ML20082S069

Provides Response to RAI Per Proposed TS Amend for Nominal PORV Pressure Relief Setpoint Vs RCS Temp for Cold Overpressure Protection Sys

ML20082S069
Person / Time
Site:	Braidwood
Issue date:	04/27/1995
From:	Saccomando D COMMONWEALTH EDISON CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM), Office of Nuclear Reactor Regulation
References
NUDOCS 9505020347
Download: ML20082S069 (14)

Text

-.

Commonwealth litison Company

/

o 1400 Opus Place Dow ners Gros e, IL 60515 April 27, 1995 Office of Nuclear Regulation U.S. Nuclear Regulatory Commission Washington, D.C.

20555 Attn:

Document Control Desk

Subject:

Response to Request for Additional Information (RAI) Pertaining to the Proposed Technical Specification Amendment for the Nominal PORV Pressure Relief Setpoint Versus RCS Temperature for the Cold Overpressure Protection System

References:

1.

R. Assa letter to D. Farrar dated April 20, 1995, transmitting Request for Additional Information 2.

D.

Saccomando letter to the Nuclear Regulatory Commission dated December 16, 1995, transmitting a Proposed Technical Specification Amendment for the Nominal POP'r Pressure Relief Setpoint Versus RCS Temperature for the Cold Overpressure Protection System Reference letter 1 transmitted the Nuclear Regulatory Commission's Request for Additional Information regarding the proposed Technical Specification amendment for the Nominal PORV Pressure Relief Setpoint Versus RCS Temperature for the Cold Overpressure Protection System which was transmitted in Reference

2. Question number 3 mentioned that in Braidwood's submittal, the instrument uncertainties have not been incorporated into the power operated relief valve setpoints for low-temperature overpressure protection curve, and that the Staff found this to be unacceptable.

We are enclosing the attached document which i

summarizes the technical bases for the exclusion of a random instrument uncertainty margin term from the pressure-temperature l

limits and the cold overpressure mitigation system setpoint.

This attachment may be useful during your review to help resolve this issue.

i i

K:nla \\tudu topral J

9505020347 950427 ADOCK0500gga, PDR P

1 g

A l'mCom CompJHy i

NRC Document Control Desk April 27, 1995 If you have any questions regarding this, please contact this office.

Sincerely, Denise M. Sac

.ando Nuclear Licensing Administrator Attachment cc:

R. Assa, Braidwood Project Manager-NRR S. Dupont, Senior Resident Inspector-Braidwood J. Martin, Regional Administrator-RIII Office of Nuclear Safety-IDNS Kinla\\bwd\\1toprai

._J

J Date:

November 14,1994 e

i To:

K.L Kofron R. Kerr i

G.K. Schwartz D.E. St. Clair j

j E.A. Broccolo M.E. Lohmann 7

t

. Technical Position on Random Instrument Uncertainty for P T Limits and COMS Setpoints.

j j-

Subject:

J j

Reference:

J.N. Chirigos and T.A. Meyer, " Influence of Material Property Variations on the Assessment of Structural integrity of Nuclear Components," Joumal of Testing and Evaluation, Vol. 6,

.j i

i No. 5, September 1978, pp. 289-295 (attached).

j

}

}.

. At the request of M.A. Gorski, Braidwood SEC, the technical bases for the exclusion of a random instrument i

uncertainty margin term from pressure-temperature (P-T) limits and cold overpressure mitigation system (COMS, also known as low temperature overpressure protection system, LTOP) setpoints have been j

summarized. Input to this letter was also provided by Russ Tamminga (Consultant), Jim Chynoweth 4

j (SMAD), and Steve Goslin (EPRI-NDE).

j introduction

~

1 P T limits and COMS setpoints are established to protect the reactor pressure vessel (RPV) from nonductile failure due to overpressurization. The determination of the P-T limits and COMS setpoints is based on very conservative assumptions and methodologies. Incorporation of a margin term to account for random j

instrument uncertainty is not required by the applicable regulations and is not necessary in light of the large j

margins already present in the determination of the P-T limus and COMS setpoints.

f incorporation of a margin term to account for random instrument' uncertainty also:

l reduces operating flexibility at low temperatures (operating flexibility as a function of the j

difference at a given temperature between the maximum allowable pressure in accordance with ASME Section XI Appendx G, and the minimum pressure necessary for proper 1

operation of reactor coolant pur p seats), and increases the likelihood of COMS actuation, endangering reactor coolant pump seals at lower temperatures and unnecessarily challenging the reactor coolant system, with no concurrent increase in protection against nonductile failure of the RPV.

l i

Anoendix G Conservatinms h

Random pressure and temperature instrumentation uncertainties are insignificant when compared to the I

margin terms already include 1in the ASME Section XI Appendix G methodology for determining P-T limits, j

which are subsequently used as an input to COMS setpoint3. These margin terms include:

)

i 1)

A safety factor of 2 is applied to tha membrane stress intensity factor (pressure). For example, a P-T limit for an allowable pressuie of 400 PSIG would actually be based on the i.

