ML20091B798
| ML20091B798 | |
| Person / Time | |
|---|---|
| Site: | Ginna  |
| Issue date: | 07/01/1983 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML082380335 | List:
|
| References | |
| FOIA-91-106 NUDOCS 9204020023 | |
| Download: ML20091B798 (16) | |
Text
. - - - _
h 7
i ENCLOSURE 1 STAFF SAFETY EVALUATION CONCERNING ANALYSIS OF
~
FLANT REspDRWDMING JAhUARY 25, TET5TEDTTIREETDT~ TUBE FAILURE
~~
~
AT TfIE~R.E. G1hkA hUCLEAR PDVER PLANT lITEM 7)_
Attachment B of the licensee response (reference 1) to the staff Safety 1
Evaluation Report (NUREG-0916, reference 2) addressed the j
thermal. hydraulic an61ysis of the steam generator tube rupture (SGTR) using the LOFTRAN computer program (reference 3). The licensee comitted to perform a detailed thermal-hydraulic 6nalysis of system behavior during the incident tu verify phenomena, including void formation.
The LOFTRAN computer program has been approved for use in Safety c
Analysis Reports for analyses of Chapter 15 design basis events
-(reference 4). -Tiese analyses include the steam generator tube rupture i
event. LOFTRAN was shown to produce conservative licensing evaluations y
by use of proper selection of input data and by use of the models i
employed in the LOFTRAN computer program.
These input data and models are not necessarily representative ';f applied to a real (best estimate) calculation.
Several limitations were identified which were significant when applied to the Ginna event. These were considered by the licensee and a number of auxiliary calculations were presented to provide more o
detailed modt111ng of localized events not treated directly in LOFTRAN..
The purpose of this evaluation was 'to verify the thermal-hydraulic phenomena during the incident, including void formation.
y l
LOFTRAN is somewhat limited by the modelling of the upper head region, steam generator secondary side, and primary-to-secondary leakage.
The-9204020023 910709 NLIA
-106 PDR
,j y, gj g / 7,'/f j go
[
l
l i
up'per head modelling assumas homogeneous, thermodynamic equilibrium 4
conditions during flashing of the upper head fluid.
Refilling of the upper head region is artificially constrained to simulate non-equilibrium behavior.
Effectively, the upper head region can not l
refill during natural circulation flow. As a result of this model.
LOFYRAN is limited in it's ability to predict void collapse in the upper head.
A faster represturization is calculated which exacerbates the SGTR event by increasing leakage and reducing HPI makeup.
Furthermore, flow into the upper head region via guide tuces is not represented.
Consecuently, the calculated upper head fluid temperature may be unrealistic for plants with sna11 upper" head " spray" nozzles._sech as "inna.
LOFTRAN is also limited by the homogeneous, saturated conditions within the seconc'ery which prometes an unrealistic 1y letnargic tube bundle region temperature response to AFW flow and_ secondary-to-primary heat transfer, in addition, these conditions result in artificially reduc'ed steam generator pressures when no steam flow occurs since the steam is effectively assumed to be in contact with the steani generator tubes.
The break flow calculations within LOFTRAN are based on conservative, i.e., maximum flow, critical flow correlations. The t
accuracy of these correlations in predicting critical flow trends over a wide range of system conditions is uncertain.
Furthermore, the break flow modelling does not consider flow resistance through the failed tube, or fluid. temperature variations between the steam generator in'et and outlet plenums, Finally, LOFTRAN does not permit reverse flow to l.
occur in the coolant loop to which the pressurizer is connected.
For l
the results presented, the pressurizer was modelled on the intact loop I
l although during the Ginna event the pressurizer was on the f aulted loop.
y l
1983: Ed Encl. July 1, 1983 i
i l3
~. _. _. _ _ _ _..
i t
l This may result in unrealistic loop flows during refilling of the pressurizer.
[
The parameters addressed for this evaluation were: (1) primary system pressure. (2) reactor coolant flow, (3) reactor coolant ten;perature, (4) i pressuriter level, (5) break flow, (6) reactor coolant voiding, and (7) steam generator overfill.
A long term recovery evaluation was diso provided, but was not based on LOFTRAN calculations.
r i
Pre-Trip System, Response The pre-trip system response analysis was performed, using, as input.
