ML20245C530
ML20245C530 | |
Person / Time | |
---|---|
Issue date: | 02/13/1985 |
From: | Shewmon P Advisory Committee on Reactor Safeguards |
To: | Advisory Committee on Reactor Safeguards |
Shared Package | |
ML20245C504 | List: |
References | |
FOIA-87-666 NUDOCS 8711030200 | |
Download: ML20245C530 (2) | |
Text
_ _ _ _ _ _ - _ - _ _ -
8'3
$ :j g NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
$g"2 b a Q PiASHINGTON, D. C. 20$56
% ,,,,, ,. February 13, 1985 ~
- I
~
To: File, Metal Components Subcommittee.
From: Paul Shewmon, ACRS Member cp.A ,
)
Subject:
Steam Line Failure at Detroit Edison's Monroe Plant On Feb.12 I visited Detroit Edison's Monroe power plant with an NRC team that consisted of Dick Vollmer (I&E), Warren Hazelton (NRR), Al l
Taboda (RES), Carl Lundeen (Prof. U. Tenn., welding) and Carl Czajkowski(BNL,weldingspecialist). Unit #1 that had the failure has been on line since 1971. The plants three other identical units came on in successive years. The plant produces 3000 Mwe (4x750), is one of the biggest power plants in the country, and is proud of their records in output, efficiency, etc.
The failure occurred at 8:07 AM which is just after a shift change, and in an area that the incoming operators walked past on their way to }
the control room to start the shift at 8:00 AM. A few minutes before l the failure a couple of engineers came past the line, noticed a steam leak, walked 100 ft. into the control room, told the operators, who went directly to check it out, saw the steam leak increasing in l intensity as they approached, and saw the line break before they got to it. Thus from detectable leak to rupture was a matter of minutes
-- so much for a reliable, slow leak-before-break, at least in steam lines!
Unit il had seen 97,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of service wnen the failure occurred.
It is baseloaded, not load following. The line operated at 1000 F, 730 psi of reheat steam. There had been few if any temperature excursions above these limits. This operation is within the ASME code. The consultant they have on the failure (as does the Mohave plant)isFailureAnalysisAssoc.(FAA). FAA ran the time-temperature-pressure history for the line through a proprietary j computer code and told them, no doubt with a straight face, that the i failed pipe had used only 2.8% of its life. From this I conclude that i whatever the failure mechanism is, it is something that neither the ASME Code people nor FAA understand or know how to put into their design calculations.
Given this unknown failure mechanism puts the utility in a poor j position to assure continued safety of operation of the plant, in my opinion. They have inspected some of the other lines, found no defects (by ultrasonic techniques), and reduced the power to reduce the pressure to 440 psi (from 730) and temperature to 950 F'. They will use ultrasonics and magnetic particle techniques for a more detailed examination of the lines at a long outage later this year.
They will look for cracks that come to the surface, and/or internal cracks. They have no assurance that such large cracks (flaws) exist long before failure, though they did in the minutes that the pipe leaked before it broke there clearly was a flaw that penetrated through the wall.
' 8711030200 87$930 FOIA
[
PDR ppg SEPE87-666
y v y oe , mw u muupawmm m swa. s ;
Wherf tNe fracture surface was first observed FAA told the utility they could see no obvious initiating defect. The fracture ran for about 20', most of the fracture ran along the fusion line with no "
visible ductility. The shape of the weld zone in the metal was - j precisely shown on the fracture surface. The last few feet of the l fracture that ended in the pipe (as opposed to that ending in the girth weld) had a thickening shear lip that tore into the basemetal.
Something had obviously weakened the fusion line region. The welding experts said this behavior was not typical for this material, which j means that samples tested in the lab. after at least 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of j creep, in the absence of steam, don't behave this way. l The NRC team members agreed that the fracture morphology was different in the Mohave plant, in that the fracture there didn't follow the fusion line, and there was more oxidation visible in the metal.
(Sometimes referred to as a 'slac inclusion' along the interpass interface.) ,
The experts in this field, i.e. the people who work with the ASME code, feel that this type of failure couldn't occur in the steam lines in nuclear power plants because such lines operate at much lower temperatures where creep is not limiting in design. I suspect they are right. However, they also agree the failure shouldn't have occurred in the fossil plants. I believe the failure mechanism operating here and at Mohave involves an interaction between slow ,
creep, steam corrosion, and the welded structure. One mechanism that I would fit this requirer.ent would be the following: localized injection l of hydrogen from corrosion, the hydrogen opens methane bubbles beneath the surface, creep load helps grow these bubbles until they rupture (merge) locally, admitting steam, which develops a local oxide region
('slaginclusions'). The type of degradation cannot be observed by UT or Mag. particle techniques until it is well advanced because it is on too fine in scale. Why I postulate this mechanism is too detailed an argument for this memo, but the same type of corrosion induced ;
porosity can occur at much lower temperatures in the carbon steels I used in nuclear plants steam lines than in the Cr/Mo steels of fossil plants. Thus until we have some real explanation of what caused the !
two failures of the last six months, it isn't clear to me that the !
nuclear plants are imune, j cc: ACRS Members R. F. Fraley D. Vollmer M. W. Libarkin l W. Hazelton J. C. McKinley l A. Taboda j
~
1 1
l l
l l
l l
- _ - _ _ _ _ _ _ _ _ .