stress intensity resulting from a pressure of 800 PSIG.

i 2)

ASME Section XI Appendix G allows the sum of the pressure stress intensity factor l

(multiplied by a safety factor of two) and the thermal stress intensity factor to be no higher than the reference stress intensity factor K shown in Appendix G Figure G-22101 (attached). The K value at a given temperature is the lower bound of all available static, i

dynamic, and crack arrest fracture toughness data, and is identical to the lower bound crack arrest K. values shown in Section XI Appendix A Figure A-42001 (attached). This is considerably more conservative than either the crackinitiation toughness, K,of Appendix l-A Figure A 4200-1 (which Section XI Appendix E permits for evaluating the effects of actual i l

overpressure events on the operability of an RPV), or the actual fracture toughness of the RPV limiting mater'al, which would be expected to fall above the K. curve.

1 1

1

~. -

" 3)

'A 20 margin on mean predicted shift is included in the Regulatory Guide 1.99 Rev. 2 method for d termination of a conservative, upper bound value of adjusted r ferenc3 ternperature. For Braidwood Unit 1 at 32 EFPY and a depth of 1/4T, this results in the addition of 56*F to the adlusted reference temperature used in the calculation of P-T limits; for Braidwood Unit 2 at 16 EFPY and a depth of 1/4T. this results in the addition of 52.5'F to the adjusted reference temperature.

4)

Stress intensity factors are calculated on the basis of an assumed flaw in the wa;l of the RPV with a depth equal to 1/4 T. The degree of conservatism associated with this

- requirement can be seen in Appendix E, which permits the use of a 1.0" initiation crack size. This is considerably smaller than 1/4T, and is consistent with industry experience in pertorming ISI of RPV beltline regions.

Illustration of Mara_in - Chirinne and Maver Refetg[g3 Figure 5 of the Reference (attached) illustrates the conservatism typical of P-T limits developed in accordance with Appendx G for a typical PWR at an end-of-life fluence.

At the lower end of the operating temperature range, where the concern for protection against nonductile fracture is highest, Figure 5 shows that a margin of at least 600 PSIG exists above the Appendix G P-T limit if the sum of the actual pressure and thermal stress intensity factors is not allowed to exceed the X, curve.

Even greater margins exist when more realistic flaw sizes and the estimated available beltline material

~{oughness is taken into account.

\\

litustration of Marain - Section XI Annardw E The Appendix E Paragraph E-1200 Acceptance Criteria can be used to !Natrate the conservatism of Appendix G specifically for Braidwood Units 1 and 2.

For Braidwood Unit 1, with a 32 EFPY 1/4T adjusted reference temperature of 159'F (56*F of which is Reg.

Guide 1.99 Rev. 2 margin), the maximum allowable pressure for a pressurized thermal transient occurring at 214*F is 2485 PSIG (the design pressure). The Appendix G steady-state pressure limit at the same temperature is 935 PSIG, a margin of 1550 PSIG.

For Braidwood Unit 2, with a 16 EFPY 1/4T adjusted reference temperature of 145'F (52.5'F of which is Reg. Guide 1.99 Rev. 2 margin), the maximum allowable pressure for a pressurized thermal transient occurring at 200*F is 2485 PSIG (the design pressure). The Appendix G steady-state pressure limit at the same temperature is 912 PSIG, a magin of 1573 PSIG.

The random instrument uncertainty of 60 PSIG and 10*F applied in the past to P T limits 5 therefore very small when compared with the margin between the Appendix G P T limits and the Appenuix E allowables.

For isothermal pressure transients, the margin available from Appendx Eis even higher. And accounting for the shift in the curves resulting from the adjusted reference temperature margin would also result in evoi, higher margins.

Illustration of Marain - Rectinn XI Annantlir G The ASME Code explicitly recognized the amount of mar 0in inherent in Section XI Appendix G P 't' limits in the 1993 Addenda. With that Addenda, Appendix G paragraph G-2215 incorporated a provision for, allowing COMS setpoints to exceed the usual P T limits by 10% as a standard practice.'

1

J d

Conclusion The regulations, Codes, and regulatory positions governing P T limits and COMS setpoints (10CFR50 Appendix G, ASME Section XI Appendix G, and NUREG 0800 Section 5.3.2 Branch Technical Position MTEB 5-2) do not require margins for random instrument uncertainties.

j Compared with the margins inherent in the P-T limits a. ' resulting COMS setpoints developed in accordance with thesa regulations, random instrument uncertainties are insignificant and additional margin need not be included.Section XI recently recognized this by incorporating a provision for allowing COMS setpoints to exceed Appendix G P-T limits by 10%.

The imposition of margin in addition to that already shown to exist reduces operating flexibility and 4

increases the likelihood of COMS actuation, endangering reactor coolant pump seals and unnecessarily challenging the reactor coolant system, with no concurrent increase in protection against nonductile failure of the RPV.

