ML20214P659
| ML20214P659 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 09/17/1986 |
| From: | Wilson R GENERAL PUBLIC UTILITIES CORP. |
| To: | Zwolinski J Office of Nuclear Reactor Regulation |
| References | |
| 5000-86-1031, NUDOCS 8609240008 | |
| Download: ML20214P659 (12) | |
Text
,
r.
GPU Nuclear as any, ew Je s y 07054 201 263-6500 TELEX 136-482 Writer's Direct Dial Number:
September 17, 1986 5000-86-1031 Mr. John A. Zwolinski, Director BHR Project Directorate #1 Division of BHR Licensing U.S. Nuclear Regulatory Commission Washington, D.C.
20555
Dear Mr. Zwolinski:
Subject:
Oyster Creek Nuclear Generating Station Docket No. 50-219 License No. DPR-16 Postulated High Energy Line Break within Emergency Condenser Penetrations On August 22, 1986, members of my staff met with you and Mr. J. N. Donohew, Oyster Creek NRC Project Manager, to discuss our evaluation of high energy fluid piping penetrations at Oyster Creek (0.C.).
The evaluation was self-initiated after GPU Nuclear became aware of a condition involving a potential failure mode for Nine Mile Point 1 (NMP1) containment penetrations.
Loads calculated for containment penetrations at NMP1 due to a postulated High Energy Line Break (HELB) in the process piping within the penetrations were determined to exceed those for which the penetrations were designed.
The higher loads resulted from use of a more accurate analysis model which included both pressure and momentum effects. Although the NMP1 and 0.C.
plants are similar in many aspects, significant differences were found in the l
penetration designs.
Therefore, the NMP1 analysis could not be directly applied to 0.C. and a plant specific analysis was undertaken.
The evaluation of the 15 high energy lines which penetrate the O.C. drywell showed that 13 of the 15 penetrations could tolerate a HELB in their normal power operation alignment.
The emergency condenser condensate return and steam supply penetrations were found to have excessive stresses if a guillotine rupture of the process line within the penetration was to occur while the system was in operation.
The issue concerning the emergency condenser piping penetrations, because of their potential failure mode, was found by GPU Nuclear to be an unreviewed safety question.
8609240008 860917 ADOCKOSOOg9 PDR P
c\\
o Q
t GPU Nuclear is a part of the General Public Utikties System
c Engineering of a modification to restore the emergency condenser piping penetrations to an acceptable stress condition during a postulated HELB has been underway for approximately one month.
The design effort has been complicated, however, due to shielding plates at the penetrations inside the drywell.
In addition, the modification cost and impact on the duration of the current Cycle llR outage would result in severe economic hardship for GPU Nuclear.
The attachment to this letter provides a detailed discussion of historical background, licensing basis, new stress criteria, outage impact and justification for operation in Cycle 11, without the modification, relative to this issue. As was discussed at the aforementioned meeting, GPU Nuclear believes that adequate justification exists for a change to the O.C. plant design basis and that modification of the penetrations is not necessary, especially in light of the proposed revision to GDC 4 for BHRs.
The attached information is provided for further NRC staff review so that additional consideration can be given to a resolution.
If the final resolution requires modification, it will be performed during the next (Cycle 12R) refueling outage.
If you should have any questions, please contact Mr. M. H. Laggart, of my staff, at (201) 299-2341.
Very truly yours, S$$
RFh)
R. F. Wilson Vice President Technical Functions RFH/pa(3984f)
Attachment cc: Dr. Thomas E. Murley, Administrator Region I U.S. Nuclear Regulatory Commission 631 Park Avenue King of Prussia, PA.
19406 NRC Resident Inspector Oyster Creek Nuclear Generating Station Forked River, N.J.
08731 Mr. Jack N. Donohew, Jr., Project Manager U.S. Nuclear Regulatory Commission 7920 Norfolk Avenue Phillips Building, Mail Stop 314 Bethesda, Maryland 20014 l
,n.-
Attachment
SUBJECT:
POSTULATED HELB IN THE EMERGENCY CONDENSER PENETRATIONS i
1.0 BACKGROUND
i i
In January, 1986, Niagara Mohawk Power Corporation (NMPC) notified NRC Region I of a potentially reportable condition involving a failure mode for the Nine Mile Point 1 containment penetrations. Loads calculated for the penetrations due to a postulated High Energy Line Break (HELB) in the process piping within the penetrations were determined to exceed those for which the penetrations were designed. These higher loads resulted from use of a more accurate analysis model which included both pressure and momentum effects.