4 normalized core power and secondary pressure, as obtained frcm the plant data recorders, to evaluate the reactor coolant temperature and pressure response. Pressurizer pressure and level calculations agreed well with it was also denenstrated that the pressure and level the plant date, arc significantly affected by the coolant temperature trends.
Post-Trip System pesponse The post-trip system response analysis was performed using the recorded intact rteam generator pressure as input to the LOFTRAN computer Because of the homogeneous equilibrium secondary side model program.
used in LOFTRAN, it was necessary to artificially steam the generator to reproduce the recorded subcooling in the associated cold leg for the transient period-f rom 7 to 16 minutes following the SGTR.
l l
Reactor Coolant Pressure l
l 1983:Ed Enc 1, July 1, 1983
)
)
A snall void may have developed in the upper head region during the initial depressurization (4 to 5 minutes) although LOFTRAN oid not predict flashing at this tire.
Following safety injection, LOFTRAN calculated a more rapid repressurization than was observed in tne plant
+
response.
This has been attributed to the collapse of en upper head void during the actual event. The LOFTRAN calculation overestimated the pretsure response by less than 100 psia for the oeriod prior to the PORV i
cycling.
An upper head void was generated during PORV cycling.
Following isolation of the f ailed open PORV, the LOFTRAN calculation showed a more rapid repressurization than was experienced at the piant.
This was attributed to the LOFTRAS limitation which inhibits refill of the upper head region void during natural circulation.
The tctual plant response, a slower repressurization, was e.ttributed to at least partial refill of the upper head region, Reactor Cool _ ant Flow The LOFTRAN calculation indicated that natural circulation through the intact loop was maintained between 31 and 4t, until reactor coolant pump (RCP) restart.
Flow stagn3 tion in the faulted loop was calculated to occur at about 45 minutes. The LOFTRAN calculation did not support significant reverse flow throu;h the f aulted loop, however, the effect of the break flow model on the calculated loop flow was uncertain. An evaluation perforned by the licensee assuming reverse flow was present could not be supported by the actual plant respceses observed.
It was concluded that sustained reverse flaw was unlikely.
These results support the existence of a counter-current type of flow regime upstream of the injection nozzle. However, LOFTRAN does not rodel this type of 1983:Ed Encl. July 1, 1983
mixing.
A review of existi.y experimental data (Creare mixing experiments, references 5 and 6) suggests that, indeed, a significant portion of the safety injection flow into a stagnant loco woulo
~
propagate upstrean of the injection norzle, and result in the type of a
counter-current flow observed.
The results of this evaluation indicated 4
l that a minimum loop flow of 21 lbm/sec existed. -
Reactor _C_oolant Temperature The inability of LOFTRAN to treat complex flow regimes, and the i
requirement that the piessuri7er be in th? Ioop without reverse flow, resulted in calculated faulted loop temperatures significantly below the observed values.
In. addition LOFTRAN was unable to predict the temperature increase observed following isolaticn of the PORV. The LOFTRAN modelling of the pressurizer in the intact loop, r,ay have crtificially promoted flow toward the vessel.
In order to estimate the expected minimum temperature, a mixing evaluation was performed assuming a sustained loop flow rate of 21 I
lbm/sec, as indicated above.
The calculated minimum temperature was 200*F, as compered with the observed value of 265'F.
F i
Pressurizer Level The pressurizer levt1 response obtainea from the LOFTRAN calculation compared favorably with the observed data, although some differences were evident. The initial decrease in level was predicted quite well.
-The data indicated the level returned on span when the charging pumps i
were started, while LOFTRAN did not predict this to occur entil the p0RV 1983:Ed Encl.
-S-July 1, 1983
. -.. -. - - =
.... ~.
was first cycled. When the vessel head water began cb flash, a rapid t
filling of the pressurizer was observed and calculated.
In the observed data the level went off scale high, while the calculated response indicated the level a still be on scar. This war. atsributed to en initially lower level calculattd by LOFTRAN, and aho my have been due to a slightly underessimated voiding of the upper head because of the LOF1RAN modelling which assumes the upper head is a homogeneous region.
Break flow.
The break flow calculation, primary-to secondary itakage, used in LOFTRAN assumed on effective break area and a modi,fied Zaloudek critical flow correlatf or. The LOFTRAN c61culated faulted steam generator pressure was underpredicted, as a result of the modelling used.
f onsequently, secondary-to-primary flow was not calculaNd by LOFTRAN when the P0RV was opened. A more detailed model was developed to assess the limitation of the LOFTRAN break flow model and the ef fects on the enalysis results.