4 if there are any questions, please call Tom Spry at 708/663-7268.

%~, B L T.D. S,ory f

/

)

S/G & RPV Projects i

i cc:

J.C. Blomgren N.J. Mares J. Hosmer K. Norris J.M. Chynoweth R.J. Tamminga D. Saccomondo H. Pontius M.A. Gorski R.E. Waninski 4

4 s

4 i

i 4

l

i

.i l

1 i

G 2212 APPENDIX G - NONMANDATORY G-2214.3 I

160 7

J

/

140 - Kh - 26.78 -1.223 emp [0.0145 (T-ATuor + 1M), h Kg ** rehrence streas intensity factor, hale

,9g

,Tl a temperature et which Kg is permit 1ed, OF yp7 = rehmnce nil ductility temperature

[

j AT 100

2. 80

/

.I k

r

)

4 60 40 20 0

-240

- 200 -160 -120

-80

-40 0

40 80 120 160 200 240 Temperetur, Relative to ATNDT IIT-8TNo d f.'F FIG. G-2210-1 8

s

flg. A 42061 1992 SECTION XI-DIVISION I i

l i

l i

220 i

200 180 160 Kg s

140

=

120

)

Kg, I 500 C

i

,s

.0 E

i

.0 it 40 RINOT 20 i

i i

i i

i l

i I

i l

i l

i

- 80

- 40 0

40 80 120 160 200 (T-RTuori,4 A92 FIG. A 4200-1 LOWER BOUND K. AND K. TEST DATA FOR SA-533 CRADE B CLASS 1, SA-508 CLASS 2, AND SA-508 CLASS 3 STEELS t

392

Authrrized Reprint froem 1

Joumal af Testing and Evaluation. Vol. 6, Ns. 5 Copyright American Society for Testing and Materials 1916 flace Street, Philadelphia, Pa.13103 1978 N. Chirigos' and T. A. Meyer' influence of Material Property Variations on the Assessment of Structural integrity of Nuclear Components REFEllENCEs Chirigus. J. N. and Meyer. T.

A., " Influence of sidered normal and are expected to occur and thnw that are more Maledal Property Vartations on she Assessment of Structural Integdty severe but are not expected to occur (accident conditions).

of Nucient Componenta," Journal of Testing aird Enoluation.

JTEVA,Yo!.6 No.5 Sept.1978,pp.289-295.

App!! cable Standa.rils and Criteria ABSTRACT: The question, "Where can the generation of new material property data reduce overly conservati e design in reactor sys.

He various applicable standards and criteria are these:

re[by id ntif ng here conse atis exb r ati (1) American Society of Mechanical Engineers (ASME) Boiler an o ete anining m%erial property variations. Several areas are reviewed to and Pressure Vessel Code,Section III, Appendix G, " Protection identify cowrvatism and the sensi;ivity of structural integrity Against Non Ductile Failure,"1974; evaluation rest,Its to the property vanations. The areas reviewed arte fluence dependence of the transition ternperature, fracture toughness, (2) Code of Federal Regulations Title 10, Part 50, Appendix G; the drop in upper shelf toughnest, the arrest toughness, the material (3) Code of Federal ReIulations, Title 10. Part 50. APE'ndix H'-

l thermal properties, and the fatigae crack gronh rate.

(4) ASME Boiler and Pressure Vessel Code Section XI,1974; (5) ASME Boiler and Pressure Vessel Code,Section XI, 1 Y WORDS: necteer reactors, mechanical pmpenies, fatigue Appendir A "Evaluationofflawindications,"1974;

, (materials), fracture mechanics, fracture toughness, structurat

, integrity, transition ternperature, arrest toughness, fatigue cvack (6) U.S. Nuclear Regulatory Commission (NRC) Ftandard Plan for Pressure-Temperature Operating Limits; and f1ou h c s (7) NRC RegulatoryGuide 1.99, Revision 1," Effects of Residual as r s IDements on Predicted Radiation Damage to Reactor Vessel Materials,"1977.

A question presently existing in the nuclear industry is, "Where cat the generation of new material property data reduce overly De ASME Ill, Appendix G, standard presents criteria and conservative design in reactor systems?" He question can be methods for prevention of nonductile failure in nudcar compo.

ansutred by identifying where conservatism crists relative to deter-nents. All loading conditions expected to occur, including normal mining material property variations in the evaluation of nuclear peration, upset, and test conditions, must be considered. I.oading components' structural integrity and by demonstrating the sensi.

c nditions not expected to occur are classed as accident condi.

tivity of the structural integrity analysis results to the property ti ns" and are not considered in the ASME lil, Appendix G.

varialons. De impact of material property variations on the analysis. Appendix G of 10 CFR 50 presents fracture toughness structural integrity (,f nuclear components can be i!!astrated by requirements and makes the, ASME l!!, Appendix G analysis evaluating the structural integrity of a pressurized water reactor mandatory. Again the conditions dassed as " accident conditions" vessel. Le limiting region of a reactor vessel that should be are excluded. Appendix H of 10 CFR 50 requires that the reactor analyzed in the esaluation of streural integrity is the beltline or vessel have a surveillance program if the end oflife Quence is core mid. plane region. His region is usually limiting because of CIPected to be greater than '10" n/cm and gives the requirements 2

the high Quence levels and resulting irradiation damage that for such a prnCram.

occurs in this region.

