ML20202G227

Forwards Comments on Metal Components Subcommittee 860227-28 Meeting in Washington,Dc

ML20202G227
Person / Time
Issue date:	03/13/1986
From:	Bender M Advisory Committee on Reactor Safeguards, QUERYTECH ASSOCIATES
To:	Shewmon P Advisory Committee on Reactor Safeguards
References
ACRS-CT-1836, NUDOCS 8607150334
Download: ML20202G227 (8)

Text

_ ~

eT-/934o' Sofety

, Follure inwstigotion i Technology & Testing cusaneca assocrues ering and Technical Advisors pgg

,1- Executlw Poric Drfw, Suite 217 Hnoxvt!!e, Tennessee 37923 Phone:(615) 690 2728 _

. - r- - '

C ~ - . . . .. .. ., uz.;;g WF,1.8%8 .

March 13, 1986

..{

-- g Dr. Paul Shewmon, Chairman Metal Components Subcommittee Advisory Committee on Reactor Safeguards 1717 H Street, loth Floor Washington, D. C. 20555

Dear Paul,

Attached are comments on the Metal Components subcommittee Meeting in Washington on 2-27, 28. If there is a need for further amplification of my views please let me know.

Sincerely, M. Bender s

.f*

860715o334 860313 I..y." """ . , , g.g ,.g PDR ACRS -

CT-1836 PDR C utified By (g -_

c .

::v . . .. -_- _- ~. --

4

,. . . J *

"hr.J.h r.: this can be shown depends on' establishing that piping

.:st.srials are sufficiently tough so that a throughwall crack will ha ctabla, even after it has reached a relatively large size, ranulting in a leak in the range where the leak detection '

-y:-tem will unfailingly expose it in time for corrective action.'. '

invostigations of stainless steel piping by Westinghouse, EPRI.

nd others have adequately established this capability for wiunry system piping systems immune from severe dynamic loads

. 4 not subject to stress-related corrosion.

he high pressure carbon and low alloy ferritic steels used in

~Ma primary steam lines and PWR secondary steam lines the

:.i crials of construction are not well enough defined to provide

.n equivalent position. It will be necessary to investigate the

.?racture properties of the piping materials, now in use, as well ta those proposed for future applications.

"ho Beaver Valley Plant is the lead installation attempting to davolop an industry position for the "broadscope" applications.

'.'ronently, the sec.ondary steam lines in PWRs seem to be the 1rincipal candidates for inclusion. Other lines like the PWR I'rassurizer Surge llnes might also be included, if they do not infringe en the regulatory application constraints.

I"r^.VSR VALLEY TEST PROGRAM M:.ver Valley is planning a test program, using piping up to 8" in nystem.

diameter, to evaluate LBB applicability in the secondary G,enerally the test program will make use of its 4.uvantory of archival piping materials. Larger size piping is not to be tested. Its characteristics will be assessed by c.t:titapolation from the smaller sized piping tests.

~

Tha extrapolation will probably be valid if:

n

1. Large and small piping are manufactured by comparable processes so that their properties are similar (not yet established).

2. Loading conditions in tests apply to actual service conditions.

3. Installation and fabrication practices are the same for large and small diameter piping.

Documeritation should be provided to establish the validity of  ;

items 1, 2, and 3 above.

l 2 of 6 I

l

l l

~

Cl3 DITECIION A c:rrelation between leak detection capability and critical (unstable) crack opening dimensions is needed to establish a tanable pcsition fer LBB in piping. The safety position is that, long before the crack r uche.s an unstable condition, it will bagin to leak so significantly that the operating organization will, with a high degree of certainty, recogni::e its condition End put the nuclear plant in a safe condition until the cracking condition is corrected.

E::perience with previously cracked lines in nuclear plants is

, that through-wall cracks have been found before unsafe crack extension occurred because leakage was observed. In a few cases, notably Duane Arnold, cracks along the inside surface of the leaking line have extended almost entirely around the inside circumference. The leak was not very large when it was discovered in those lines. Whether the crack was approaching instability in any of these cases is not established. However, there is sufficient evidence from these events to raise doubt as to whether the leaks would be large enough to be reliably detected.

