ML19247E427
| ML19247E427 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse, Oconee, Crystal River, Rancho Seco, Crane  |
| Issue date: | 04/28/1981 |
| From: | Eisenhut D Office of Nuclear Reactor Regulation |
| To: | Case E, Harold Denton Office of Nuclear Reactor Regulation |
| Shared Package | |
| ML19247E428 | List: |
| References | |
| TASK-2.K.2.13, TASK-TM TAC-45198, TAC-45202, NUDOCS 8105110175 | |
| Download: ML19247E427 (6) | |
Text
.
- '8cg'
'/'
o UNITED STATES
- I [ E[p. e g[,j NUCLEAR REGULATORY COMMISSION
, WO.,,.8/,
t
- e. ASH NGTON D. C. K555
.,a.~
,a e
S fdb J
+,
$,5 2 b Ibb)
- es,*
MEMORAND'M FOR: Harold R. Denton, Director J
Office of Nuclear Reactor Regulation Edson G. Cr.se, Deputy Director Office of Nuticar Reactor Regulation FROM:
Darrell G. Eisenhut, Director Division of Licensing
SUBJECT:
7F.EP3.AL SHOCr. TO PWR REACTOR In response to E. Case's note to me dated April 16, 1981, we have tocrdir.ated an interdivisfor.a1 technical review of the reactor vessel fracture issue to deter:nine if irrnediate licensug actions should be required.
Our preliminary revie > has concluded that alih6 ugh no immediate action is recuired for operating reacters, the staff should continue to evaluate this issue in the near future with the actions identified herein.
b0Y u s.nnu t.,U Divisiond{. Licensing stret or varrein o.
cc:
NRP DIV DIRs d
gjpilIOITS
~~
s s
FRELIMINARY ASSESS?:E!!T CF THERP. L SHOCK TO PWR RE'CTOR PRESS"RE VESSELS
... n -.. r-. e..
.*1aRwaas6.Vf During the past few months the subject of react:r pressure vessel thermal sh0ck has received increased attention by the NRC.Itaff. Most recently, en March 31, 1931, NRC representatives met with the Fressurized Water Reacter (PWR) industry Regulatory Response Groups and the PWR reactor manufacturers. In acdition, concerns have been raised regarding the safety of operating reactors.
In order to determine whether any immediate licensing action is necessary relative to the potential for thermal shocks in pressurized water reactor (PWR) pressure vessels, the staff has evaluated (1) the types of transients or accidents that could lead to everucoling of the reactor system; (2) experience to date with transients that have occurrec in U.S. PWRs; (3) the' pr bability that such over:coling events will occur; and (4) the capacility of reactor vesseis to withstand these transients.
Item a fetused on the likelihced of a flaw existing in a reactor vtssel (RV),
the copper c:ntent c7 RV welds, and the extent of RV irrad'.ation (fluence).
EACKGROUND Severe react:r-system overcooling events which 00uld be followed by re:ressurication of the RV can result from a variety of causes. These include instrumentation and centrol system malfuncticns and postulated ac:idents such as small-break loss ~of-ccolant eccidents (LOCAs), main steamline breaks, or feedwater pipe breaks. Rapid cooling of the RV internal surface causes a temperature distribution atross the RV wall.
This tem:erature distribution results in thermal stress, with a maximum tensile stress at the inside surface of the vessel and a compressive stress at the cutside surface. These stresses rtine with the hoop stress caused by the internal pressure in the vessel. The ragnitude Of the thermal stress depends en the temperature differences across the RV wall.
- s lonc as the fracture resistance of the RV material remains high, such transients will not cause failure. After the fracture toughness of the vessel is reduced by r.eutr:n irradiation, severe thermal transients c:;id cause fairly rr.all flaws near the inner surface to initiate --
and result in -- significant cracking. The vessels cf ecocern are these w'th a hist:ry of high radiation ex;csure, which are made of material that
..as a high sensitivity to radiation danace (such as those r,ade with weids of high copper c:ntent).
S w
1,
- F:r failure t: Ot:ur, a number of centributing factors must te present.
There f auters are:
(i) a rea:ter vessel flaw of sufficient size to pre;ecate, (2) high copper content, (3) a relatively hign level of irradiation, (4) a severe over coling transient with repressurization, and (5) e resulting crack of such size and location that the ability of the RV to maintain : ore cooling is affected.
EVALUATICN The staff review of overecoling events and their probabilities included a review of the Office of Reacter Research (RE5) stucy en Overt: cling events at Babcock & Wilcox (B&W) plants (Ref. I attached); a survey of operating (experience on Westinchcuse (W) anc Combustion Engineering (CE) plants Ref. 2); a review cf availabTe accident analysis in Final Safety Analysis Reports (F5ARs) and in vender topical reports; and a
- relimir.ary probabilistic analysis performed by tne Civision of Safety Technology (OST) (Ref. 3 attached). The preliminary results of these evaluations indicate that there is a pr bability of about 10-3 per reacter year that a ELW-designed plant will experience a severe cver
- 0Lling transient similar er creater in ragnitude :: that experienced at Rancho Seco en March 20, 1978. This transient is the cost severe over:colina transient experienced by any PWR in the U.S.
This probability of 10-3
- per reactor year includes contributions frc steam cenerator control system ralfunctions (the dcminant Centributor); scali-break LOCAs; main steamline or feedwater line breaks; and complete less of feedwater fl0w.
