The warning came in an email to the supervisor of a complicated project at the Crystal River nuclear plant.
"I just want to reiterate my concern ... ''
It was March 9, 2009. To replace aging steam generators, Progress Energy was about to cut a big hole in the concrete building that shields the nuclear reactor.
Charles Hovey was an experienced construction foreman who had worked on similar projects at other nuclear plants. Progress, he observed, planned to use a different procedure to cut into its containment building.
"I have never heard of it being done like this before and I just want to express my concerns to you one last time.''
Progress considered Hovey's point, then went ahead with its plan that fall.
Disaster followed. Workers discovered a crack in the wall while cutting the 25-by-27 foot hole through the 42-inch-thick concrete. A later repair attempt created a second crack. The plant has been shut down since. Progress will not restart the plant for at least two more years — if ever.
Price tag for the fix: $2.5 billion and rising.
Since the accident, Progress officials have argued that customers should be forced to help pay for the damage since no one could have reasonably foreseen the problems.
This "could not have been predicted," Progress spokeswoman Suzanne Grant wrote to the St. Petersburg Times.
But the Times has found that others echoed Hovey's concerns; that Progress hired subcontractors to perform critical tasks even though they had no experience with similar projects at other nuclear plants; and that it used a markedly different procedure in getting ready to cut into the containment wall than was used at other plants.
A nuclear plant containment wall is the final barrier keeping deadly radiation from the reactor from reaching the atmosphere. Inside the wall are metal "tendons'' that tighten around the structure, reinforcing its strength. Before cutting into the concrete to remove the steam generators, engineers loosen the tendons in the area they plan to cut.
Hovey's concerns centered on how many tendons Progress planned to loosen and in what order.
At other nuclear plants, engineers usually loosened 70 to 80 tendons.
Progress hired Sargent & Lundy to determine how many tendons on which to focus. The company had never performed such an analysis for a nuclear containment wall, according to documents obtained by the St. Petersburg Times. The engineering firm first recommended 97.
Too many, Progress said.
"We said, 'Hey, that's a lot of tendons,' " John Holliday, a contract employee for Progress in charge of the work on the containment building, recalled in a recent deposition. "Can you go back and take another look at this?"
The next proposal: 74, within the range loosened at other plants.
Holliday wasn't satisfied. "De-tensioning the tendons is a very expensive and time-consuming effort," he said.
The engineers were told "to put on their thinking caps and determine if there is an alternative method of analysis that we could pursue that would result in a lot less tendons being de-tensioned," Holliday said.
The final result: 65 tendons would be loosened.
The procedure for de-tensioning prescribed by Sargent & Lundy for Crystal River differed from that used at other nuclear plants.
• At other plants, workers generally did not loosen one tendon and then the one right next to it. They did it nonsequentially. At Crystal River, the tendons were loosened in sequential order.
• At other plants, workers de-tensioned all the proposed tendons and then cut the wall open. At Crystal River, only 27 of the 65 tendons were loosened before the cut was made.
It was this markedly different procedure that Hovey questioned seven months before Progress enacted its plan.
'No doubt' it was okay
Progress tried to save money on the project from the start.
Just two companies, Bechtel and SGT, had managed all the previous 34 steam generator replacement projects at U.S. nuclear power plants. Of those, at least 13 had involved cutting into the containment building.
All 34 projects were successful.
Progress expected the Crystal River project to cost about $230 million. For the management portion of the job, it got bids from both Bechtel and SGT. The lowest, from SGT, was for $81 million.
Progress officials rejected that offer. They could save $15 million by self-managing the project.
Although Bechtel didn't get the job of managing the project, Progress did hire Bechtel to manage the construction. Hovey sent his e-mail that raised concerns about the plan to several of Bechtel employees. As his concerns about Progress' plans circulated, other Bechtel employees began asking questions.
"Why are we doing tendons different here than all other jobs?" site supervisor John Marshall asked in an e-mail sent to Sam Franks, another Bechtel supervisor.
Gary Goetsch, a supervisor with the company hired to prepare the containment building for cutting, had worked on 11 similar jobs. He said the Crystal River job was "the first and only one'' to use the procedure adopted by Progress officials, according to notes of his interview with analysts from Performance Improvement International, the firm hired by Progress to determine what went wrong.
And he knew the plan was a bad idea, according to the notes. (The Times could not ascertain whether Goetsch raised his concerns with Progress before the accident.)
But Holliday, noting the questions raised by Hovey and the Bechtel supervisors, ordered Sargent & Lundy to rerun its computer analysis. According to Holliday's deposition, that analysis said everything would be okay.
"You can practically do a hand calculation to show'' that Sargent & Lundy's procedure was safe, Holliday said. "There was no doubt in my mind it was okay.''
On July 31, 2009, Holliday's boss, Dan Jopling, sent an e-mail that said, "As we discussed in the past, the engineering approach to containment analysis used by Sargent & Lundy is significantly different than SGTs. I am satisfied the Sargent & Lundy approach is technically correct and will withstand scrutiny."
The project went on.
'Big chunks falling off'
In October 2009, Mac & Mac Hydrodemolition cut into the containment building. Holliday noticed a break in the concrete but decided to continue, thinking it predated the ongoing work.
An hour later, water began pouring through the wall. And the scene was not like what they had imagined.
"Hydro-demolition resulted not in the small pieces (you) would expect, (but) big chunks falling off the wall," Holliday told analysts from Performance Improvement International.
PII concluded that the tendon de-tensioning, particularly the sequence Progress used, and removal of the concrete caused the building to crack.
"A primary factor was the number of de-tensioned tendons that were located in a row," the analysts noted in their report.
The Crystal River nuclear accident is already one of the most expensive in U.S. history. Progress thinks insurance will cover most of the repair. But it wants its customers to pay $670 million.
For that to happen, Progress will have to convince Florida's Public Service Commission at a hearing scheduled for June that it acted in a reasonable and prudent manner in removing the steam generators.
Grant, the spokeswoman for Progress Energy, declined to answer specific questions from the Times, adding that "it is appropriate for us to only address specific questions related to these matters in the appropriate regulatory arena under the schedule set forth by the (Public Service Commission)."
But Progress can point to the "root cause analysis" it ordered from PII.
"Post event research and modeling … determined that very rigorous application of typical industry tools would not have been able to accurately predict'' that the containment wall would crack.
In other words, the accident would have happened to anybody cutting into the Crystal River containment building.
Still, Progress will likely have to answer pointed questions about its conduct.
• Why did it decide to be the first to manage the project itself?
• Why did it hire a firm, Sargent & Lundy, inexperienced in this type of work to come up with the engineering analysis?
• Why did it follow that plan, given that it was so different from the procedure used at other plants?
• Why did it hire a company, Mac & Mac, that had no experience cutting into a nuclear containment wall?
• Should the concerns of Hovey and the other experienced workers have been enough to get Progress to change its plans?
The Public Counsel's Office, which will represent customers at the June hearing, is convinced that Progress made errors in judgment that it needs to be held accountable for.
"We believe that the evidence is going to show that Progress was not prudent in the way they went about this repair," said J.R. Kelly, the state public counsel. "If we prove our case … the ratepayers should not be responsible for any of the costs flowing from the repairs."
Ivan Penn can be reached at firstname.lastname@example.org or (727) 892-2332. Follow him on Twitter at twitter.com/consumers_edge and find the Consumer's Edge on Facebook.