He ASME XI standard gives code requirements for in service To assess the reactor vessel integrity it is necessary first to In5Pection, and Appendix A of Section XI requires that defects provide a basis for the assessment by considering the applicable revealed during inspection be evaluated for necident loading r Wattis and criteria used to design, evaluate, and operate a mnditions as well as for fatigue. The NRC standard review plan melcar plant. Then, with these standards and criteria as guides, presents guidelines for the NRC reviewer of pressure temperature the evaluation of the sensitivity of the reactor vessel uructural operating limits and invokes 10 CFR,50, Appendix G, on all integrity to various material property assumptions can be made, plants, old as well as new, and also requires demonstration of His sensitivity is assessed in this report for two sets of plant the adequacy of the p'snt under accident (thermal shock) cundi-conditions: those nuclear plant teactor conditions that are con-ti ns. De vari us regulatory guides give suggestions as to how to account for different effects. For example, R.G. l.99 gives

'Mana,r:ct and eng:ineer, respettively. Westing:hr use flectric Corpora.

guidance on how to acenunt for the embritcling effects of neutron tion,1.O. Llos 355, Penn Center lhulding 2. Piustiurgh, Pa.15230.

irradiati.m.

(> 1978 by the American Society for Testing and Idaterlats 0000 3973/78/0000 02S9500.40 209

=

i JOURNALCF TESTING AND EVALUATION l

, 296 k> to the calculated stress intensity factors mrmalCondition Plant Opcistion l

he standards and cr.teria applicabic during normal plant value delined by the Km curve, that is, the reference fracture i

cperat,(on are Appendixes G of both ASME.fli and 10 CFR 50.

toughness.

The specific Appendix G cvaluation of heat.up and cool.down

%c prmespal anatcrial property of interest is the material fracture curvein ASME.lli, in nuclear power plants will be discussed separately in a later l,

toughness, which is given in the code as the A in Appendix G, and as equations for mutation toughness K. andg;,

i By considering the various design loading conditions such as of arrest toughness Ka in Appendix A of ASME.XI. 'lhe Ku steady state operation and step load change in power, the stress i

ASME.XI corresponds to Kia in ASME.llt. Appendiz G.

Figure I gives a plot of Ki. and Kn (or K,n) as obtamed from i intensity factor is calculated for each the ASME.lli, Appendix G. Temperature is plotted relative to the i, appropriate safety factors are applied, and th reference transition temperature, RTsor. De code gives n intensity factor is compared with the K ncurve at the ternperature i

i guidance as to the value of toughness on the upper shelf, which is where the particular loading condition occurs. An example of assumed here to be 220 MPa m 's (200 ksi in.). It is felt that the results of this type of analysis is shown in Fig. 3 for an analysis j

i this value is supported by existmg datat however, the actual valueof the beltline region of a four loop reactor vessel at end of. life a

taken is not too important for the task at hand, which willinclude.. g),,,,.cc conditions.

he points plotted in Fig. 3 represent the stress intensity factors

\\

]

an assessment of varying shelf toughness.

calculated for individual loading conditions. De position of the reference toughness curve (K n curve) has been established by two i

Ruence Dependence different methods: Method A, % power ARTaur. fluence de.

An important aspect of identifying the toughness curves to be pendence and Method B, % power ARTsur fluence dependence.

curve, for Even though there is a significant difference in the K a applied to a given situation is to account for the effect of irradi.

i ation. Bis can be accomplished by employmg socalled trend these loading conditions, in this particular example ample margin l

cunts that give the change tr. reference temperature with fluence still remains with the curved portion of the K ai curves.

j at given copper and phosphorus levels. Regulatory Guide 1.99 e

j presents such curves as does Ref1.

OUNDT I*II As can be seen by examining these doeurnents, there is a dif.

l I

ference in the fluence dependence assumed on property changes.

Ris difTerence is illustrated in Fig. 2. This figure presents AR Taar j

i fF) as a function of increasing fluence. Two fluence trends are lj shown in Fig. 2: one curve is ARTar P otted as a function of the l

m

1. b m al

..sqdare root (% power) of the fluence and the second curve is 1

I ARTwr plotted as a function of the cube root (% power) of the

,,3 l

fluence. The impact of different a RTuor fluence curves on reactor

" **

• 3 vessel structuralintegrity can now be assessed.

4 j

l i

ASME lli, Appendit G Assessment l

%e Appendix G assessment involves determining the stress Fluence (NICM2 ), y,y) fi intensity factor owing to the design loading conditions acting on an assurned reference flaw usually taken as one fourth of the wall i

thickness (%T)in depth. Appropriate safety factors are applied F10. 2-change in reference armperature whh /fuence.

