PRODUCTION STRING ON THE MACONDO WELL
EXPERT REPORT OF GLEN BENGE
ON BEHALF OF
THE UNITED STATES OF AMERICA
August 26, 2011
Evaluation of the cementing on the 9 5/8 x 7” production string on the Macondo Well
(Complete report at the above link)
- Summary of Conclusions
This review is an evaluation of the available information regarding the production casing cementing job on the 9 7/8 x 7” production string on the Macondo well, and a determination of the key factors related to the failure of the cement to provide isolation in the well. Included in the report are discussions of the key decision points and their impact on the cementing results.
In my review of the data from the Macondo well, based on personal experience and industry practice, I have concluded the following with respect to the production casing primary cement job:
Cement did not isolate the formations in the Macondo well. The failure of the cement to provide wellbore isolation or act as a barrier can be summarized as: 1) inadequate design of the cementing slurry or job to address the requirements of the well; 2) failure of the cement slurry to perform as expected; and 3) failure of the cement slurry to be properly placed in the well.
For the cement to fail to provide a barrier in the Macondo well, it was either not present across from a producing formation or it was not set and able to act as a barrier to flow, or both. Channeling allowed for a flow path in the annulus for formation fluids. Even with a flow path in the annulus to the casing shoe, the cement left inside the casing had the potential to provide a wellbore seal inside the casing. For it to not have provided a seal, the cement was most likely not set because of contamination, temperature effects or both.
The BP wells team was well versed in cementing. These BP personnel were the final decision makers and were empowered to accept or reject the advice of both the BP internal cementing expert and Halliburton. The BP engineers chose to accept additional risks when designing the cement job with the awareness that remedial cementing work could be done at a later date. Those additional risks included using a leftover cement blend not appropriate for foamed cementing, using a foamed cement in a synthetic oil- based mud (SOBM) environment, limiting cement volume and selecting a reduced number of centralizers.
BP’s slurry design for the Macondo well was inappropriate and not suited for a foamed cement application. BP compromised the quality of the slurry design by utilizing a dry blend leftover from the Kodiak #2 well. The dry blend leftover from the Kodiak #2 well contained additives that were not suitable for foamed cement. This resulted in a suboptimal foamed cement design.
A conventional (unfoamed) slurry could have been used to cement the Macondo well, which would have eliminated the significant risks associated with the use of a foamed cement on the production string.
The laboratory testing performed by Halliburton was incomplete and did not adequately evaluate the slurry. Because the cement was not originally developed as a foamed cement system, BP should have required more testing to confirm the appropriateness of the slurry design for the well. Free water, settling and unset foamed stability testing should have been part of the testing program. BP started the cement job without these key laboratory test results and without a complete set of tests on the slurry actually pumped on the Macondo well – the slurry containing 0.09 gal/sk SCR-100 retarder.
BP also did not follow its internal guidelines for testing cement slurries for deepwater wells. BP’s recommended practices specifically identify temperature as a major risk factor that can lead to cementing failures. Use of its internal guidance could have helped BP identify weaknesses inherent in the suboptimal cement design.
Laboratory testing of a slurry design is dependent upon using the correct temperature of the well. Failure to use a correct temperature can lead to a slurry that sets too quickly, or one that will be over retarded for well conditions and does not gain strength when needed. Halliburton performed lab testing of the Macondo well’s base slurry’s strength at 210° F, and that temperature was reached four hours after the tests were initiated. The testing for foamed compressive strength was performed at 180° F and showed no strength development for at least 24 hours.
It takes time for a well, especially a deepwater well, to recover to near bottom hole static temperature (BHST). Based on the available temperature data reviewed, the Macondo well took significantly longer than four hours to reach 210° F after completion of the production casing cement job. In the case of the Macondo well, the negative pressure test was performed less than 18 hours after the cement was in place. Based on all of the information I reviewed, it is my opinion that, at the time of the Macondo well negative pressure test, the cement was not set. If the cement was not set by the time of the negative test, it would not have been possible for the cement to provide isolation in the well. Once the flow began during the negative pressure test, if the cement was not set, any potential isolation from the cement was permanently destroyed.
Foamed cement should not have been used in the SOBM environment that was present at the Macondo well. The destabilizing effects on foamed cement by SOBM are severe and can lead to a job failure. The risks of failure are so severe that I have not, nor will I, recommend using foamed cement in an oil-based mud environment.
The job design was inadequate for the cement to be placed properly in the well. Poor centralization, use of the base oil pre-flush, limited pre-job circulation and low pump rates virtually assured the cement integrity would be compromised. Lack of proper centralization of the production string increased the potential for channeling, thus leaving an un-cemented area in the annulus.
The production casing used six centralizers spaced throughout the bottom portion of the cemented interval. A pre-job OptiCem run by Halliburton recommended the use of 21 centralizers1 and indicated an increased chance for channeling if that number was reduced.2
Inadequate pre-job circulation did not allow for breaking up of gels in the mud or circulation out of any formation fluids that may have entered the mud while the casing was being run. Coupled with poor centralization, this lack of circulation eliminated opportunities to maximize the circulatable volume of mud in the wellbore.
Finally, the use of base oil, particularly in a weighted mud system, can enhance channeling and is not recommended. BP’s use of base oil increased the chance for channeling due to its very low density and viscosity. The base oil’s low density also reduced the total hydrostatic pressure in the annulus.
EXCERPT FROM PAGE 36
When the positive pressure test was performed 12 hours after the cement was in place, the cement was not set in the well.
When BP performed the negative test on the Macondo well less than 18 hours after the cement was in place, the Halliburton foam crush data indicated the cement was not set.
EXCERPT FROM PAGE 37 of report referenced to Waiting on Cement Time (WOC), the fact that it is critical to obtaining a good cement job.
- Post Cement Job Well Operations - Wait on Cement (WOC) Time
When BP performed the positive pressure test 12 hours after the cement was in place on the Macondo well, the OT&C data79 indicates the cement was not set, with the Halliburton data showing the foamed samples also were not set. When BP performed the negative test on the Macondo well less than 18 hours after the cement was in place, the Halliburton lab data showed the foamed cement was not set.
After the cement is placed in a well, it must be allowed to set. This is commonly referred to as wait on cement, or WOC, time. The WOC time is determined by evaluating the strength development data from the lab reports. Industry practice calls for waiting until the cement has 500 psi strength before continuing well operations.
When cement is relied upon as a barrier for well control, the same requirements placed on any other piece of well control equipment should also be placed on the cement. A barrier must be tested to be relied upon in the well. In the deposition of Thomas Roth of Halliburton, he stated:
“[O]nce cement is placed in a well, before it can be depended upon to be an effective barrier as part of a well control program, it must be confirmed.”80