The staff estimates that-the probability of such an over coling event in CE er W-designed reacters is lower, pernaps by an order macnitude, than f:r ELT-Eesigned reactors. This difference is based on design differences and cn Operating experience.
In the 1973 Rancho Seco transient, reactor pressure was raintained at a The fairly high level (1500 psic to 2100 psig) throucheut the c:cidewn.
minimum tem:erature of the reactor teclant (230'F) during tre transient was high enouch to r.aintain the material touchness of the reactor vessel.
Moreover, this eval ~uation leads the staff to believe that if this transient were to be repeated at Rancho Seco cr any other B&W-designed facility within the next few years, the RV failure weuld still be unlikely.
Nonetneless, the possibility cf vessel failure as a result of an ever-c cling event cannot be cc:aletely ruled out.
If an avert oling event sucn as that at Rancho Seco were to occur, based on the nany factors
- ertinent to an analysis of vessel failure,the staff wcuid expect much less than ene failure in the current population f reactor vessels. Even fer the vessel with the worst material prc;erties, the staff would not ex;ect a failure.
e
3-The staff conclusien is supported by the analyses of the Rancnc Seco event performed by the Oak Ridge National Laboratory (CENL) (Ref. A attached). The CRNL analyses indicate that the threshold irradiaticn levei for crack initiation (that is, small cracks growing t: larger ones 40*F assuming ccr.servative initial material properties such as RTNDT =9 1
to and' cop;er centent of 0.35%) would be in the range of 0.75 x 10 1.7 x 10l9 neutran/cm2 The highest fluence to date in a B&W-designed facility is less than half the minimum value listed above. It would, therefere, be several years before any B&W-cesigned facility reached its threshold irradiation level.
Some reactor vessels in CE and W facilities have somewhat higher fluences; however, other mitigating factors -- such as icwer values of initial NDT -- provide a significant margin to failure should an overcocling RT event similar to that at Rancho Seco cccur.
en.-, m.0..c nr). c..r C n,~u.v. h: an. -we, c
.1 ~ ~
a a
sa-As a result of its evaluations to date, the s_:aff has concluded that the probability ;f a severe overcooling transient (similar in magnitude to tne Rancho Seco event) is relatively low. Fct S&W-designed reactors chis probability is estimated to be about 10-3 per reactor per year, and for W-and CE-designed reactors, it is icwer, pernaps by an order In additien,, the staff has concluded that, based on of magnitude.
prese.nt irradiaticn levels at aperating reacters, RV failure frem such AcccrdTngly, the staf f believes that no inmediate an event is unlikely.
licensing acticns are required en 0;erating reactors; however, the staff
-eccmments that the felicwing actions be taken:
Request industry representatives meet with the NRC staff in the 1.
near future to discuss:
industry ; regress since the March 31, 1951 meeting a.
b.
bases for centinued safe cperation the letter of April 10, 1951 fr 'm D.L. Basdekas to c.
Chairman Cdall (Ref. 5 attached).
The Divisicn of Licensing has recuested such a meeting with the
?WR venders and Owners Grcups. The nesting is scheduled for Acril 29, 1981.
TSe staff shculd continue to refine its understanding of this safety 2.
This c:ntinuing assessmant, taken together with inferration
- cncern.
being prcvided by industry Dsners Grcu:s (including the Owners Group Action Plan due May 15,1951) should permit the staff to define wha e
9
actions the industry and the NRC.aust take to resolve this safety c:ncern. The stafi's eff:rts during this short ter: should include, but n:t necessarily be limited to, tne fc11: wing areas:
a.
Oevelopment of a better understanding of overcooling transients anc accidents. Factors to be examined or addressed in this continuing evaluation would include:
(1) human facters considerations (2) refinements in the analysis of the prchability of such events occurring, including considerations of overcooling events more severe than at Rancho Seco (3) an understarding of improvements in instrumentation and control systems implemented since the event at Rancho Seco and other evercooling events and the effects of these improvements en the prcbability cf overcociing events.
b.
Development of a b ".ter understanding of the potential for and effects of RV :nermal shock including:
(1) a cate;crization of the susceptibility of operating RVs to cracking as a result of rapid cooling, censidering the combination of irr~adiation lereit, vessel impurity content, and existing flaw sizes.
/
(2) a sensitivity study of the effects of fluid mixing and the development of realistic sodels and assumptions.
c.
An assessment Of further recuirements and of the overall contributicn to safety of potential improvements.
d.
An overall integrated assessmen' and report of cenclusions and recommendations developed in connection with the above items.
e 4
m O
i
..........e r.t r : -.. c.
1.
" Insights on Overcooling Transients in Plants witn the B&W NESS,"
M. Taylor to 5. Fabic, dated October 29, 1980.
2.
Nuclear Power Ex;erience 1980, Bernard J. VerraI Publisher; Nuclear F wer Experience, Inc., Encino, CA.
3.
Trecuency cf Excessive Cooldown Events Challenging Vessel Integrity.
A. Thadani to G. Lainas, dated April 21, 1981.
4.
Parametric Analysis of Rancho Seco Overcooli.',g Accidents, ORNL letter, R.D. Cheverton to M. Saginis (NRC, RFS), 3/3/81.
5.
Letter from D. Basdekas to The Honorable "erris K. Udall, dated April 10, 1981.
m O
e M
O