ML20033B982
| ML20033B982 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 11/24/1981 |
| From: | Heider L YANKEE ATOMIC ELECTRIC CO. |
| To: | Eisenhut D Office of Nuclear Reactor Regulation |
| References | |
| FYR-81-155, GL-81-21, NUDOCS 8112020544 | |
| Download: ML20033B982 (9) | |
Text
O
- *^*"* ' **'**
YANKEE ATOMIC ELECTRIC COMPANY 2.C.2.1
~
FYR 81-155 ffh 1671 Worcester Road, Framingham, Massachusetts 01701
, YANKEE 60 t
9.
A (TIg 6
November 24, 1981 OLO t th t_
DE01 198 4 Tg
]b.MM,7 h United States Nuclear Regulatory Commission Washington, D. C.
20555 3
Att ention :
Mr. Darrell G. Eisenhut, Director N
Division of Licensing
References:
(a) License No. DPR-3 (Docket No. 50-29)
(b) USNRC Letter to All Licensees, dated May 5,1981 (Generic Letter No. 81-21)
(c) YAEC Letter to USNRC, dated August 21, 1981 (WYR-80-96)
Subject:
Natural Circulation Cooldown
Dear Sir:
Enclosed please find our response to your request in Reference (b) to provide the following information:
1.
A demonstration (e.g., analysis and/or test) that controlled natural circulation cooldown f rom operating conditions to cold shutdown conditions, conducted in accordance with your procedure, should not result in reactor vessel voiding; i
l 2.
Verification that supplies of condensate grade auxiliary feedwater are sufficient to support your cooldown method; and l
3.
A description of your training program and the provisions of your I
procedures (e.g., limited cooldown rate, response to rapid change in pressurizer level) that deal with prevention or mitigation of reactor vessel voiding.
l An analysis has been completed to further support the information submitted in Reference (c).
Based on the results of this analysis, we believe j
the possibility of upper head voiding occurring during natural circulation cooling at the Yankee Plant is remote:
design features of the reactor vessel allow for relatively efficient cooling of the upper head region. Furthermore, operators are trained to recognize symptoms of the St. Lucie event and are prepared to implement proven methods for safe recovery. Procedural i
restrictions are imposed to ensure that natural circulation cooling can be 8112O20544 811124 PDR ADOCK 05000029
$0 p
, bl l l
~.
Mr. Darrell G. Eisenhut, ' Director November 24, 1981 Page 2 performed safely. ~ Additionally, supplies of condensate-grado feedwater have been verified as more than sufficient to support the Yankee Plant natural circulation cooldown method.
We trus t this information is satisfactory; however,. if you have any ques tions, please contact us.
Very truly yours, YANKEE ATOMIC ELECTRIC COMPAN'l y ff!
U L. H. Heider.
Vice President Enclosure COMMONWEALTH OF MASSACHUSETTS)
)ss MIDDLESEX COUNTY
)
Then personally appeared before me, L. H. Heider,' who, being duly sworn, did state that he is a Vice President of Yankee Atomic Electric Company, that he is duly authorized to execute and file the foregoing reques t in the name and'on the behalf of Yankee Atomic Electric Company, and that the statements
~
therein are true to the best of his knowledge and belief.
/
VY i
Armand R. Soucy Notary Public My Commission Expires September 7, 1984' i
I 1
I.
4 Enclosure Natural Circulation Cooldown.
at the Yankee Nuclear Power Station I. Introduction The subject of natural-circulation (NC) cooling was previously addressed in Reference E-1, which stated that the possibility was considered 4
remote that Yankee Nuclear Power Station (YNPS) would ever experience an event similar to the 6/11/80 occurrence at the St. Lucie Plant. Reference E-1 further stated the belief that, even if fluid voiding was postulated to occur in the upper head region during NC cooldown, there is no substantial adverse impact on plant safety. This same conclusion was repeated in Generic Letter No. 81-21, which emphasized the NRC's conclusion that there was no " direct safety concern" during the St. Lucie event because adequate core cooling was maint ained. Pursuant to concerns over the St. Lucie event stated in Generic i
Letter No. 81-21, information is provided in the following sections to show for YNPS that reactor vessel design features, in conjunction with special operator training and procedural controls for NC cooling, ensure that t
undesirable void formation in the upper head region can be prevented.
i YNPS design features provide for effective NC cooling of the upper head region by maintaining adequate subcooling there during forced-circulation operations; and by establishing flow communications during NC cooling between the core inlet annulus, upper head, and core upper plenum regions. A detailed analysis, performed for YNPS using the RETRAN computer code, shows that NC cooling is both effective and controllable. This analysis is discussed further in Section II.
In addition to the YNPS reactor vessel design features that provide e ffective cooling of the upper head region fluid, the maximum NC cooldown rate I
allowed is 20 F/hr.
Notably, the NC cooldown rate for the first three hours during the St. Lucie event was 60 F/hr, Reference E-2.