It was concluded that the LOFTRAN model, with the exception of reverse flow, provided a reasonable estimate of the break flow. The calculated lower pressure in the faulted loop steam generator results in an overestimate of the primary-to-secondary leakage, which is a conservative result when applied to SAR licensing aulyses concerning radiological consecuences.
Reactor Coolant Voidirro I
Upper head void, although not calculated by LOFTRAN, prior to RCP trip was estimated to be less than 132 cubic feet. Any steam bubble in the
. upper head while the RCPs were running would have been quidly l
l 1983:Ed Encl. July 1, 1983
i
'l cundensed.
The observed response indicated this was the c;se. Upper i
head voiding was both indicated and calculated when the PORY was opened.
The side of the void was estimated to be approximately 305 cubi-feet, the upper head volume. The observed response indicated et least partial refill of the void, while LOFTRAN effectively inhibits refill.
The size of the void when safety injection was terminated could not be determined from available data. A mass balance evaluation suggested that a maximum void of 125 ;ubic feet could have been present at this time.
Steem Generator Overfill The LOFTRAN csiculation indicated overfill of the faulted steam generator and steam line, resulting in lif ting of the steam generator safety valve, somewhat earlier than was observed. This was attributed to early termination of steam relief to the condenser in the LOFTRAN calculation (in order to better simulate the transient response),
resulting in about 11.000 lbm more inventory being in the steam generator.
In addition, more primary to secondary leakage may have becn calculated by LOF1RAN. The combined effect was the earlier filling of i
i the secondary side volumes.
1 l
L,ong Term Recovery The LOFTRAN calculation was terminated when safety injection was terminated since the homogeneous equilibrium model on the secondary side l
overestimated the primary-to-secondary pressure differential and therefore 'eakage through the failed tube.
Summa.) and Conclusions 1983:Ed Encl. July 1, 1983 i
Lu.
____.__._s 4
There are a number of limitations inherent in the t.0FTRAN compJter program. These limitations, in particular the homogeneous equilibrium mode) *and one dimensional flow models, limit the capability of LOFTRAN to calculate complex transients, such as the Ginna incident. These limitations are the result of the conservative nature of L0rTRAN, as developed for SAR licensing analyses.
The licensee clearly identified these limitations and performed auxiliary calculations to support and supplement the LOFTRAN results.
These auxiliary calculations employed standard mass and energy balance techniques to address the limitations in the LOFTRAN results. These calculations were also reviewed by the staff and found to be acceptable.
These analyses support the verification of the system phenomena, including void fonnation, as required by NUREG-0916.
The staff finds the information provided by the licensee acceptable for the evaluation of the Ginna SGTR event of January 25, 1982.
f I
I l
1983:Ed Encl. July 1, 1983
,m
,g...
wor.
g.,-a---
-,..w--
- n. y vs w-m e.
,m
.m..-
-.m.m.....-e
,~e=.
~im..-
I REFERENLES 1.
Docket No. 50-244, " Response to Safety Evaluation Report-NVREG-0916 Steam Generator Tube Rupture incident R. E. Gir.na Nuclear Power Plant Docket No. 50 224
- 1etter from J. E. Maier to D. M. Crutchfield, November 22,1982fDCS8211290410).
2.
NUREG-0916, "NRC Report on the January 25, 1982 Steam Generator i
Tube Rupture at R. E. Ginna Nuclear Power Plant," USNRC, April 1982.
1 3.
L. A. Campbell, et.al., "LOFTRAN Code Description " WCAP-7878 Rev.
3, January 1977.
4 Safety Evaluation Repo. t, "LOFTRAN Code Description," memorandum from R. W. Houston to G. C. Lainas, June 27, 1983.
5.
J. A. Block, " fluid Thermal Mixirg in a Model Cold Leg and Downconier with loop Flow," CREARE,,Inr, Hanover, New Hampshire, 4
EPRI-NP-2312 April 1982.
6.
P. H. Pathe. " Transient Cooldown in a Model Cold leg and Downconsr,";CREARE, incl., Hanover, New Hampshire, EPRI-NP-3118 (Interim Report), May 1983.