GPUN became aware of the NMPC analysis through the Oyster Creek (OC) i General Electric site representative.
In February, 1986, GPUN voluntarily initiated an investigation of the OC plant containment penetrations. Although the Nine Mile Point and OC plants are similar in many aspects, significant differences were found in the penetration designs.
The specific design features of the OC plant provide additional strength over the Nine Mile Point design.
Therefore, the NMPC analysis could not be directly applied to OC and a plant specific analysis was undertaken.
J The GPUN analysis involved hydraulic and mechanical modeling as well as evaluation of system operation for the 15 high energy lines which penetrate the OC drywell.
The initial evaluation was completed by July, 1985.
The evaluation showed that all OC penetrations could tolerate a HELB in their normal power operation alignment. The emergency condenser condensate return and steam supply penetrations were found to have l
excessive stresses, however, if a guillotine rupture of the process line within the penetration was to occur while the system was in operation.
This result necessitated a further review of the exact conditions under which rupture could occur since the emergency condenser is provided with a rupture detection system which automatically isolates the system upon l
detection of high process flow. This evaluation was completed in early l
August, 1986.
In parallel with this evaluation alternatives for eliminating the potential for penetration failure were investigated.
GPUN met with the NRC staff on August 22, 1986, to discuss the evaluations performed to date. With the exception of the emergency condenser piping, all penetrations were found to be acceptable by GPUN through a 10CFR50.59 evaluation. The issue concerning the emergency condenser piping penetrations, because of their potential failure mode, was found by GPUN to be an unreviewed safety question.
L 5952d
1.1 LICENSING BASIS Section 3.8.2.4.3 of the OC FSAR update describes the primary containment penetration design. According to this description, the penetration design condition includes:
... the pipe connected thereto asumes
[ assumes] (sic) to rupture, at random, on either side of the drywell."
The description goes on to state: "For the loading on the penetration assembly from a given pipe, the jet force loading condition resulting from the equivalent of a guillotine-type pipe break, applied at any angle, shall be defined as:
1 x nominal line pressure x cross-sectional line flow area."
Strictly interpreted, the above criteria is met for all OC pipe penetrations.
The implied criteria, however, is that all containment penetrations be capable of accepting a guillotine break in the process pipe. Applying this interpretation, the loading equation specified is incorrect in that it does not include momentum and additional pressure effects which would accompany a process line rupture within the penetration.
This latter interpretation is the basis for GPUN's determination that the potential failure of the emergency condenser penetration.3 constitutes an unreviewed safety question under 10CFR50.59.
1.2 ANALYSIS METHODOLOGY:
1.2.1 Calculation Method:
The force on the drywell steel liner was calculated by combining pressure and momentum effects, both in the process pipe and the guard pipe.
Fluid pressures, densities and velocities were calculated considering homogeneous, steady or transient flow using multi-node computer solution or Fanno Line methodology, which considered friction losses, actual geometry, flow source from both sides at the break, when appropriate, and a pipe separation during the break exceeding one fourth of the pipe diameter.
1.2.2 Allowable stress values: OC FSAR states that the allowable stress values for jet impingement loads due to HELB on or in the penetration shall not exceed 0.9 Sy (yleid strength).
This is based on the hydrostatic test requirements for a Class A vessel l
per Paragraph N-714.2 of the 1965 Edition ASME Code, Sect. III, which states:
l "If the minimum test pressure defined in N-714.1 is to be exceeded at any point in the vessel by more than 6 percent, the upper limit shall be established by the design engineer using an analysis which includes all loadings that may exist during the test.
The calculated primary membrane stress intensity shall not exceed 90 percent of the i
j specified minimum yield strength of the material."
