ML19254F454
| ML19254F454 | |
| Person / Time | |
|---|---|
| Site: | Trojan File:Portland General Electric icon.png |
| Issue date: | 11/04/1979 |
| From: | Broehl D PORTLAND GENERAL ELECTRIC CO. |
| To: | Engelken R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION V) |
| References | |
| NUDOCS 7911090461 | |
| Download: ML19254F454 (7) | |
Text
f bc: LER Distribution List D*
e PcrtlandGeneralBectricCon eniy le w w
3 v
DonadJ.BroeN
<-e Vce Rescent November 4, 1979 Trojan Nuclear Plant Docket 50-344 License NPF-1 Mr. R. H. Enge hen, Director U.S. Nuclear Regulatory Commission Region V Suite 202, Walnut Creek Plaza 1990 N. California Blvd.
Walnut Creek, California 94596 Deer Sir:
With reference to prior discussions and your letter of October 22, 1979 relat:ve to leaking tubes in "A" and "D" steam generators at the Trojan Nuclear Plant, this letter is in response to your request to provide the NRC with (1) the results of our investigation to determine the cause of the dnfects identified, and (2) the basis for our conclusion that power operation may be safely resumed.
The decision to remove the Trojan plant from service on October 12 was based upon observation of primary-to-secondary leakage rates over an extended period of tire. Although the leakage rates were well below limits established by the Technical Specifications, an outage was scheduled to locate and plug leaking steam generator tubes at 'a time convenient for power operations scheduling.
Leakage History Following the return to service of the Trojan plant in early January 1979 titer an extended outage, a small primary-to-secondary leak was detected (1-2 gpd). This leak was identified by the observation of small quanti-ties of radioisotopes in the main condenser off-gas.
By measurement of radioisotope concentrations, estimates of the magnitude of the leakage were made. However, with leakage at this level, it was not possible to identify the leaking steam generator to any practical degree of accuracy.
In the 4-month period to April 27, 1979, the leakage increased in magni-tude to approximately 3 gpd. At that time the unit was removed from service for a maintenan e outage. Due to the small magnitude of the leakage, no attempt as nade to locate the leak. The unit returned to y
service on July 3,1979, and leakage was again detected.
.7 g
u b, rs<.w s e sc w m gnoco e
Portland GeneralBechicCcwsy Mr. R. H. Engelken November 4,1979 Page two The magnitude of the leakage increased slowly to approximately 50 gpd on September 5, 1979 when the unit tripped from full power when a reactor coolant pump motor power circuit grounded., When returning the unit to power on September 6,1979, a safety injection actuation event occurred from a high stammMne flow coincident with low steamline pressure signal that resulted in a high differential pressure (>1650 paid) being applied across the steam generator tubes. This transient occurred when trouble-shooting the turbine EHC System during startup and resulted in all four turbine control valves to fully open with the turbine at about 200 rpm.
The transient was terminated by a safety injection, turbine trip and reactor trip. The unit returned to power on September 9, 1979, with primary-to-secondary leakage measure. at approximately the same value as before the September 5, 1979 trip.
The mit remained at full power until October 2,1979 when a trip from full power occurred due to the inadvestent closure of a main steam isolation valve. Primary-to-secondary leakage had f.acreased to approxi-mately 80 gpd prior to che October 2,1979 trip. Tie unit returned to power and operated at full powar until October 12, 1979 when the current outage began. Leakage at this time was measured at 90 to 100 gpd.
Location Testing - Prior to Shutdown During this lengthy operational period prior to the October 12, 1979 shutdown, extensive testing was done to identify the leakage with a particular steam generator. Samples of steam generator water from the blowdown system were taken, a well as concentrated samples (resin columns) from the main steam _ -
.2.
By late August this testing indi-cated that the most probable contributor to the primary-to-secondary leakage was the "D" steam generator, with possible 1er': age in the "A"
steam generator. As the magnitude of the leakage increased, more accur-ate location could be determined. At the time of the shutdown, it was estimated that approximately 80 percent of the total leakage was from the "D" steam generator, and the remainder from the "A" steam generator.
This estimate was later confirmed by identification of the leaking tubes.
to id'ntify the leaking steam Experience has shown that it is difficult e
generator below 50 spd leak race, and that leakage of approximately 75 gpd in a steam generator is necessary to assure that the proper steam generator is selected prior to removing the unit from service for repair.
vn7 176 O
e e +
y e
W C18tletal BOCitiC Cvugicsif Mr. R. H. Engelken November 4, 1979 Page three Testing Results for Steam Generators A and D Following shutdown on October 12, 1979, the primary system sides of the Steam Generators (SG) A and D were opened to insp1tet for tube leaks.
One tube in Row 1 of SG A, and four tubes in Row 1 of SG D were observed leaking water. The eddy current (ET) examination disclosed that an additional Row I tube in each steam generator provided an ET indication in the same general area (U-bend) as the leaking tubes.