~
t 0?
l 1983:Ed Encl. July 1, 1983
--<--e
.-,,.-m-
~--nw.-+--s,-+--
- w v
w.,
-w.,E,--
-.-w.,-
ev.r,-
rr wvw e ve w w e-w r-e
.w---w www e--
m e-.m e r au - r e-4
ENCLOSURE 2 STAFF SAFETY ~TVIEUATfCW CONCERN 1hG REACTorCEUliTTUMT IITUTATGkTFETERI A AND REACTOR COOLANT PUMP RE5TTJTTiaTUTXT0rTC l @ lNNA NUCLEAR P,07E TkU KT~{TTER5 6 & 9)_
Attachment C of the licensee response (reference 1) to the staff Safety Evaluation Report (NUREG-0916, reference 2) ad:Iressec' titernate reactor coolant pump (RCP) trip criteria.
Attachment D (reference 1) addressed toe RCP restart requirer ents following a steam genentor tube rupture (SGTR).
The alternate RCP trip criteria reviewed included: (1) tht current
-criterion of reactor coolant system (RCS) pressure,below 1285 psia
~
(including instrumentation uncertainty), (2) reactor coolant subcooling, (3) a secondary pressure dependent RCS pressure value, (4) reactor vessel icvel, and (5) retctor coolant pump electrical current.
The reactor vessel level and reactor coolant purap current methods were dismissed because of the need for substantial equipment modification and the need for extensive analytical and expcrimental ef forts.
Several LOFTRAN annlyses were performed for a spectrum of SGTR events to assess the margin to RCP trip following a SGTR.
It was concluded that the secondary. pressure dependent RCS pressure method provided the most potential for preventing RCP trip for a SGTR.
It was also concluded that this method was beneficial only if instrement uncertainties were evaluated for normal containment conditions.
If abnormal containment condition occur (increased pressure and temperature), then RCP trip a
could be expected, based on increased instrume.nt uncertainties.
Normal 1983:Ed Encl. July 1, 1983 1
. ~... - - -
4 containment conditions are expected during a SGTR.
It was therefore concluded that %henever the need for pump trip is ecdressed in the procedure, the operator would be required to evaluate the containment condition and to select the appropriate criteria depending upon containment conditions.
[e.g.,fornormalcontainmentconditionuse secondary pressure dependent RCS pressure; for abnorni condition use 1785 psia RCS pressure].
This approach would prevent RCP trip for a design basis SGTR, while still providing for a required pump trip in the I
event of a LOCA".
'n response to Generic Letter 83-10d (reference 3), the licensee (reference 4) has-indicated that tne resolution for RCP trip will be addressed in a Westinghouse Owners Group (WOG) submittal, scheduled for December 1983. The RCP-trip setpoints will be incorporated in Revision 1 of the WCG Erargency Response Guidelines,. scheduled for July 31, 1983.
The llcensee has comitted to implement the revised criteria into the existing emergency procedures and provide operator training within 2 months of receipt of the revision (provided the necessary instrumentation is currently available). Based on the information provided in reference 4, the staff believes the necessary instrumentation is currently available in the plant.
Pased on the information provided in Attachment C, and on the licensee's commitment to the-WOG resolution of the pump trip issue, the staff concludes that the licensee is cognizant of the RCP trip issue and.is in conformance with the requirements of Gener'c Letter 83-10d.
The licensee has also evaluated the RCP restart criteria to assess the 1983:Ed Encl. July 1, 1983
\\..
potential fo. coolant flashing and loss of pressurizar control during pump restart following a SGTR.
The issue of concern is loss of pressurizer level ( and unavailability of pressurizer hetters to control pressure) resulting from the collapse of a sufficientiv large steam bubble in the vessel upper head af+.er pump restart. The collapsed bubble draws water from the pressurizer and ieduces the reactor coolant subcooling.
The current emergency operating procedures in place at Ginna were reviewed to determine if indicated pressurizer leve,1 and reactor coolant subcooling would be maintained follow RCP restart after a SGTR event.
It was concluded that the reactor restart criteria are sufficient to ensure both indications are maintained.
In some cases, for large enough steam bubbles, the level may decrease below the minimum level required for operation of the heaters.
In these cases guidance is p-ovided to restore level using normal charging and safety injection pumos.
It was also concluded that the current RCP restart criteria tay not be appropriate for other accidents or multiple f ailure events where safety concerns exist.
Based on the information provided in Attachment 0, the staff concludes that the Ginna RCP restart criteria are sufficient within the context cf the steem generator tube rupture emergency operating procedures to ensure that indicated pressurizer level and reactor coolant subcooling would be maintained.
