Advanced modelling software analysed the cementing conditions for BP’s Deepwater Horizon oil well as unstable, days before the blast that killed 11 oil rig workers and let millions of barrels of oil spill into the Gulf of Mexico.
Halliburton, the company that carried out the cement job, used its own modelling software called OptiCem, to support arguments that more stability was needed for the piping and cement. High pressure within the well eventually caused it to blow out, spewing oil.
The dispute over the modelling highlights the complications over cementing analysis, and how closely software modelling and human experience are aligned. The arguments were detailed in the findings by the US government Oil Spill Commission, which found problems with the cementing and poor managerial decisions, but do not pin the accident on one factor. BP declined to comment.
An OptiCem test on 15 April, five days before the blast, stipulated that from Halliburton’s point of view, 21 ‘centralisers’ needed to be added to the well bore. The centralisers are used to provide space around the oil pipe casing within the well, as cement is poured around it, and are a vital part of safe drilling.
BP initially adhered to the OptiCem software test and ordered 15 extra centralisers. But when technicians on the rig received the extra centralisers they mistakenly decided the new centralisers were the incorrect type. At this point BP proceeded with the drilling anyway, with the six centralisers, deciding another known technique of injecting cement in other places would work.
Halliburton representatives used the OptiCem software models to warn against the drilling with only a small number of centralisers. The company strongly advised of a potential “severe gas flow problem” in the well.
BP today declined to comment, but said in a Bloomberg story in August that in order “to deflect attention away from its potential role in the well blowout, Halliburton has tried to focus the public’s attention on the number of centralisers used by BP in the Macondo well”. It insisted Halliburton had not deemed the cement job to be unsafe with the six centralisers, and questioned the cement slurry mix.
According to the Oil Spill Commission’s findings this week, Brian Morel, drilling engineer at BP, wrote an email to Brett Cocales, another BP engineer, as the drilling proceeded, saying: “Who cares, it’s done, end of story, we’ll probably be fine”.
He added that a decision by BP’s well team leader, John Guide, to proceed with the drilling and cementing with the lower number of centralisers, was right on the “risk-reward” ratio.
The commission did not conclude whether the number of centralisers was wrong. But it highlighted the strong dispute over modelling and questioned why OptiCem was not rerun at this point.
At a hearing in July, BP’s well team leader, John Guide, explained the decision not to go with the software’s recommendations. “The model is – first of all, it’s not accurate all the time. ...I put very, very little faith in the model because it’s wrong a lot.”
BP still drilled with “no direct indicators of cement success” and no cement evaluation log, the Oil Spill commission said. The company conducted a separate negative pressure test, an oil engineering test designed to show whether the casing and cement would hold against significant pressure, and isolate potentially dangerous hydrocarbons.
The test was failed, but was – for an unexplained reason – deemed a “complete success” by both BP and rig owner Transocean at the time, a presentation on Monday said.
The report said there was “no evidence” to suggest BP had put cost before safety. But Fred Bartlit, the lawyer who led the investigation for the US Oil Spill Commission looking into the disaster, said there were still no answers as to why BP closed the well after the failed negative pressure test.
Commission co-chair William Riley noted “what appeared to be a rush to completion” at the drilling site. He added that “one must ask where the drive came from that made people determine they couldn’t wait for sound cement, or the right centralisers”.
There was also “poor communication” and a “muddled line of authority” at the well, the report noted. Questions also remain over key automated systems in the blowout preventer, which sits on the well head. Forensic examinations on the BOP are continuing.