A.
Visual Ermmination Visual aramination for leaks was accomplished by looking at the tube sheet through the manway. With the secondary water level approximately 3 ft above the tube bend and nitrogen overpressure, visual inspections were made and the leaks observed when the nitrogen blanket was pressurized at 50 lb and 10 lb in Steam Generator D, and at 45 lb and 10 lb in Steam Generator A.
Reported observations are: Steam Generator A, Row (R)1-Column (C)68 e 15 to 20 drops / min from the cold leg; Steam Generator D, Rl-C6 e 50 to 100 drops / min from the cold leg, R1-C62 # 5 to 10 drops / min from the hot leg, R1-C70 e 5 to 10 drops / min from the hot leg, and El-C91 e 1 drop / min from the cold leg.
When the water level was dropped below the top tube support plata no leakage was observed.
B.
Eddy Current Examination In Steam Generator D, a group of twenty-four (24) tubes made up of the four observed leakers, and adjacent tubes in Rows 1 and 2, were examined both at 100 and 400 KHz with a differential coil probe (two coils of equal number of turns in electrical balance). The probe size used was 0.650-in. diameter. The aramination proved excellent for inspecting the straight length of tube and the areas at tube supports. Ernmination with the differential coil, however, was uninformative in the U-bends.
There is no evidence of dentin 8 An examination technique using only one coil in the probe (absolute method) was then developed, with which we were able to observe indications from the areas where leakage was suspected (U-bends). Wa found an indication in each leaking tube except R1-C91 (which had the lowest observed leak rate of e 1 drop / min).
All of the other tubes in Row I and three Row 2 tubes adjacent to each of the leaking Row I tubes were examined using the
'oj
} }? {
~a pot &Trd Geilead BectiicCcisip==if Mr. R. H. Engelken November 4, 1979 Page four 0.650-in. diaaeter absolute probe. For all ET examinations in Steam Generator D, the tube area inspected was from the hot leg tube end to the first tube support (No. 7 support) on the cold leg side of the U-bend except for Tube R1-C49, 50, 51 and 52 which could not be avamined with the fixture in-place. These tubes were hand probed from the hot leg side half way around the U-bend.
In Steam Generator A, all Row I tubes (94 tubes) were avamined frott the cold leg side using the absolute coil at 100 132.
The probe size used was.0.650-in. diameter. The probe was pushed around the U-bend to top tube support on the hot leg side in nineteen (19) tubes including Rl-C68, the leaking tube. Due to difficulties encountered.in passing the probe and connecting cable through the full U-bend without damaging the sensor, we discontinued the process of passing the probe all of the way around the band for the remaining seventy-five (75) Row I tubes, stopping at the first resistau.. point past the apex. Tha ET indication in Tube RI-C68 verified the leakage seen from the cold leg. An indication was observed on the hot leg side of Tube Rl-C38 which is one of the tubes we could not pass the probe completely through the U-bend to the tube support plate.
This tube is not a visually observed leaker. Examination of the SG A tubes from tha hot leg side past the aper is in process.
C.
Ball Plug Gaging Thirty-seven (37) tubes in Row I were avamined with a 0.720-in.
diameter ball plug gage. The ball plug passed completely around the U-bend in all of these tubes including the leakers, thus indicating no ovality exceeding the manufacturing tolerances.
Corrective Actions The visual and ET examins. tion has identified the location of small perforations in the U-bend area of five Row I tubes and provided ET indications in the U-bend area of two additional Row I tubes. These seven indications together with the suspect leaker in SG B chat was plugged in 1978 constitute approximately 2 percent of the Row 1 tubes, and at least the seven indications have exhibited signs of distress in the U-bend area. We believe that this experience is sufficient to warrant further investigation into the cause o1 the problem for the knowledge and benefit of the industry as well as Trojan. The type of investigation necessary to determine the cause will involve entry into the steam generator above the top tube support plate and removal of a number of tubes for inspection and analysis. This type of program cannot
"/ &
7 178
P0fthM1cf Ge:1eral BecFicCamp= eor Mr. R. H. Engelken November 4, 1979 Page five be developed and results provided in a time frame consistent with our expected schedule for return to power in later November or early December 1979. Indeed, we would fully expect the conclusions from such a program to be considered at length before any actions involving the Plt.nt would be deemed appropriate. Keeping the unit out of service while a research program is developed, samples obtained, analyzed and conclusions debated is not in the public interest since alternative energy would need to be obtained from oil-fired generation with limited availability and very high cost.
We have considered the alternative of plugging all Row I tubes, and thereby permanently reducing the available number of tubes. However, the safety considerations discussed below dr., not suggest that such action is necessary, or even appropriate considering the lack of any technical justification. Rather, we believe the most appropriate course of action is to complete ET examination of Row I tubes in SG A and to follow this with ET m mination of Row I tubes in SGs B and C.