Frecture Toughness (ksi 6) k'I N-KIR 300 230 4

200 j

250 i0 8

160

1. . 6 200 a,ee ss.tes e a sosesel0 os 140

,.ng,,,

{t.o s imes**d al

- [' *' mot * '"N 120

~150 en.* t o K ses i

100 nt,,,

{

W'H** 81 30 i

100 30 i

40 50

- Ku.see.i.ss:

f.-

..fo onas[s.avmo,. iso],

20 f

!l

!I I

I I

I i

e i

I t

i I

f 0

150 500 50 0

50 100 150 200 250 200230 285 350 400 450 500 550 600 Temperature (*F)

Tempesetute RTH0T (*F1 FIG. 3-Temperature corn versus Kon for four loop reactur vessel FIG 1-Reference frarture soughness curws if Ast he."

es the in kthre l1 Asi in. " = l.1 MPa m *; *C = (*F - 22)fl.6,\\.

= 1.1 Wa m *; 'C = (Y - )))t1.gl.

"i i

4 we w

CHIRIGOS AND MEYER ON NUCLEAR COMPONENTS 291 There is, however, a less satisfactory situation relative to the reactor design that operates at low average coolant, tempera.

proximity of the individual K ealues to the assumed toughness at ture T i

g.

the shcif. Here is ample margin relative to the assumed value of The criticality limit is determined by adding a 22*C (40'F)

{

i 220 MPa m"2 (200 ksi.in.' 3) and less, of course, at the value of margin to the maxirnum pressure limit to prevent brittie fracture l

182 to 187 MPa mP2 (165 to 170 ksi in.8'2). which represents based on ASME lll. Appendix G. Note that in ti,is case reactoe i

e the termination of the Km cune in Appendix G of ASME lit.

operation is marginal based on Method A (% power dependenec)

It should be emphasized that there is considerable conservatism and impossibic based on Method B (% power dependence). Rus.

in the calculated stress intensity factors, arising mainly from the it can be seen that the Quence dependence of.1RT,wr is a sig-factor of 2 on pressure stress intensity and the assumption of a nincant parameter that must be established, it is essential that

%T Daw, as required by ASME ill. Appendix G.

the proper Quence dependence % power versus % power iluence The upper shelf values portrayed in l'ig. 3 are considered ade-dependence, be established since it can determine w hether reactor quate, in panicular for plants built to controlled copper limits.

operation is possible at end-of life conditions, with the assumption, i

Results from on going tests will help estabhsh upper shelflevels for of course, that the margins presently required will be applicabic in older plant.s with relatively high copper content.

the future.

It appears that there is a more critical need to establish the effect

1. cat Up and Cool Dower of fluence n the shift in transition temperature rather than the effect of nuence on upper shelf toughness.

Typical heat up and cool-down limit curves are constructed An indication of the conservatism in the ASME.!!!, Appendix G, by the methods of ASME Ill, Appendix G. for specified heat up limits can be seen in Fig. 5. This Ogure presents the maximum and cool down rates, that is 33 *C/h (60 *F/h) cool down rate.

pressure limit at end oflife conditions for a three loop vessel A time period of reactor operation is chosen for which heat up having Ngh (0.30%) copper content.

and cool down limit curves are to be determined and the RTunr is %e lowest curve is that obtained by using the ASME Ill, j

then identined. De stress intensity factor K, for thermal stress is Appendix G, methods. The subsequent curves result by:

calculated at a specified coolant temperature relative to the RTwor. De K a is determmed from ASME Ill. Appendix G and (1) relaxing the factorof 2onK,(pressure),

i (2) using theK,instead of theKsacurve, then the a!!owable reactor system pressure is calculated which i

satisfics the relation:

' Indicated System Pressure (PSIG)

K a > 2K, pressure + K, thermal i

4800 I

I This procedure is repeated for temperatures covering the range No Themal Stim

) j j

of reactor coolant temperatures of interest and the results ar I8 presented as curves of maximum system pressure versus system 4400 g g i

temperature, as shown in Fig. 4. The curves in Fig. 4 are repre-

~

I sentative curves for a three loop, high copper vessel having an 4000 hkiM ' ktr end of life Duence, A

~

f Too end-of life limit curves are shown in Fig. 4, one determined 3600 l l by the % power fluence dependence, the other by the % power I l

~

fluence dependence of shift in RTaor. The case chosen was for a 3200

- Acm.s k c l g i

0.5 inch rtaw l

I K,c um.

c 2800 -xtc Cam.

