AP Wire Story on Japan Reactor
John D Hawkinson
sufficiently to write a reply.
Assuming the AP article got the age of the reactor right (and screwed up the
power rating and the type), the affected reactor is:
Owner: Tokyo Electric Power Company
Unit: Fukushima Daiichi 2 (there's also a Fukushima Daini 2 owned by the
same company, but the power listed is way too high....)
Power: 760 MW(e), net to grid
Type: Boiling Water Reactor
Initial Criticality: May, 1973
Commercial Start: July, 1974
Reactor Supplier: General Electric
Generator Supplier: General Electric
Architect/Engineer: Ebasco
Constructor : Kajima
O.k. General Electric BWRs have TWO reactor circulation pumps located outside
of the reactor pressure vessel. The circulation pumps force water through
piping into the vessel and thence into "jet pump nozzles" located midway down
the vessel. Jet pumps assist the flow of water through the core--a lot of
core cooling can occur even without the assistance of the reactor pumps just
from natural circulation (convective flow), but the pumps allow the reactor to
operate at a higher power level.
As best I can tell from the AP article, an operator inadvertently tripped one
of the two operating reactor pumps, while the reactor was at power.
O.k. What probably happened next is the reactor protection system (RPS)
detected an interruption in the flow of one of the two reactor coolant
recirculation loops (makes sense, one of the pumps just turned off!!). This
event triggers a reactor scram, and simultaneously the "trip logic" signals
for the start of the emergency core coolant system (ECCS actuation).
When the ECCS is activated, the High Pressure Core Spray System (HPCS) is
activated and prepares to provide makeup water to the reactor vessel IF NEEDED.
Valves are aligned to feed water into the reactor vessel, pump motors are
started and run up to speed, spray nozzles are directed to remove decay heat
from the core, one of the emergency diesel generators starts and is run up to
speed (just in case electrical power should be lost), and a variety of other
core coolant systems are held in readiness while the control room operators and
the plant computer sort out the situation. If the situation is actually
getting worse, the plant computers can take further actions to safeguard the
reactor core (mostly by flooding lots and lots of water on it), or the control
room operators can over-ride the plant computer and manage the incident
manually (or as what probably happened in Japan, tell the plant computer the
trip was accidental and call off the ECCS activation).
The design philosophy behind all this is fairly straightforward--it's better to
scram the plant and start ECCS immediately than delay flooding the core with
water if it's ever needed.
Since the AP article didn't mention anything else happening, like the pressure
vessel automatically depressurizing, water actually being shot into the reactor
vessel, or reactor steam being dumped to the supression pool (huge water vats
inside the containment building that cool and condense steam), I'd guess the
plant operators realized the error and called off ECCS activation within the 27
seconds between the trip signal and time the plant computers would have
autonomously started injecting water into the reactor vessel.
In the U.S., the plant operators would probably have been required to notify
the U.S. Nuclear Regulatory Commission of the incident within 4 hours as a
"non-emergency event", per 10 CFR 50.72b(2)(ii): "Any event or condition
that results in manual or automatic actuation of any Engineered Safety Feature
(ESF), including the Reactor Protection System (RPS)." If it got to the point
where the ECCS actually injected water into the vessel (i.e., the event went
beyond 27 seconds), then the report would be made to the NRC within one hour.
It definatly would NOT have reached even the lowest threshold of the Emergency
Classification Level scheme--Notification of Unusual Event (NUE), and thus
off-site notification would not be required. (Personal opinion: It might,
however, be darn smart public relations to make a press release, especially in
a nuclear-phobic community).
NOTE: Usual disclaimers--my opinions and interpretations are my own, and
shouldn't be construed as reflecting ISU or anyone else.
Stephen Behling
In article <1992Oct7.0...@news.iastate.edu>, ha...@iastate.edu (John D Hawkinson) writes:
[...]
|>
|> Assuming the AP article got the age of the reactor right (and screwed up the
|> power rating and the type), the affected reactor is:
|>
|> Owner: Tokyo Electric Power Company
|> Unit: Fukushima Daiichi 2 (there's also a Fukushima Daini 2 owned by the
|> same company, but the power listed is way too high....)
|> Power: 760 MW(e), net to grid
|> Type: Boiling Water Reactor
|> Initial Criticality: May, 1973
|> Commercial Start: July, 1974
|>
|> Reactor Supplier: General Electric
[...Interesting hypothetical scenario deleted...]
Just thought I'd add my two cents about the Japanese nuclear power
program. I know John A. did a post assuming Japanese built only PWRs.
In Japan, they have both Boiling Water Reactors and Pressurized Water
Reactors. General Electric built the early BWRs solo, but since then
Hitachi and Toshiba build BWRs. Westinghouse built the early PWRs, but
since then Mitsubishi Heavy Industries builds PWRs.