It would be screwhat risky to base the LBB case on identification of small leaks unless there.is a reliable and sensitive monitoring system. In the recent steam power plant event at Monroe and elsewhere the time between cbserved leakage and instability was very short. The failure conditions may not be comparable to those in nuclear power plants, but the case for reliable response has not been made in the light of this

• experience. Both the leakage rate and the instability determi-nation depend on correct prediction of through-wall crack dimensions. If the crack is long and tight it may leak very little and still be close to instability. The crack depth and aspect ratio (depth to length relationship) are a function of fracture properties ~of the materials. As Mr. Rodabaugh observed during the meeting', we do not have good information about the A106 ferritic piping materials in use during the first two decades of nuclear power plant and perhaps, insufficient knowledge of present-day materials. This makes the problem of predicting the size at which crack instability occurs difficult.

For a conservatively safe position on LBB it might be necessary to establish leak detection requirements on the basis of the smallest possible crack instability dimensions in a configuration where leakage is restricted by a very small critical orifice configuration. This will need careful study--EPRI claims some test results related to leakage predictions but that work needs critical review.

3 of 6

n. . .

~

C 3. : .:Cf.7. N ~: J. . LON: :.0 C;' EIPE BTJ.".T.5 Ic rc.T c m.n M ' -!?c br a.ks are net of dirceu concsrn from the v.t.. j e f.P. t , J cura dantgo.

ro:: cutrple, a double cnded pipe M: ih th "lP. kcin st'u m outlet line vill not by itself lead

.. ?  : V ..l.. c r/ O .it.g e . If sufficient nakeup WEter is available

. , unnun-n.u; -M eteen lect thrc:'gh the broken line, the system

. ill depr;.c.nnti- . in the same Etnner as if the J.DS valves are

,n;ictienf.og (.h1 MS valves may, in f act, work simultaneously)

..':ultia., iu u 8.y a depressurizction condition with the reactor

ubaritJ. ,1. ' chi:: vould make the event relatively innocuous if thera vco enfficient time for corrective action.

a thm lo . nystnuts only the secondary side of steam generater is vul a ..'ull and, unless the breEk leads to external equipment 1.uo u i - o louc of all secondary coolant inventory without ECCS n 0 6b i l.i. h y , th, core would not be threatened. A probabilistic

.~.galysi o? the break consequences may show that for most events nu emca damage threat is negligible. Thus, a " fairly good leak M hicti +." ccpc.bility, combined with low probability consequen-

c. i.e. l. th 1, thy result in an adequate case for LBB when either, h P ta t1 m :, ua'; not.

! 't i E.', C .cff ic considering the introduction of probabilistic I n .1. y c 0 for such purposes. They chould be. encouraged to

v. . !. " . c. Ein Lpproach because it may lessen the need to depend aa hice 3.it.bility 1

in the leak detection systen.,

Xd .

t etcchion capability is needed beyond that which

ou . c . I. y .1xists the available techniques need to be fully e inv b;t.bd . Electrically energized sensing devices placed j u.ad , 1 piping in';ulation cannot be maintained conveniently and

h. L g t. .. f. a t toduce serious operational difficulties.

. Sensors 4 it : tha insulttion would be operationally more acceptable.

riu l c.he:: t.h o H2O Staff nor the industry personnel has a good handl.a on what L.ight be provided.

LM AGC i: CENT METHODOLOGY i

Tht FmF:inghouse (W-H) presentation of LBB Assessment Methodology was i.1.1.nstrative of the approach which any licensee should follow in vaTpnding to the LBB Broad Sc6pe Rule. Because the analysis, ita t l, la high1.y detailed, the numerical results of the W-H work i

wonin uo% ha very meaningful, if used to establish a sharp line o f A a. .;kahlon between allowable and unallowed unrestrained c o n c; i c. . .. a n . Hovt.ver, as a means of showing the effect of relcv:.t ; factors, some analysis of this type should be required for .? vu nuclect installation.

