Shearon Harris: One of
the most troubled reactors in the U.S.
In
August, NC WARN revealed that EIGHT system failures at the Shearon Harris
nuclear plant have occurred since April. Progress Energy stated there were
only FOUR.
(See list on page 2)
The Discrepancy:
The NRC uses a complicated and public-unfriendly formula for
classifying and reporting “SCRAMS”, also called “reactor trips,” “unplanned
shutdowns,” etc. In a nutshell, a multitude of human and mechanical
problems, along with acts of nature, can require the reactor to be shutdown
automatically or manually in order to avoid overheating.
Some of this year’s eight system failures at
Harris occurred while the reactor was off-line for refueling in April and
May, and as it was being powered back up afterward. Some of these are not
classified by the NRC as a SCRAM, even though if they had occurred with the
reactor running, each would have almost certainly caused it to trip.
All eight failures exceeded the NRC’s “event” reporting level, and each
required back-up safety systems to engage. Even when the reactor is
off-line, the cooling system must exhaust large amounts of excess heat from
the reactor core and waste pools.
In fact, the margin for safety can be far lower during many points of a
refueling outage than during normal operations; the reactor vessel and
containment building are often open, other safety systems aren’t always in
place, and handling of fuel creates particular risks. NRC studies show that
one particular 24-hour function during refueling involves risk equal to that
incurred during 37 days of normal plant operations. We believe the most
serious of Harris’s eight failures occurred at that time
(see April 26 below).
From NRC reports and Progress statements, it appears that the
causes for most of the emergency shutdowns remain unknown. Each failure
diminishes safety margins, and opens the possibility for further errors that
could lead to core damage.
The industry average for narrowly defined
SCRAMs is one per 27 months.
So even if one only compares Harris’s four actual SCRAMS in 2003, along with
four more last year, Shearon Harris’s problems are running ten times over
the industry average since 2002, when there were four additional SCRAMS.
As UCS Nuclear Safety Engineer David Lochbaum
points out, Three Mile Island had 12 recurring cooling system failures –
which plant owners didn’t understand or correct – in the year preceding the
April 1979 meltdown.
“Non-Nuclear” Failures?
Progress Energy told
the media the recurring problems are on the “non-nuclear” side of the plant
and did not involve the cooling system protecting the reactor. If that were
true, the Nuclear Regulatory Commission would not have sent a special
inspection team to Harris.
The various loops of the cooling system are interconnected, and together
must exhaust large amounts of heat in order to avoid damage to the reactor
core. At Harris, that cooling system also serves the high-level waste
pools.
History’s two worst nuclear accidents started on the
“non-nuclear” side. Turbine testing led to the 1986 Chernobyl explosions.
The “non-nuclear” condensate system at Three Mile Island failed, leading to
partial meltdown of the nuclear core.
System Failures at Shearon
Harris in 2003:
April 26
A relief valve failed, causing interruption of core cooling early in the
refueling process. A special two-week investigation by an NRC team found
that had cooling not been restored, water inside the reactor would have
boiled in 29 minutes. NRC said workers could have closed containment
hatches in 25 minutes to prevent radioactive steam being released into the
atmosphere, then started a pump to replace the water being boiled off from
around the reactor
core.
May 16
Restart of the reactor after refueling was delayed when a back-up cooling
pump was activated during testing of the primary cooling system.
May 18
A turbine trip activated back-up cooling pumps and caused shutdown of the
reactor as it was being powered back up.
May 20
Failure of operating main feedwater pump activated back-up cooling pumps and
caused shutdown of the reactor.
May 22
The reactor was at 72% power when it tripped again due to vibrations in the
main cooling water system.
June 14
Pump malfunction in the main feedwater system led to shutdown of the
reactor.
June 15
Human error led to failure in the same cooling system as the plant was
preparing to be brought back on line.
August 17
Pump failure in the condensate portion of the cooling system causing a
second pump to fail. Plant operators shut the reactor down in
anticipation of an automatic shutdown. |
