Unlike their American counterparts, not all control room operators in Japan have access to plant specific training simulators. Instead, according to a report by NPR, they use “generic” simulators that are similar to, but not identical to their plant. This difference may have contributed to the difficulties operators had at Fukushima Dai-ichi when responding to complex events that followed the catastrophic earthquake and tsunami on March 11, 2011.
The differences between a generic simulator and a plant-specific one are often in precisely the systems the Fukushima operators were struggling with: electrical power supplies, cooling water, and building ventilation. These differences matter little when training for events within the design basis of the reactor, but when events stray outside the design basis, include “cascading” failures, and involve severe accident response, generic simulators can’t accurately model the events. When training for situations that exceed the capabilities of a simulator, instructors are left with fewer, less realistic options like classroom training or “table top” walkthroughs of operator actions.
In my 20+ years as a nuclear plant senior reactor operator, instructor, and manager I spent thousands of hours on plant specific simulators as a student, instructor, and management evaluator. I can say with certainty the differences between a generic and plant specific simulator matter greatly in the quality of training received by operators! This higher quality training results in operators with greater proficiency in dealing with plant upsets and complex events. In fact, training on generic simulators is a practice the USA abandoned more than 20 years ago. This shift was but one of many improvements in training for nuclear workers adopted following the 1979 accident at the Three Mile Island plant. Today no nuclear plant can be built in the USA without it’s own plant specific simulator, and no operator can be granted a reactor operator’s license without exhaustive training and examination on a plant specific simulator.
My peers from other nations often have not agreed. Many I have met view plant specific simulators as a costly extravagance, and they still rely on generic simulators. In addition, generic simulators are often located at some central or distant location rather than on the plant site. This makes simulator training less frequent and more time consuming because of the travel involved. As a result, operators get far less simulator time per year than their counterparts with local plant specific simulators.
In the USA, nuclear plant control room operators routinely spend one of every five or six weeks in full time training. Each training week is typically composed of 20 hours of simulator time and 20 hours in the classroom, and includes both written examinations and evaluated scenarios in the simulator. If a team or individual fails to pass an evaluation they do not return to operate the plant until they have been remediated and pass a subsequent re-evaluation. The result is more than 100 hours of training and evaluation per year on plant specific simulators for every operator. Compare this to commercial airline pilots who typically attend only one week of simulator training per year, and you begin to get the picture of how highly trained US nuclear plant operators are. When taking into account initial qualifications, additional training for subsequently higher levels of responsibility, and routine “continuing training”, a typical nuclear plant operator in the USA spends about one-third of their career in training!
While it is too soon to draw conclusions, I suspect the international response to the events at Fukushima Dai-ichi will include improvements in the quality and quantity of operator training, and increased access to plant specific simulators at nuclear plants around the world. Even if official findings do not include this recommendation, nuclear utilities without plant specific simulators would be wise to consider such an investment at each nuclear plant they operate.
John Wheeler
#1 by Jim Fredericks on July 10, 2011 - 6:10 PM
If any good comes of the Fukushima experience, it will be that many so-called costly extravagances will be understood as bargain-basement, common sense precautions. How does the cost of a plant specific simulator compare to the cost of meltdown exacerbated by poorly trained workers?
#2 by Al on July 11, 2011 - 10:26 AM
Although I fully agree that plant-specific, full-scope simulators are highly beneficial for training competent operators, I have serious doubts that it would have helped in the Fukushima accident.
For starters, I don´t know of any plant simulator being used for training under core damage conditions. Even the term “full scope” is relative, since many local actions (for instance disabling interlocks, connecting jumpers…) that are required during ECA procedures are not physically reproduced due to the lack of the actual racks, cabinets, etc (i.e., the instructor “simulates” the action from his console)
In a Fukushima-like scenario, where the loss of DC due to flooding effectively killed I&C, left the MCRs in the dark very soon, and the post-earthquake-and-tsunami environment made field operations extremely difficult, I would judge the benefit of simulator training to be slim.
Operators (including field operators) would have needed a mixture of Navy Seals’ and nuke Navy damage control training more than anything else. Given the circumstances they faced, they performed exceptionally well (as recognized by Weightman, the UK Chief Nuclear Inspector that led the IAEA Fact Finding Mission). After reading the TEPCO report about first response to the accident, I can only agree.
Despite having been a nuclear advocate for a long time, I cannot stop feeling uneasy with some statements coming from the US, implying that the Japanese were sloppy and such things could not occur in the US. I understand it is an effort to reassure the public, but it is unfair, self-deceiving and in the long run counter-productive.
#3 by Blubba on July 12, 2011 - 5:17 AM
I also question whether plant specific simulators would have made much of a difference. It is still far to early to make assertions that if the plants had been owned, operated and regulated by Americans that the outcomes would have been substantially different but there are areas that merit exploration. For starters, the Japanese regulatory structure and relationship with licensees appears to be more analogous to the old Atomic Energy Commission than the NRC. The question is what regulations and requirements are imposed on US plants that are not at Japanese reactors that would have made a difference. Also, the industry and NRC have put a lot of focus on reviewing SAMG and EDMG commitments in response to Fukushima. It would be important to understand the extent to which the Japanese plants had analogous preparations and how well they performed. I’ve also been told that the Japanese have not been strong adopters of probabilistic risk assessment tools compared to Americans and Europeans. I would be interesting to examine what an internal flooding model would have revealed and what plant changes that might have driven in response or provide insights into why Units 5 and 6 did not completely lose diesel power while the other units did. The Fukushima Daini plant just down the coast experienced substantially the same earthquake and tsunami. Did their operators train on plant specific simulators as opposed to the operators at Fukushima Dainichi?