Each publicly owned utility has chosen one type of reactor. So Tokyo
Electric Power Co. builds BWRs and Kansai Electric Power Co. (Osaka-Kyoto-
Kobe area) build PWRs. Other BWR utilities are Chubu, Tohoku, and Chugoku.
Other PWR utilities are Hokkaido, Kyushu, and Shikoku.
Another point about ECCS actuation in BWRs...they do not run with borated
water (or the stuff would plate out on the fuel rods). The ECCS system is
borated, so if it does get accuated, the boron would have to be flushed out
before normal operation could be resumed. This could cause operators to
try other recovery techniques before ECCS is activated. [Note, I have no
knowledge about the actual plant technical specification and procedures
in this area.]
I would assume the full details will be coming out in the trade journals
in the next few months. If someone would remind us at that time, we could
post the complete story.
-steve behling
Mike Smith
>Sorry for the delay in posting...took awhile to unscramble the AP news article
>sufficiently to write a reply.
Fine with me.
An interesting description of what most likely did happen deleted.
If this is even remotely right, then the AP article isn't, as some
have said, analogous to someone using their brakes to avoid an
accident. It is more analogous to someone (accidentally) taping
their brake pedal and triping the 'cruise control' to shut off,
seeing the speed start to drop, and resetting the cruise control...
i.e. there wasn't even a risk of accident involved, only a risk of
an accidental shutdown of a perfectly fine plant ...
>It might,
>however, be darn smart public relations to make a press release, especially in
>a nuclear-phobic community).
And especially if the press are going to completely mis-interpret the
meaning of an ECCS trip, that may not have even happened...
More and more I'm coming to think that the word 'Emergency' needs to
be taken out of the cooling descriptions ...
--
E. Michael Smith e...@apple.COM
'Whatever you can do, or dream you can, begin it. Boldness has
John De Armond
> Another point about ECCS actuation in BWRs...they do not run with borated
>water (or the stuff would plate out on the fuel rods). The ECCS system is
>borated, so if it does get accuated, the boron would have to be flushed out
>before normal operation could be resumed. This could cause operators to
>try other recovery techniques before ECCS is activated. [Note, I have no
>knowledge about the actual plant technical specification and procedures
>in this area.]
Here's the rundown on the reactor protection system. GE came about as
close to a standardized design as any so most plants adhere to this
scheme.
The borated water system you speak of is known as the Standby Liquid
Control system and is the last ditch system that would be used only
in extreme emergencies. As the inside joke used to go, when the
SLC button is pushed it also drops a ticket to the local concrete
vendor to start hauling in the concrete to fill in the reactor.
The first system is called the HPSI (prounounced Hipsy) system.
This consists of a pair of steam driven pumps located inside the
secondary containment. The steam comes from the reactor. The pump
takes suction from the torus (MkIII) or supression pool (MkIV) and
injects the water on top of the core. Anything that actuates the
ECCS also causes a containment isolation so the excess steam flows
from the reactor to the torus/supression pool and forms a closed loop.
The system is actuated when power is dropped to a fail-to-open valve on the
steam line. The pump runs until the steam runs out. At that point
the LPSI (Lipsy) system takes over. One pump is typically steam driven
and the others are electric. They operate the same as the HPSI except
they are higher volume, lower pressure pumps.
If the LOCA is so severe that all reactor pressure is gone, the Core
Spray system takes over cooling. This consists of several multi-thousand
horsepower electric pumps that take suction from the torus/supression
pool and pumps into a sparger header inside the reactor and over
the core. All these systems can be reconfigured to take suction from
the condensate system or the river.
I've heard nothing more about the particulars of the incident in question.
I still can't imagine what an operator could do that fits even some
reasonable interpretation of the "facts". Since almost any SCRAM in
a GE reactor causes containment isolation and thus activates the ECCS,
the incident could have been as minor as a technician bouncing a reactor
level transmitter. This is a common occurance because the trip setpoints
are close to the operating points. A physical jar or even the act of
valving one of the 4 transmitters out for calibration can jog the other
three into generating a SCRAM.
The way the reactor protection system works in a BWR is the "one
out of two taken twice" concept. That is, there are 4 channels
of reactor protection organized into two trains of redundant
pairs. A single trip from each train causes a half-scram. Any
trip on the other channel causes a full scram. So for example,
a reactor water level transmitter can half-scram Train A and
then something as innocuous as putting a channel of Train B in
calibration mode (generates an automatic half-scram) would trip
the reactor. Operators work REAL hard to minimize the state of
half-scram :-)
John
--
John De Armond, WD4OQC |Interested in high performance mobility?
Performance Engineering Magazine (TM) |
Marietta, Ga |Interested in high tech and computers?
j...@dixie.com |Write me about PE Magazine
Need Usenet public Access in Atlanta? Write Me for info on Dixie.com.