We will permanently plug all lov 1 tubes that exhibit leaks or indications in the tube bend area prior to returning to power. The time between now and the refueling shutdown next spring v111 bn. sed to develop a program, in conjunction with Westinghouse, and possibly with involvement of EPRI and the Steam Generator Owner's Group, to research the cause of the problem. Following shutdown next spring, we would expect to remove representative samples of the tubes of interest for subsequent examination and analysis in an effort to determine the mechanism causing the distress.
We believe that this course of action is reasonable while recognizing the uncertainty as to whether the mechanism causing distress is limited or progressive. This belief is based on our experience to date as these very small leaks have developed coupled with review cf the conservative analysis of possible accident consequences.
Safety Considerations 1.
To date, Icaks have only occurred in six steam gene.rator tubes over the 4 yr of Pbnt operation. ':he currently existing leaks developed gradually over a considerable time span.
2.
During visual inspection the maximum measured individual tube leak rate is only 50-100 drops ' min with nitrogen blanket pres-i sur:. as high as 50 psi indicating that the largest leak area is very small.
3.
Burst tests of U-bend portions similar to Row I steam generator tubes conducted by Westinghouse indicate that the burst strength at these locations is greater than in the straight sections of
" 7 179 ea w
e
1 Portland Generd BectricCc Tparty Mr. R. H. Engelken November 4, 1979 Page six the tube. Testing performed on U-bend samples having defects indicates that the burst pressure for this portion of the tube is 40 percent greater than that for the straight portion of the tube. Required pressures were several times that which could occur from a worst-case accident.
4.
The Plant trip and safety injection incident of September 6, 1979 resulted in an estimated primary-to-secondary pressure differential of at least 1650 paid in the steam generators. No increase in the measured leakage was observed as a result of this incident, indicating that an incipient failure-threshold at these conditions was not present.
5.
'The steam generator tubec are conservatively designed as described in Trojan FSAR Section 5.5.2.
6.
The ET n :mination did not show any tube wastage.
7, An emergency procedure in place at Trojan addresses the action to be taken in the event of the occurrence of a steam generator tube rupture.
8.
The radiological consequances of multiple tube rupture accidents is less than evaluated in the Trojan FSAR.
In Section 15.4.3 of the FSAR, a complete severance of one tube was assumed to occur with a concurrent fuel defect level of 1 percant. Technical Specification 3.4.8 restricts the fuel defect level to cpproxi-mately 0.27 percent. Assuuing operation with 0.27 percent fuel defects, it would take tLe simultaneous complete severance of nearly four tubes to exceed the FSAR consequences (which are will below 10 CFR 100 criteria). This is unlikely to occur if all first-row tubes with leaks or ET indications are plugged prior to operation.
Further, the current fuel defect it. vel of Trojan is averaging approximately 0.021 percent. For this reactor coolant activity it would take the simultaneous complete severance of nearly 47 tubes (about 12 percent of the total first-row tubes or about 50 percent of one steam generator first-row tubes) to exceed the FSAR consequences. This is extremely unlikely.
9.
A main steamline break accident requires consequential tube leakage much ste.ater than the 10 gpm assumed in the FSAR to exceed the FSAR radiological consequences. Based on the Tech-nical Specification 3.4.8 limit of 0.27 percent fuel defects (see Item 8), a consequential tube leak of about 37 gym would be
~7 180
.g.
- he.en e.
w..m+
-u----
e e-
Pbrtiand General BectricCorre Mr. R. H. Engelken November 4, 1979 Fage seven bounded by the FSAR analysis. For the current Trojan fuel defect level of 0.021 percent, a consequential tube leak of 480 gpa would be bounded by the FSAR analysis; this corresponds approximately to the simultaneous occurrence of a main steamline break and a complete severance of a steam generator tube (see FSAR Figure 15.4-51).
The combined probabilities of both of these low probability events occurring simultaneously is felt to be extremely low, especially considering that all first-row tubes with either leaks or indications will be plugged, and considering that the steamline break would have to be a break that occurs in the very short piping run between the Containment and the main steam isolation valve.
We believe that there is no compromise in the Plant safety by the actions proposed and that the Plant accident analyses continue to be conserva-tively assessed. Conseqrently, continued operation of Trojan as planned will not constitute an undue risk to the health and safety of the public.
i Therefore, we believe that the interests of the public will best be served by the course of action proposed.
Sincerely, M/
DJB/KM/DIH/WSO/4kk232 c:
Mr. Lynn Frank, Director State of Oregon Department of Energy Mr. A. Schwencer, Chief Operating Reactors Branch #1 Division of Operating Reactors U.S. Nuclear Regtlatory Commission 7
181 e
4