I s/4 T ri.w Pressure (psi)

- o.51nch riaw (Kgg+x )

3a00 g

gy 2400 - (kiu + ktil o.i e...

a I

Appendix G, l

2800 u.n at trPY No Safety ra-tor

~

a u.m.* e-2000 kic Cum.

g txp gg

\\ lII

~ *'" "**

2400 l /< 1 liHf 1600

'kia

• ks78

^oe' add o f

2000 2 k gg + Kgyl i

f 1200

/

/

1600 cemceur Lunds-[

,/

Temperature Scale

/

Method A 800

/

1200 j

7, 400 Method a e00 too aco soo soo soo soo oo soo 400 oI I

I I

I I

I I

i i

I I

i i

i o

0 200 400 600 800 200.t00 0

100 200 300 400 500 600 700 Temperature (*F)

Indicated System Temperature ('F) rlG. 4-Operating limits aceveding to ASAIElil. Appendit G F1G. 5-Reactne wssel cool-Jawn limir curws. AT/st. masimum ll psi = 6.9kPa: *C = (*F - 32)/l.5; UyY = rjproavfullpumryrurs.

3)*C/h (60*F/h}llpsig = 0.11 Ara:1in. = 2Mmm: *C = (*F - 321/l.xl.

l JOURNAL.OF TESTING AND EVALUATION

,.'297 l

(3) relaxing the assumed critical flaw size from %T to 25.4 because they can produce high thermal stresses in the reactor i

and 12.7 mm(I and % in.), and, finally, vessel. Td evaluate the reactor vessel under thermal shock condi-(4) basing the curve on an assumed " actual toughness," which tions l11 a desesiption of the accident transient, usually pit-can be expected to fall above the code Kw curve, a value which sented in terms of the coolant pressure and temperature as a i

may result from the testing of specimens contained in reactor ves-function of time into the transient, is first developed. From this set surveillance capsules.

Information the temperature distribution, or the temperature pro-file, through the reactor vessel wall at fixed times into Ihe transient j

Figure 5 illustrates explicitly the conservat. ism that exists m. the is calculated. At a given time into the transient, the temperature i

}

current methods for defining maximum pressure limits. The lower profile is used to deterrnine the thermal stress distribution through q

temperature scale resuhs when the teughness curve is shifted by h wall. The thermal stress and the pressure stress (if any) are

]

the % power fluence dependence as opposed to the upper scale, then used to determine the stress intensity factors for a postulated trhich is based on the % power Huence dependence.

flaw of varying depth through the wall. He calculation is usually Figure 5 demonstrates the need to establish the corrret method performed at a given time into reactor life, typically at end of life.

of extrapolating the effect of fluence on.iRTer, and it also illus-which pru ides the maximum fluence at the inside diameter of trates the need to assure that shiftmg the reference toughness curve the vessel. The distribution of Iluence through the vessel wall, in 4

by use of Charpy data is mdced properly conservative. This can be conjunction with the temperature profile at a C ven time into the i

accomplished by conductmg irradiation tests that include both transient, then permits the determination of the fracture touC -

j h

Charpy and fracture mechanics specimens that yield a direct ness, Kw and Ku, as a function of position in the vesse! wall at a measure of fracture toughness and comparing the shifts m tough-g, gg g

g.

g ness curves. Support for the contention that the method of ac-transient.

counting for the effects of irradiation on toughness usmg Charpy The results of such a calculation can be represented in a plot of dita to shift the K ai curve is conservative can be seen in Rg. 6.

stress intensity factor as a function of a/r, where e is the erack l

In Fig. 6 the dynamic fracture toughness obtamed from 12.7 mm depth or position in the wall of thickness t. Such a plot is shown 1

(%.m.) thick wedge openmg-loaded samples irradiated m tw b Mg. 7.

different surveillance capsules are plotted. Dese data, although In Fig. 7 the variation of K, Kw, and K, through the vessel i

i limited, mdicate that the toughness is significantly above the. wall is presented for 600 s into the transient. Where K a Ku frac-i Ka curve, which was positioned by us, g a % power fluence ture initiation will occur, and if initiation occurs, arrest will occur m

i dependence on Charpy data. He dashed curve was drawn mth whenK < K t

the shape of the K, curve and positioned by the point at T -

It can be sin in Fig. 7 that the K curve has just n!issed the

)

i i

RTwor,= 3o*C (100"F). The dashed curve is that used to define lower part of the Kw curve and that initiation will not occur until "retual toughness in Rg. 5.His plot also suggests that the shelf some crack depth greater than 0.6 e/t. Obviously, a slight change j

2 toughness for irradiated (-7 x 10'8 n/cm ) material is greater in the position of the Ku (orK,) curve could have lead to initiation J

than 193 MPa m (175 ksi.,in. ).

at flaw sizes less than 0.1 a/s.

