I wanted to return to Jeff Donn's piece from earlier this week concerning nuclear safety inspections. As I mentioned in a post from last night, I've been working with NEI's Jim Slider on taking a closer look at it. Jim is an old pro who has worked in the nuclear industry for more than 35 years, beginning his career at NRC's Office of Nuclear Reactor Regulation as a design basis accident analyst.
One of the first things I asked Jim about concerned the figures that Donn cited in his report. According to the story, the GAO report AP obtained a copy of said that NRC cited 10,776 "low-level" violations and 257 "higher-level" violations at U.S. nuclear plants between 2000 and 2012 (see page 3 of the AP story that was posted at ABCNews.com for the aggregate numbers).
Those data points puzzled Jim because according to the Reactor Oversight Process (ROP) that NRC instituted in 2000, NRC categorizes inspection violations from 1 (most significant) to 4 (least significant). The terms "higher-level" and "lower-level" simply aren't used. Given that AP refused to share the GAO report with us, we don't have any way to resolve this discrepancy.
The ROP was instituted to help plant operators and NRC identify problems with the highest level of safety significance in order to prioritize corrective actions (work on violations rated "1" first and take care of the rest in order of safety priority).
Now back to the raw numbers. There is one thing we can be sure of -- that when you add up 12 years (!) worth of statistics you can compile some scary numbers. I'm sure plenty of readers blanched at the figure of 10,776 "lower-level" violations cited by NRC inspectors. But when you realize that those violations were tabulated for 12 full years across 104 U.S. reactors, the numbers look a lot less scary.
Let's just do the math:
One of the first things I asked Jim about concerned the figures that Donn cited in his report. According to the story, the GAO report AP obtained a copy of said that NRC cited 10,776 "low-level" violations and 257 "higher-level" violations at U.S. nuclear plants between 2000 and 2012 (see page 3 of the AP story that was posted at ABCNews.com for the aggregate numbers).
Those data points puzzled Jim because according to the Reactor Oversight Process (ROP) that NRC instituted in 2000, NRC categorizes inspection violations from 1 (most significant) to 4 (least significant). The terms "higher-level" and "lower-level" simply aren't used. Given that AP refused to share the GAO report with us, we don't have any way to resolve this discrepancy.
The ROP was instituted to help plant operators and NRC identify problems with the highest level of safety significance in order to prioritize corrective actions (work on violations rated "1" first and take care of the rest in order of safety priority).
Now back to the raw numbers. There is one thing we can be sure of -- that when you add up 12 years (!) worth of statistics you can compile some scary numbers. I'm sure plenty of readers blanched at the figure of 10,776 "lower-level" violations cited by NRC inspectors. But when you realize that those violations were tabulated for 12 full years across 104 U.S. reactors, the numbers look a lot less scary.
Let's just do the math:
- 10,776 "lower-level" violations / 12 years = 898 violations per year / 104 reactors = average of 8.6 "lower-level" violations per year per reactor or less than .75" lower-level" violations per month.
- 380 "lower-level" violations/12 years = 31.66 per year /2 reactors = average of about 15 "lower-level" violations per year, or around 1.25 per month.
Comments
The term "violation" is being misused by Donn for dramatic purposes*. The GAO report is clear enough that almost all of these are findings, not violations. Green findings will be almost all of the "non-escalated" category.
--------
* Drama attracts readers, who knew?