! 5952d i
This value was a reasonable selection as an allowable at the time the plant was built, considering the state of the art then and the method of calculating the loads.
This allowable only addresses the primary membrane stress intensity and does not address other stress categories in the penetrations. Also, as there were no specific allowables stated in the code for jet impingement loads and 0.9 Sy keeps the primary membrane stress intensities in the elastic range, the criterion assured both structural and deformation integrity of the drywell boundary.
Due to a more refined and realistic methodology for calculating HELB load, GPUN considered it was appropriate to use more current criteria considering primary local membrane and bending stress intensities in containment vessels. Consequently, GPUN used the allowables for service level "D" of the 1983 ASME Code Section III Subsection NE, Figure NE-3221.5, and the Definition of Service Level "D" in Paragraph N-3113.4. Thus, the allowable stresses will be 1.5 Sy or 1.8 S.., whichever is less, where Sy is the-yield stress and S.,
is the allowable stress intensity.
2.0 JUSTIFICATION FOR CONTINUED OPERATION As discussed above, the existing condition of the eme.gency condenser condensate and steam piping has been determined to constitute an unreviewed safety question.
The safety significance of the probability of a break occurring at the specified location of concern is quite low, and second, a pipe failure at this location would show leakage and be detected long before catastrophic failure could occur.
Each of these factors is discussed in detail in the following section.
2.1 Application of Leak Before Break:
The modification to mitigate the consequences of HELB in the emergency condenser penetrations will be impractical to implement during the 11R outage due to physical plant configuration and impact on outage duration.
Therefore, the mitigation of pipe breaks will be addressed using Leak Before Break evaluation.
This approach is taken to assess the probability of a guillotine pipe rupture occurring.
This is accomplished using a combination of augmented Inservice Inspection (ISI), fracture mechanics analysis and leak detection as follows:
2.1.1 Inservice Inspection - Augmented inservice inspection was l
performed on emergency condenser piping both inside and outside containment during the 11R outage.
The purpose of l
the inspection was to detect the presence of intergranular i
stress corrosion cracking in stainless steel weld joints, in response to NRC's Generic Letter 84-11. The method of examination was ultrasonic (UT) testing. The acceptance criteria was to limit flaws to 107. through-wall per ASME code Section XI.
The inspection results are shown in attached Table 1.
It should be noted that no IGSCC was found in the piping inside the drywell. The weld inside the penetration I
was not inspected due to inaccessibility.
! l 5952d i
..m
~
2.1.2 Fracture Mechanics Analysis - Leak Before Break analysis was performed for emergency condenser piping outside the drywell.
This analysis is applicable to the piping inside the drywell since it is fabricated of the same material and has lower primary and secondary stresses as compared to outside piping.
The lower stresses have been determined using the recently completed IE Bulletin 79-14 re-analysis for condensate return and the steam piping.
The maximum primary plus secondary stress inside the drywell is 69% of the maximum primary plus secondary stress outside the drywell for condensate lines. This value is 88% for the steam lines.
The fracture mechanics analysis was performed using NRC developed criteria called " Palisades Criteria" which permitted licensees to perform a safety assessment as an alternative to system modifications (Ref. 4.1). The analysis was performed assuming that a 2t (t - pipe wall thickness) long circumferential through-wall crack was present at the most highly stressed weld location.
The crack growth rate and resultant leak rate were calculated for one month intervals until the crack length exceeded 90* of the pipe circumference under ASME Section III Level D loads and extreme conditions, such as hangers and snubbers assumed ineffective.
The analysis has indicated that there is a tendency to develop through-wall cracks in a circumferential direction.
The cracks exhibited excellent stability even when the crack length equaled 90* of the pipe circumference.
For the purpose of the analysis, it was conservatively assumed that crack instability will occur when the crack length exceeds 90* of the pipe circumference.
2.1.3 Leak Rate Calculation and Leak Detection - The results of calculations showed that the leak rate from the crack would be approximately 1 gpm at the start and it will increase to greater than 5 gpm before the crack grows to an unstable i
length. This leakage will be detected as unidentified leakage in the drywell which is continuously monitored during power operation.
This leakage is sufficiently high enough to l
be detectable by the available leak detection methods.