4 of 6 l

l J

y - . , . _ + r.-. ..a .>w - . . - - - - - . , - - - - - ._w.a ,e .s- u _a - - , ,s n . - -

4 l .

Oventually a rance .cf values for the irportant parameters will nozd to be established in crder to ccmbine the conditions of concern logically and in a statistically neaningful way. Some i cansitivity analysis would be helpful for this purpose.

, The EPRI work, which has largely fccussed on stress corrosion t

induced cracking, is not very relevant to theg Eroadsccpe Rule, rince corrosien effects ara excluded as a permissible condition.

Until we understand the corrosion effects better it may be 3 necessary to continue to provide restraints in' critical l locations, as the NRC Staff intends.

?

i. I.lthough it has been sometime since I examined the EPRI Reports

] on leak measurement, I do not recall that their work has provided ,

j . a basis for a high' degree of confidence in leak measurement i techniques for the ~ range of conditions rel.svant to the

, Broad-i scope rule. If the.EPRI correlations are to be used, they should be reviewed by competent flow analysts who are familiar with the i failure conditions to be considered.

j t

The ASME Code requirements for design, manufacture, fabrication

! cnd installation of piping covered by the Broadscope rule may 4 give adequate control of material properties for piping installed l to Section III or the equivalent B31.7 Piping Code but, much of i the existing piping was installed under the older B31.1 rules.

) The latter needs thorough evaluation if it is to be included.

• JET IMPINGEMENT PROTICTION

• r

~

The Jet Impingement issue has not been properly presented. It is '

i cimost impossible to design impingement protection that would resist the direct forces of high pressure steam or feedwater pipe break conditions. The real purpose of the protection is to protect against loss of function where redundancy has not been provided or where redundancy is contingent on screening critical equipment from fluids released by the pipe break. ,

! Impingement baffles'that protect the equipment can be adequately '

l effective if they merely prevent hot fluids from reaching important componentp or from causing electrical short circuits.

i Shelter rather than ' strength resistance is the important require-

ment in most cases.

I

The LBB Broadscope Rulemaking document ought to include a discussion intended to clarify the environmental qualification issues associated with pipe breaks.

L h) il 5 of 6 i

'ii; >

L-

i 2

l j CE O.T.~ T.' '.'IZ3 CIACKING IN SENSITIZED AUSTENITIC STAINLISS STIELS j

Tra .*.e'h t r:a concern for 0::ygen control to prevent cracking of

' . .ati':!. .2d" stainless steals see=s to be minir.i::ed by the use of hydrag:n- injection for EE syster.s. So far, all signs point to i r riniv .: rnsults. The alternative is to maintain very low oxygen co.9t:nt in the EE coolant (<20 ppb of oxygen). Most BE plants i.t.v;.. fcund thic to be difficult.

I C '.rrently, the only drawbacks to the hydrogon application are the

cdditional hazards of storing and handling hydrogen inventory and i the Nit

. ogen 16 radiation caused b j .

Without hydrogen injection, the N 1 ywould carryover be held into up the as an turbine.

oxide

! fer:r, probably nitrate, in the primary coolant liquid until it i

decayed, but with excess hydrogen injection it seems to be ,

l carried out to the steam turbine as noncendensible ammonia before it dect.ys to a stable element. Some installations, but not all,

htve had difficulty with the radiation problem. The reason is 1

anomalous but may be affected by the amount of excess hydrogen provided. If it can be sufficiently low the bulk of the N16 r.ight b: held in the coolant until the radioactivity decay tr: occur hr.s eliminated the radiation problem.

!i

! I? hyd.rogen injection is not used, nuclear grade stainless steels

I (c

:tre low carbon <.035%) may still avoid cracking because they j d> not become sensitized. But this has not been proved by j ,

-::i-i:i ;* experience.

iI

.i

• ii i

I 9

i

)

i l

4 i,

f 4

l i

J j

I j 6 of 6 lf l

I >

i .

j! .