This rate of cooldown at St. Lucie, in conjunction with using auxiliary spray via charging pumps to cool the pressurizer, was a major contributing factor that caused voids to form in the upper head region following loss of component cooling water. When consideration is given to YNPS for its favorable design features 1
i
)
for eff ective NC cooling and the much more stringent procedural limitiation on NC cooldown rate, an event similar to St. Lucie's could not reasonably be an ticipated. The YNPS operator training program and NC cooling procedure, which specifically address the St. Lucie event, are further discussed in Section III.
Appropriate procedural control of NC cooldown, coupled with efficient heat removal for the upper head region, will permit expeditious cooldowns to shutdown cooling system entry points where condensate grade feedwater supplies are no longer needed.
In contrast, for St. Lucie, a proportionally larger volume of feedwater is required because stagnant fluid conditions in the upper head region require cooldown times nearly twice as long as for YNPS. Thus, supplies of condensate-grade auxiliary feedwater necessary to support NC cooldown are relatively small f or YNPS.
A calculation wae performed to demonstrate adequacy of YNPS feedwater supplies for NC cooling and is discussed in Section IV.
II.
REIRAN Assessment of NC Cooling Without Upper Head Voids A computer code simulation of controlled NC cooldown was performed using the RETRAN-02 computer code. A model of YNPS was specifically created for this purpose to include in significant detail the reactor vessel in te rna ls. The purpose for such model detail was to allow accurate predictions of the buoyancy-controlled NC flowrates following termination or loss of forced-circulation cooling.
Special attention was devoted to properly represent metal mass, flow pathways, and couponent elevations.
The REIRAN evaluation consisted of simulating (1) a reactor scram from full power simultaneously with (2) tripping remaining main coolant pumps powered from off-site sources, and (3) subsequent cooldown at the procedural limit of 20 F/hr to shutdown cooling entry conditions of 300 psig and 330 F.
During the depressurization, pressure-temperature limits in the Technical Specifications were observed analytically by using a conservative schedule for pressurizer spraying. Normally, system depressurization during controlled cooldown is achieved via cycling spray flow to the pressurizer so that higher subcooling levels are ensured. Pressurizer level control is maintained via charging flow into a main coolant loop.
i Procedural limitations during NC cooling were thus tested analytically to verify that no void formation would occur in the upper head region. The procedural guidelines verified by the RETRAN evaluation were:
(1) minimum subcooling requirement that highest reactor e ssel temperature be at least 40 F below saturation; (2) maximum cooldown rate requirement that 20 F/br not be exceeded; and (3) complianr:e requirement that pressure-temperature limits of plant Technical Specifications be observed.
Results of RETRAN analysis for NC cooling confirmed the information provided by YAEC in Reference E-1.
Namely, results showed that direct injection from the reactor vessel inlet annulus into the upper head region of nearly 1% of the total system flowrate maintained temperatures there during normal operations at near - T n
ns.
Dudng forced cheulation C
cooling, enthalpy transport in the upper head occurs via upward flow from the vessel inlet to the upper head through the upper head injection nozzles, proceeding downward through the guide tubes to the upper plenum. Following termination of system flow via main coclant pumps, reversal of the upper head circulation pattern occurs due to buoyancy forces within the reactor vessel.
Enthalpy transport from the upper head during natural circulation cooling occurs via upward flow from the core upper plenum through the guide tubes, and then by downward flow toward the vessel inlet annulus via the injection nozzles.
The principal mechanism for upper head metal heat transfer is low-flowrate natural convection cooling, which is supplemented slightly by metal-to-fluid heat condt.ction.
In the RETRAN analysis, ambient heat losses from the reactor vessel were conservatively neglected and the initial upper
- head fluid temperature was conservatively taken as T instead of T
- C The upper head fl.uid temperatures were shown to closely track T during cooldown, so that upper head fluid subcooling levels could be adequately monitored fo" compliance with procedural requirements. Depressurization rate variations that could be expected to occur during NC cooling were shcyn to have only a small effect on upper head fluid temperatures.
The elapsed time predicted for h ' cooling te shutdown cooling system entry conditions were about 6.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> for T and about 9.8 hrs. for the ave upper head temperature. In comparison, because of relatively poor fluid communication for their upper head region, the elapsed time for St. Lucie's NC cooldown to shutdown cooling is about 25 hours2.893519e-4 days <br />0.00694 hours <br />4.133598e-5 weeks <br />9.5125e-6 months <br /> for upper head fluid temperature. The considerably shorter NC cooldown times for YNPS over St. Lucie reduce significantly the required amounts of condensate grade auxiliary feedwater to support this evolution.
III.
Operator Training for Natural Circulation Procedures Operator training focuses on the St. Lucie event by emphasizing the need to maintain the upper head region fluid subcooled at all times.
Furthermore, a more stringent cooldcwn rate limitation is required for NC cooldowns, in comparison to other t,;,es of forced circulation cooldown procedures. This special requirement for NC cooling is attributed in this procedure to " low flow in the reactor head region [that] will cause this region to remain at a temperature higher than the main coolant temperature
during NC cooling.