If similar plots for longer times into the transient are con-a *===iont of VesseIIntegrity Under Accident Conditions sidered, as presented in Rg. 8, it can be seen that the initiation

)

ry Sm8U r my larSe cPendng on W wu Kw h

]

5 C

Accident conditions are those severe conditions which, although not espected to occur, are considered in the reactor vessel design 600 Seconds Stress Intensity (KSI/UI)

Dynamic Fracture Toughness (ksi/d 300 250 8

225 200 l

/~N 8

200 ns

,/

f

\\

~

e 1

f 32s f

$00

/

100

/

IA 75

/

e cweae a 60 A ew.w.e 25 I

I I

I I I I I I ! 8 ! '

O O

300 100 0

100 200 300 0.0 0.2 0.4 0.6 0.8 1.0 (T.RTNOT), 'F Relative Temperatu'*

Fractional Distanco (a/t)

FIG 6-surmsure niese lI asi in." = 1.1 blPe.m*: *C = (*F -

F1G. 7 Thermal shock evaluation far 600 s into the svensient (1 32VI.Rl.

ksi.in." = 1.1 bil'a mh).

s

~ ~~-~~^ ~~'

~C CHIRIGOS AND MEYER ON NUCLEAR COMPONENTS I

l

' inown. Actually, in this series of curves, the K value never Effect o/Thermo/ Properties i

reaches the K u value unti1 the upper shcif, chere the critical

. g g

g gg l*

stze (mmimum crack imustion depth) is very large in terms of the wall thickness. As illustrated here the resu';ts are very sensitive to g g

g g;

the locanon of the toughness cune. Obsiously, the snethod of ness, and upper shcif toughness. We have not considered un.

l shifting the uninadiated toughtiess curves can have a very signifi.

cant impact on the calculated crmeal flaw sin.

.M i M iso idi rimhmi The relat,ionship of the actual vessel toughness to the assumed design Ku curve can also be very significant m determming the ggg g

actual critical flaw sue determined for the vessel. A small change l

in toughncu can change the mtcrsection pomt from a very small t in Fig.10 a temperature profile that results from an assumed a very large value of a /t (or vice versa),

reduction of 25% in thermal conductivity, because of irradiation, is compared with the base case, that is, no reduction in thermal conductivity.

l Efect ofIhe UpperShelf Toughness Value it is known that neutron irradiation will reduce electric,a! con.

De sensitivity of the calculated critical flaw size when Ki > Ku ductivity in metals and it is reasonable to assume a reductmn m in t! c upper thelf region is shown in Fig. 9. Here, as before, the Nnnal con ty. To the kst d the authon knowkdge the j

standard shelf value has been taken as 220 MPa mb2 (200 inaal properties of irradiated reactor vessel steel have not been i

ksi in.8'2), which, for this case, yields a, (minimum or critical measured. He 25% reduction is arbitrary and was chosen to flaw size) of ~ 0.68 s/t-very large. If the shelf were taken as 275 S

MPa m"2 (250 ksi in.u2) there would be no e, value, which for all

'I '

• I '"ge occurs in the temperature distnoution near N

practical purposes is not much diff'erent than 0.68 a/r. If the the vessel.inside d.iarneter, but towards the sessel outside diameter shelf were 182 MPa m"2 (165 ksi in.2), there is, of course, there is about a 28'C (50*F) difTerence in temperature.

a reduction in the a, value to 0.54 a /r, again, very large.

In arriving at this plot, a 25% reduction in conductivity of both Dus in the case of an accident evaluation the exact value of gg the upper shelf is not scry significant, ne opposite is true for the the eiTect of the clad is negligible and the results portrayed are djus rIt s I at if initiation had occurred at about

"* "I"Y ***,pletely to the assumed reduction 'un base metal igu e al

• • ""***'I' 0.08 a/t, antst could be expected at a/r ~0.30, that is, where e thermal stress profile resulting from this, temperature pro.

K falls below the Ki, curve Thus, the validity of the arrest con-n in Fig.11. In this figure it can be scen that a signifi.

3 e s cept and the location of the arrest curve can be extremely impor.

1. ant base m med sims has muW fmm the M reh tant, in particular when the initiation flaw size turns out to bc

t. ion in conductivity, increased tensile stress in the inner region,

',,,jg*

and increased compressive stress in the outer region. A fracture 600 Seconds 700 Seconds a

600 Seconds O

i Stress intensity (KSidiii) 300 Eg a

C 800 Seconds 900 $econds

,8 8

200 I

/

4 X

/

[I

/

100 7"aA 1000 Seconds 1100 Seconds i I I I I f f I' l o0.0 0.2 0.4 0.6 0.8 1.0 Fractional Distance (a/t)

FIG. 9-Thermal shoch rmluation; efm of upper shelf (I hal in."

FIG. S-Thermal shoch emination for longer times inte she transient.

= 1.1 Mra m).

)

1 I

i

)

I 1

1

< "294 JOURNAL OF TESTING AND EVALUATION mechanics analysis would be required to evaluate the impact of envirnnment-water versus air [2]. At low frequencies the data chinges in thermal condQetivity on the reactor vessel integrity, appear to differ by a factor of about 20 owing to environment.

as shown in Fig.12. His Ogure is schematic and is intended only ( illustrate the difference caused by the environmental ctTect fatigue Crack Crowtir and the large conservatism contained in fatigue calculations when Data obtained to date for reactor vessel steels indicate a large no consideration is given to the presence nf the cladding, which dependence of crack growth rate on loading frequency and prevents access of the water to the base metal.