Two independent methods are available for unidentified leak detection.
These are: (a) drywell unidentified leak rate integrator which is monitored every four hours and, (b) leak rate recorder which provides continuous leak rate indication in the control room. Any abnormal increases in this leakage (a 2 gpm increase in a 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> pericd or a maximum of 5 gpm) will require plant shutdown by Technical Specifications.
Additionally, sufficient time (on the order of several months) exists to take appropriate actions (i.e. shutdown and inspect) between the time of leak detection and the time that a crack would grow from that point to an unstable length.
_4_
5952d l
l
p 2.2 Compensatory Action:
Emergency condenser steam and condensate piping penetrations will be visually inspected during unexpected outages when the drywell is de-inerted and accessible.
This will assure that a small crack has not initiated inside the penetration.
2.3 Probability of Break Occurrence: The probability of guillotine type rupture in the pipe is considered very low. A break of this type in the specific penetration location is considered even more unlikely.
The evaluation of conditions under which HELB will be of concern are described below:
2.3.1 Steam Supply Line - In the standby mode, (outside containment isolation valve on condensate return line closed) the calculated load on the drywell shell is within allowable for a HELB in the penetration. When the emergency, condenser is in operating mode, the calculated load will exceed allowable values, if HELB occurs in the penetration.
However, the probability of occurrence of this condition is considered ve,ry low as explained below:
2.3.1.1 The normal stresses in the pipe at the penetration are approximately the same for the conditions when the emergency condenser is in the standby mode or in the operating mode.
Therefore, no step change in stresses will occur when an emergency condenser is initiated when HELB would be a concern.
2.3.1.2 The time of operation of emergency condensers is very short.
The average number of scrams per year during the last ten years equals 3.5 scrams / year.
Although not all scrams require use of emergency condensers, this scram rate was used for conservatism in the analysis.
The amount of time the emergency condensers are used during each scram is approximately 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> for cool-down.
The emergency condensers are used in an on-off mode during this cool-down by opening and closing the i
condensate return valve. Out of this 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> cool-down reactor pressure is high enough so that the load value for postulated HELB will exceed allowable values for approximately 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />.
- Thus, approximately 3.5 x 1 - 3.5 hrs. is the time when emergency condensers are used per year (8760 hrs.)
or less than 0.04% of the time for which a hypothetical pipe break could jeopardize containment integrity.
Based on this short time of operation, l
5952d l
l
_~
~
the probability of an HELB in the penetration during i
=
emergency condenser operation is considered very low as explained below:
Probability of-HELB = 10-*/yr.
4 (Ref. NASH-1400, median value)
Failure frequency (based on expected emergency condenser operating time) -
3.5 scrams x 1 Hr. x 1 Year x 10-4 Ruptures year scram 8760 Hrs.
Year 3.995 x 10-* per year It should be noted that the HELB probability as given in WASH-1400 is generally applied to all high energy lines without consideration for the specific location of the break.
Ref. 4.2 provides probabilities of line break at specific locations.
This analysis for.large piping in Pressurized Water Reactors-(PHR) indicates that estimated probability ofalargeLOCAinducedbyplanttransientsand earthquakes is 6.7 x 10-'. This value will increase by.several orders.(per discussion with industry sources this is in the range of 10* to 2
5 10 ) of magnitude if Intergranular Stress Corrosion Cracking (IGSCC) is considered in the piping.
Based on this analysis,-the application of probability given in NASH-1400 is considered to be a conservative estimate.
2.3.1.3 As described in SRP (Standard Review Plan) 3.6.2, "An operational period is considered.short if the fraction of tin,e that the system operates within the pressure-temperature conditions specified for high energy fluid systems is about 2 percent of the time i
that the system operates as a moderate energy fluid system." SRP further state's that "through-wall cracks instead of breaks may be postulated in the
+
piping of those fluid systems that qualify as high energy fluid systems for only short operational periods but qualify as moderate energy fluid systems.
for the major operational period".
e 4
l-5952d
..s This SRP statement sets the precedence that for HELB
'=
which has a consequential effect for a very short duration of operation, a guillotine break need not be considered.
