Usually, I'm a fan of "body counts", because they are the only way we have
to quantitatively assess risk. (Nuclear energy generation seems to be less risky
than some other activities.) However, when we want to assess if cultural or
procedural problems exist in a particular industry, we have to compare apples to
apples.
Some activities, such as skiing, mining, dam building, fishing, forestry
... require people to expose themselves to nature. These tend to be more risky
than activities that take place in totally man made environments (clerical,
teaching, hair dresser, NPP operator...). The more one is exposed to nature, the
more likely one is to encounter true accidents or "acts of god". You can reduce
the number of such events (fishermen can check the weather forecast before going
out), but they can't be eliminated (weather forecast is wrong).
Where was the act of god in the Tokaimura "accident"? Clearly there were
severe procedural and cultural problems there. Now we should assess if these
issues also exist in other parts of our industry. Is anyone else writing
procedures because regulators are requiring procedures? Are these procedures
written with the intent of satisfying lawyers and regulators, rather than being
practical and safe operating procedures?
----- Original Message -----
Sent: Wednesday, March 20, 2002 2:08
AM
Subject: Criticality accident and power
deaths
I wrote:
What number of accidental criticalities and
deaths would be needed to qualify as "many"?
Related responses
During two of the years I lived in the
Northwest, there were four fatal accidents (a total of seven deaths) in
chemical facilities handling SO2, and I understand that is approximately the
continuing rate of such fatal accidents. I think 2 fatal
accidents a year qualifies as "many." The rate of process criticality
accidents during the years 1955 to 1965, when most of them occurred,
qualifies as "many" but the accidents rate dropped dramatically. I
have not looked up any accident statistics for other industries..
There are about 10 fatal skiing accidents a
year in the U. S., and that qualifies as "many."
and
add the 121 workers that were killed during
construction of the Snowy
Mountains hydroelectric scheme in
Australia.
and
The Good News about Radiation by
John
Lenihan published in 1993 by Medical Physics
Publishing (out of print)
From p. 103 of TGNAR by Lenihan DEATHS BETWEEN
1969 AND 1986 RESULTING FROM SEVERE ACCIDENTS TO GENERATE ELECTRIC
POWER
Energy
Source no. of
severe accidents number of deaths*
coal
mining
62
3600
oil (refinery fires,
transportation)
57
2070
natural gas (fire, explosion)
24
1440
hydroelectric (dam failures)
8
3839
nuclear
1
31
* immediate deaths only
and
2. Characterizing an average 0.4 accidents per
year as "not many" is not justifying them either. I thought it was an
unusually good record of industrial safety, given the state of knowledge in
the 1950s and 1960s
I add:
Hoover dam - 96 industrial fatalities during the construction of the dam.
Industrial fatalities includes deaths from drowning, blasting, falling rocks
or slides, falls from the canyon walls, struck by heavy equipment, truck
accidents, etc. Industrial fatalities do not include deaths from the heat,
pneumonia, heart trouble, etc.
http://www.hooverdam.usbr.gov/History/fatalities/fatal.html
AND
OSHA Preambles
Electrical Power Generation, etc. (29 CFR
1910.269)
V. Regulatory Impact Assessment
A. Introduction
The Occupational Safety and Health Administration (OSHA) has determined
that there is a significant risk to the health and safety of workers who are
exposed to the hazards of electric power generation, transmission, and
distribution. . . ..
http://www.osha-slc.gov/Preamble/Elect_data/ELECTRIC5.html
In 1999, there was an inadvertent criticality in Tokaimura Japan. I do
not see how much further afield from power production you can get than the
bucket of fuel at the plant in Japan. The two deaths from this event PROVED to
me that those in charge in Japan learned nothing from the previous inadvertent
criticality accidents. Two sites (of many I found) below seem to support
that opinion:
http://www.iaea.org/worldatom/Documents/Tokaimura/iaea-toac.pdf
http://www-rcf.usc.edu/~meshkati/tefall99/toki.html
After reviewing these sites, is there ANYONE on radsafe who could not
have predicted this accident would occur? No, it was only a matter of
time.
So, what have we learned?
The power industry is a heavily industrialized process. People are killed
in industrial accidents - including the production of power - nuclear, hydro,
or fossil.
That the industrial safety at locations such as Hoover dam (built in the
1930's) and the Snowy Mountains hydroelectric scheme (1950's?) was not what it
is in 2002 (or in 1999).
Now, perhaps there is someone who can tell me how the above "statistics"
can be related to the inadvertent criticality and the 2 deaths in Tokaimura
Japan (1999)?
How can we compare industrial accidents from the 1930's or other
accidents (such as skiing accidents) to a totally avoidable accident in 1999?
To paraphrase Mr. Lipton - Comparative body counts are even less
useful.
Perhaps we should stop the comparative body counts and accept that not
everyone who needed to learn from the criticality accidents of the past
learned enough to prevent them (in the present).
After the Tokaimura accident no one can believe that an average of 0.4
accidents per year is "not many." The accident at Tokaimura clearly showed
that there are too many NOW.
Before the vaccine for smallpox, smallpox was a major health concern.
Following the development of the vaccine, smallpox was still common. In the early 1950s - 150 years after the introduction of
vaccination - an estimated 50 million cases of smallpox occurred in the world
each year, a figure which fell to around 10-15 million by 1967. Last natural
case, 1977. A fatal laboratory-acquired case occurred in 1978. Over what
period should we take an average? 2000 BC to 1951 AD? 300 AD to 1270 AD? 1900
to 2000? 1980 to 2000? Each gives vastly different "deaths per year" values.
However, for any average before 1977, the rate will be above zero. So based on
the rates before 1977 a new smallpox case in 2002 could be viewed as "not
many?"
What does this prove? You can "prove" just about
anything IF you get to use the statistics and rates of your choosing. Do rates
of death from skiing or construction of Hoover Dam have any comparison to
criticality deaths? Not in my mind.
Paul Lavely
<lavelyp@uclink4.berkeley.edu>
Just an old HP living in a world run by
administrative law.
--