The presence of an environmental effect suggests the possibility of an in pile effcet since the radiatioa environment can break Temperature (*F) down the water and lead to species which may control the en.

hanced crack growth rates observed when tests are conducted in 400 water. his is an area that should be evaluated.

0.75 K J Dere is a dcarth of data on crack growth rates peninent to actual reactor vessel evaluations, that is, irradiated material tested in a water environment. Limited data exist for irTadiated t.0 K 300 material tested in air (3.4j and no data exist for testing in a water environment. Here does not appear to be an cfTect of irradiation

. on crack growth rates when the testing is performed in an air environment.

200 Since the present analytical methods. which do not consider the presence of the clad, are conservative for reactor vessel fatigue l

crack growth evaluations, the analysis results could be improved 200o seconds After Large LOCA by obtaining and using more pertinent fatigue crack growth data

]

100 obtained in vacuu m or air environments.

I I

I I

I 00 0.10.20.30.40.50.60.70.80.91.0 da/dN a/t Vessel Wall Thickness (Inches / Cycle)

FIG. 10-Temperature prof;Ies: effect of thermal conductMay l'C = (Y - 32)/I.8; LOCA = lowof coolant accidensl.

Axial Thermal Stress (ksi) 1 45 W ater Fa tor l

4g g

Environment 20 I

35 0 75K 4

A

i 30 Environment y

25 20 1.on i

15 10 Stress Intensity Range 1

5 FIG.12-Schematic diagram of crack growth rete wrsus stress l

O intensity rence it in. - 214 mm).

-5 i

10 Summary i og 1 200o seconds Several areas have been reviewed to identify conservatism in the 7

After Large LOCA 0.75K determination of matcrial property variations and ihe sensitivity of

-2O I I I I l l I

I I

the stmetural integrity analysis results to the material propeny

-25 j

0 0.2 0.4 0.6 0.8 1.0 variations. De areas covered were fluence dependence of the transition tempe:ature, fracture toughness and the drop in upper

• a/t Vessel Wall Thickness shelf, the arrest toughness, the thermal properties, and the crack l

FIG. It-Thermal stress profiles: r#wt of thermal conduc May growth rate. Majnr interest is in denning the fluence deperulence (I Asl = 6.9 Afib: 1 OCA = fuwof content scrident),

of irradiation embrittlement. In particular for thosc evaluations i-

295

,.'CH ftigOS AND MEYER ON NUCLEAR COMPONENTS i

Acknankdgments

~ ' that require extrapolation into the future with resulting higher The authors wish to express their appreciation to J. M. Krampe fluence. Accurate methnds for predicting changes in transition temperature and dcGmng fracture toughness for irradiated and O. Mccuwis for their contributions to the acchnical ba material are needed. Indications are that knowledge of toughnessgg; is the low toughness region of the fracture toughness curve is more important than toughness at the upper shelf.

References When the assessment of a vessel indicates small Daw sizes for reliance on arrest to demonstrate (1J Bamfont. W. H. and Duchalet. C. B.. " Method for Frac initiation and consequent Andpis of Nuclear Reactor Vesseh Under Sescre lhermal Tran, adequacy, then the arrest toughness for irradiated material must sients," WCAP 8510. Westinghouse Commercial Atomic Power Division, Westinghouw Hectric Co.. Pittsburgh. Pa.. July W70.

be accurately known.

Finally, the thermal prop.rties of irradiated reactor vessel12) Damfoid. W. H., Moon D. M. and Sett K. V., "Esaluation of material should be evahtsted since an irradiation ciTect can affect Critical Factors in Crack Grnsth Nate Studres of LWH Steels evaluation of vesselintegrity undcr accident conditions.

presented at the Fourth Water Reactor Sarcty Information Meeting, G4thersburg, Md., Oct.1976.

f Aslong as cladd.ing is not considered.m crack growth cvalusu,on, (3) Shahinian. P., Watson. H. E., and Hawthorne J. R., " Fatigue then the behavior of irradiated material in a water environment Crack Growth Resistance of Several Neutron irradiated Pressure must be established, and it should be determined whether an Vessel Steels and Welds." Transactions of ese American Society of 1

MccAunicalEnginccrs. Vol %. Senes If.1974, pp. 242-248.

in pile effcet on crack growth rate in a water environment exists.

[4] James. L A. and Williarrs, J. A. "The Effect of Temperature and Also,.f co,nsideration is to be given for the clad, more crackNeutron Irradiation upon she l'asigue Crack Propagation ficharior ot' i

growth data in vacuum or air should be obtained to eliminate the ASTM A533B, Ctau I Stect," HEDL.TME.72132, llanford present conservatism.

Engineering Development Laboratory, Atomic Energy Commission.

Washington. D.C., Sept.1972.

l e

l

/

s s

O e

.O e

4 i

k

^^

^ ^ ~ ~ ~ - - -

-