1983:Ed Encl. July 1, 1983
{
REFERENCES 1.
Docket No. 60 244, " Response to Safety Evaluation Report -
NUREG-0916 Steam Generator Tube Rupture Incddent R. E. Ginna N'uclear Power Plant Docket No. 50-244 " letter from J. E. Muier to D. M. Crutchfield, November 22, 1982 (DCS 8211290410).
2.
NUREG-0916, "NRC Report on the January 25, 1982 Steam Generator Tube Rupture at R.E. Ginr.a Nuclear Power Plant, USNRC, April 1982.
3.
Ger,eric letter 83-10d, " Resolution of TM1 Action item II.K.3.5, Automatic Trip of Rehetor Ovotant Fu. ps," let ter f rom D, G.
Eisenhut to Licensee, February 8,1983.
4.
Docket No. 50 244, " Response to Generic Letter 83-10d, Automatic Trip of Reactor Coolant Pumps R. E. Ginna Nuclear Power Plant Docket No. 50-244," letter from J. E. M6ier to D. M. Crutchfield, April 22, 1983.
4 J
o 1983:Ed Enc 1, July 1, 1983
O (O
ENCLOSURE 3 STAFF SAFETY EVALUATION CONCERNING EVAldKTION OF POTENTIAL STAGNATION TRANSIENTS AND SCENAR101"FOR THE R. E. GlhNA NUCLEKR POWER PLANT (ITEM 18 & 19)
Attachment F of the licensee responses (reference 1) to the staff Safety EvaluationReport(NUREG-0916, reference 2)addressedpotential situa*: ens which could lead to loss of natural circulation following reactor coolant pump (RCP) trip. Loss of natural circulation can result in cold safety ir.jection water entering the reactor vessel downcomer and increase tne likelihood of having a pressurized tnermal shock (PTS) event.
The situaticns addressed were: (1) inadequate core neat generation (decay heat fractions less than 0.5 percent of full power), (2) loss of reactor coolant system inventory (small break LOCA), (3) inadequate symmetric heat heat removal (loss of heat sink), and (4) non-synnetric heat 7emoval (overcooling in one steam generator, due to steam line break or steam generator tube rupture resulting in the isolation of one steamgenerator).
Following a period of loss of natural circulation, starting the RCPs, using the PORV to depressurize the reactor coolant system, or a subsequent break in the hot leg or reactor vessel upper head region can s
draw the cold safety injection water into the downcomer. During extended periods of loss of natural circulation (on the order of 20 minutes), safety injection will result in low downcomer temperatures.
The staff has evaluated PTS for Westinghouse plants in SECY-82-465 (reference 3) bated on the Westinghouse Dwners Group PTS program 1983:Ed Encl. July 1, 1983
(reference 4).
This evaluation included the consequences of the events described above.
In SECY-82-465 the staff concluded that the risk associ'st 'd with PTS, including events which lead to iuss of natural circulation, is acceptable for plants with nil-ductility trantition reference temperatures below the screening criterion value of 300'F
~
(circumferential welds).
The reference temperature calculated for R.E.
Ginna was 213*F (as of December 31,1981).
The Ginna SGTR event of January 25. 1982 was also evaluated in SECY-82-465.
It was concluded that for a reference temperature below 378'F, no PTS r" lated vessel f ailure would occur for this event.
~
a t
1 i
l r
l 1983:Ed Encl. July 1, 1983
REFERENCES 1.
Docket 50-244, " Response to Safety Evaluation Report HUREG-0916 Steam Generator Tube Rupture incident R. E. Ginna Nuclear Power P'lant Docket No. 50-244 " letter f rom J. E. Maier to D. M.
Crutchfield, fiovember 22,1982(DCS8211290410).
2.
NUAEG-0916, "fiRC Report on the January 25, 1982 Steam Generator Tube Rupture at R. E. Ginna Nuclear Power Plant," USNRC, April 1982.
3.
SECY-82-465, " Pressurized Thermal Shock (PTS)," Noveriber 23, 1982.
4 "Surinary of Evaluations Related to Reactor Vessel Integrity," WOG Letter 0G-70, from 0.D. Kingsley to H. Denton, May 28, 1982.
~
1983:Ed Encl. July 1, 1983
- - - - - _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ ____ _ ____ _ _______________________ _ ______ ______ ___ ____ _ _ _ _ _ _