2.3.1.4 LThere are pipe break detection differential pressure switches in the flow measuring elbows of the steam and condensate lines of the emergency condensers. -A high differential pressure, equivalent to a line.
flow three times the. rated flow in any elbow in the affected loop will automatically initiate closure of all isolation valves in that loop. Upon closure of the' valves the loss of coolant inventory will be reduced.
2.3.2 Condensate Return Line - The emergency condenser condensate return line is normally below 200*F in standby mode and a HELB in the penetration can be tolerated in this mode since a single blowdown will. result.. However, there is a potential that~a HELB in the condensate return penetration may cause i
emergency condenser initiation based on low-low reactor water level.
If this initiation occurs,.the resulting double blowdown into the affected penetration results in a load that-would exceed the allowable for the drywell shell.
For this condition, through-wall cracks instead of break will be postulated and leak before break analysis will apply.
3.0 EVALUATION OF CORRECTIVE ACTIONS f
Corrective actions during the llR outage cannot be provided due to the following reasons:
3.1 Complication of Design Effort - There are shiel' ding plates at the penetration inside the drywell. There are 3 plates, each 1-1/2" thick and 32" diameter at each penetration. These plates are welded to the drywell steel shell.
The plates are located just upstream of 4
the elbow on the piping.
The configuration of:these plates is complex and it will take substantial effort to design a modification to accommodate HELB in the penetration. A simple fix to install a flow restrictor similar to NMPI is not feasible for the Oyster Creek penetration design due to the existence of these plates.
J i
i 5952d
3.2 Cost of Modification - The cost of the modification to install flow restrictors is estimated at $600,000. 'In addition, there will be outage related cost due to the extension of the outage for approximately 5 weeks.
3.3 Impact on Outage - Currently, OC is in a refueling mode.
During refueling, the area near emergency condenser penetrations is not accessible. Also, presently, there is no scaffolding in the area.
close to the penetrations inside the drywell and it will need to be installed to permit access for installation. Modification installation is complicated by the existence of shielding plates at the penetrations. The reactor hydro test and the integrated leak rate test (ILRT) cannot be performed before or during the installation and has to be performed before restart.
Based on those limitations, the outage, which is currently scheduled to end October 12, 1986, would have to be extended by about 5 weeks to install the modification.
3.4 Radiation Exposure - Due to the complexity of installation and the difficult access to this area, it will take significant exposure to accomplish this modification.
It is_ estimated that approximately 40 Man-Rem exposure will be required for this modification.
Based on the above discussion,.it will be impractical to install a modification to address HELB in the emergency condenser penetrations during the 11R outage.
Engineering will continue to develop a modification.
However, it is felt that the evaluation presented above is adequate to justify a design basis change.
This position is further reinforced by the proposed revision to GDC 4 for BWRs. A resolution to this issue will be achieved prior to restart from the next (12R) refueling outage.
The modification will be installed if eventually required.
Appropriate changes to the FSAR will be incorporated in the next update (for 1986) to document evaluation results and the utilization of new stress criteria. However, conclusions regarding emergency condenser penetrations will not be incorporated pending resolutior. of the design basis change vs.
modification issue.
l l
t 5952d 1
+,
4.0 ~
REFERENCES:
4.1 Letter dated 10/15/84, P. B. Fiedler to D. M. Crutchfield, SEP Topic III-58, Pipe Break Outside Containment.
4.2 NUREG/CR-2189, Vol. I through 9, Probability of Pipe Fracture in the Primary Coolant. Loop of a PWR Plant.
-4.3 Standard Review Plan, SRP 3.6.2.
j i
f 5952d
~
s TABLE 1 EMERGENCYCONDENSERPIPINGINSERVICEINSPECTIONRESULTS TotalNo.of No.ofWelds Results Welds Inspected Steampiping InsideDrpell 25 6
NoIGSCC OutsideDrpell 63 20 1IGSCCcrack*
Condensatepiping InsideDrpell 29 6
NoIGSCC OutsideDrpell 64 20 NoIGSCC
- Mayhavebeenmisinterpretedduring10Rinspections,i.e.
dispositionedasundercut
.