In addition, the YNPS emergency procedure cautions the operator that unexplained pressurizer level increases during NC cooling are indications of insufficient subcooling in the upper head. Use of pressurizer heaters to maintain sufficient subcooling is also discussed. In the remotely possible event that symptoms of upper head voiding are observed, operators are trained to resort to the same feed-and-bleed method that was su:cessfully used at St. Lucie to safely recover from this occurrence.
According to independent analysis performed by Westinghouse, Reference E-3, the procedural method for aC cooling at YNSP will adequately prevent undesirable void formation. This study, performed by the Westinghouse Owners' Group, concludes that there are "no real safety issues attributable to upper head flashing" during NC cooldown.
The RETRAN analysis described in Section II above was reviewed against and shown to be reasonably consistent with the Reference E-3 Westinghouse analysis, when YNPS design features are accounted for. Operator training in NC cooldowns at YNPS reflects this contemporary understanding and effectively uses the hindsight provided by the St. Lucie's experience.
IV.
Verification of Adequate Supplies of Condensate-Grade Feedwater The Technical Specifications for YNPS require that at least 85,000 gallons of water mast be available in the storage tanks for primary water (PWST) and demineralized water (DWST). Maximum capacities for these two tanks are 135,000 gallons and 30,000 gallons, respectively. They normally will contain about 120,000 gallons and 25,000 gallons, respectively, for a combined total far in excess cf the Technical Specifications' requirement.
A calculation was performed to verify that this supply of condensate grade water was adequate for an NC cooldown evolution.
RETRAN results, discussed in Section II, showed that in order ta reach the shutdown cooling system entry conditions NC cooling must occur for about 6.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> to achieve the necessary T reduction and about 9.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> for the necessary reduction of upper head region temperature. The decay heat energy that must be dissipated during a 10 hour1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> period af ter scram is approximately 320(10)6 Btu, per Reference 4.
During the cooldown, an additional 62(10)6 Btu must be dissipated in order to reduce moderator temperature and remove stored heat from the main coolant system. Approximately 55,000 gallons of feedwater will be consumed during the 10 hour1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> cooldown (20 F/hr). Note, however, that the shutdown cooling system can be placed into operation much earlier, at about 6.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> according to RETRAN analysis, when T has been reduced to the entry point temperautre of 330 F.
Thus, 55,000 gallons is a conservatively large value for purposes of verifying the adequacy of feedwater supplies. A 15 hour1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> cooldown (13 F/hr) would require about 69,000 gallons of feedwater.
In either case, the Technical Specifications' requirement is sufficient.
Moreover, the normal condensate-grade water supplies in the DWST and the PWST provide nearly twice the amount of feedwater required for even a 15 hour1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br /> cooldown.
V.
Summary and Conclusions During NC cooldown at the maximum rate allowed by procedure, main coolant system pressure can easily be maintained above the saturation value for upper head region fluid. Also, the minimum subcooling requirement imposed by procedure will ensure that voiding in the upper head region is extremely unlikely. RETRAN analysis confirms that adequate fluid mixing occurs in the upper head during NC cooling, so that fluid temperatures there will track T,
relatively closely during cooldown.
Operator training is conducted on the principles of NC cooling, with emphasis on the St. Lucie experience.
In addition, the emergency procedure contains specific diagnos ti.
information so that symptoms of upper head void formation can be recognized. In the unlikely event that these conditons are observed, operators are trained to resort to a feed-and-bleed cooldown method that was success fully used at St. Lucie to safely escort the plant to shutdown cooling conditions.
Analysis also confirms that for NC :ooling purposes more than enough condensate grade water is required by the Technical Specifications. More than twice as much water is available that :ould reasonably be consumed during NC c oold own.
In conclusion, the possibility is remote that upper head voiding could occur during NC cooling at YNPS. Design features of the reactor vessel allow for relatively efficient cooling of the upper head region. Operators are trained to recognize symptoms of the St. Lucie event and are prepared to implement proven methods for sa fe recovery. Procedural restrictions are imposed to ensure that NC cooling can be performed safely.
Sufficient supplies of condensate grade feedwater are available.
VI.
References E-1. WYR 80-96, WYR 80-26, Letter from W. P. Johnson to NRC, Void formation in Vessel Head During St. Lucie Natural Circulation Cooldown Event on 6/11/80, August 21, 1980.
E-2. NSAC 16, Analysis and Evaluation of St. Lucie Unit 1 Natural Circulation Cooldown, prepared by NSAC/INPO, December, 1980.
E-3. OG-57, Memo from R. W. Jurgensen to NRC, St. Lucie Cooldown Event Report, prepared for Westinghouse Owners' Group, April 20, 1981.
E-4. NRC Branch Technical Position ASB 9-2, Residual Decay Energy for Light Water Reactors for Long Term Cooling, Revision 1, 11/24/75.
.