- e

... .....n_,......... . . . ~ .

._ - me . ---V..e- -

<e *-. * * - * - - - * *

~ ' ~ ~- - --

.et.t OC.*:Z.IA.7.T ON PRESSURIZED THETE.AL SHOCK DISCUSSIONS AT THE ACRS METI.L OC:12hERTS SU3001-O'.ITTEE KEETING IN WASHINGTON, DC, ON 2-27,28, 1986

-rei ?.:ad. by M. Sender, Querytech Associatas ,

1.tr h 13, 198G Thc :,cassurized thermal shock (PTS) issue continues to claim the

! t tt.d. ion of the NRC staff. The most recent rapid cooling event at th+. SHUD Rt.ncho Saeco nuclear pcwor station has reawakened interest in Lhe Pressurized Thermal Shock Rule. In the discussions at the 0-20 Subcommittee meeting the NRC Staff provided some updated infor:ation concerning the status of the issue.

RIS" CRITERIA FOR ACCEPTABLE PRESSURIZED THERMAL SHOCK CONTROL Th: :3C Staff is using the work done for them by ORNL as a method-ology basis for judging the public risk form pressurized thermal shoch in PWR Nuclear Power Units. The work on the whole is mostly an event probability sequence analysis using PRA methodology that makes .

use of materials property information and crack detection capability as t. basis for judging crack initiation potential for probabilisti-cally defined event sequences. The NRC Staff intends to use severe cera dama'le probability from such events as a way of judging the PTS i threat. In principle the concept is sound if the numerical core melt probability values are valid and the consequences are properly l established. -

Frca a brief reexamination of the ORNL work I concluded that the work

) was not sufficiently clear to accept the numerical values directly as j a judgement basis. However, if used, as the NRC intends, to assure

that a threshold for PTS caused accidents is sufficiently conser-I vative to provide time for appropriate control meacures it is a

! usable methodology.

1

The controversial points about the nethodology are (a) the time-of-response that can be given credit in the a

alysis (b) the engineering improvements that have been implemented to reduce PTS probability and '

(c) the flaw detection and propagation cha

acteristics related to PTS

, events.

1 l FLAW DETECTION i

"j of these, " flaw detection" seems to be tna least well established information. Dr. Vagins of the NRC Staff indicated that he had succeeded in obtaining some sections of pressure vessel materials from the Limerick 2 vessel and was expecting to perform an NDE examination for comparison with metallica11y inspected sections of l the material to enhance our understanding of detection capability.

I This is highly important to the vessel integrity issue and should be strongly supported by the ACRS. Also fracture toughness testing of representative portions of this vessel might be useful, although it

! should only verify what is already known.

i 1 of 2 l

l

)

- /7

~

For conparisen svaluation p rposes cidcr vessel tatsrial night be i

tvciitble. As a means cf strengthaning our knowledge of actual shcp f abrication quality it would be desirable to erpand this effort to i:ivs. tigate older vessels in a similar way.

VZOf. 2L INTEGRITY INFORIGTIONAL NEEDS A chcrtcoming of the FTS progra::t continues to be its lack of a cystematic r.cthod of displaying information relevant to the PTS 1 iccue. The NRC Staff indicated that it did not have a strong interest in developing an organized set of information for pressure

vossels as a tool for assessing the status of PTS information, and j preferred to depend on the ASME Boiler a'nd Pressure Vessel Code l conformance for this purpose.

1 I do not believe the requirements of Section III of the Code are cufficiently definitive to support this public safety position. It i would be very desirable to establish a definitive list of information needed to establish the case for acceptable PTS risk and to require

] cubmittals of such information from the Licensees in a form that is I amenable to analysis for completeness as well as performance adeque.cy .

!]i -

4 o .

i .

II l

i t

i

! 2 of 2 Go

..~ _ _ - _ - . - - . - - , . . . . _ _ . ,. , - ..-.-_- .- .-_-. - , . - , - . - _ - - . -