ML20138H895

Rev 0 to Leak-Before-Break Evaluation of Reactor Coolant Loop, Final Rept

ML20138H895
Person / Time
Site:	Trojan File:Portland General Electric icon.png
Issue date:	06/30/1985
From:	Bak W, Emerson R, Grubb R ABB IMPELL CORP. (FORMERLY IMPELL CORP.)
To:
Shared Package
ML20138H881	List: ML20138H876 ML20138H895
References
01-0300-1395, 01-0300-1395-R00, 1-300-1395, 1-300-1395-R, TAC-06596, TAC-6596, NUDOCS 8510290174
Download: ML20138H895 (38)

Text

,

Leak-Before-Break Evaluation of the Reactor Coolant Loop Final Report Submitted to:

PORTLAND GENERAL ELECTRIC COMPANY Prepared by:

Impe11 Corporation 350 Lennon Lane Walnut Creek, California 94598 Impe11 Report No. 01-0300-1395 Revision 0 8510290174 851025 PDR ADOCK 05000344 June,1985 P pm

Page 2 of 38 IMPELL CORPORATION REPORT APPROYAL COVER SHEET Client: Portland General Electric Company Project: Trojan Nuclear Plant Job Number: 0300-023-1356 R: port

Title:

Leak-Before-Break Evaluation of the Reactor Coolant loop R; port Number: 01-0300-1395 Revision 0 The work described in this Report was performed in accordance with the Impell Quality Assurance Program. The signatures below verify the accuracy of this Report and its compliance with applicable quality assurance requirements.

Prepared By:

~ ' ^ ^ ~ ~

Date: #!24h5 Robert Emerson, Supervising Engineer Reviewed By: d[

Walter R. Bak, M r, Date: 4/2///d

/ '

Applicd Mechanic tion Approved By: -

< Date:gr d9 T

' Robert L. Grubb, Manager, Advanced Engineering Division REVISION RECORD Rev. Approval No. Prepared Reviewed Date Revision 1

. Portland General Electric Rep 3rt No. 01-0300-1395 Revision 0 Page 3 of 38 TABLE OF CONTENTS Page 1.0 Introduction 4 2.0 Methodology and Numerical Results 5 2.1 Criteria 5 2.2 Description of Piping System 5 2.3 Crack Detectability Under Normal 6 Operating Conditions 2.3.1 Crack Leakage Analysis 2.3.2 Leak Detection Systems 2.3.3 Detectable Crack Size 2.4 Fracture Mechanics Analysis 8 2.4.1 Loading Conditions and Critical Locations 2.4.2 Local Stability of Postulated Crack 2.4.3 Limits of Applicability 2.5 Material Properties 13 2.5.1 J-Integral and Tearing Modulus Data for Base Metal 2.5.2 J-Resistance Curves for Welds 2.6 Net Section Evaluation 14 2.7 Summary of Results 14 3.0 Conclusions 15 Tables 16 Figures 24 References 29 List'of Symbols 30 Appendix A Description of Computer Programs 32 Appendix B Sensitivity of Leakage Detection 37 Systems i

Portland G nzral Electric R: port No. 01-0300-1395 Revision 0 Page 4 of 38

1.0 INTRODUCTION

A leak-before-break evaluation has been perfomed to define the potential for pipe rupture on the reactor coolant loop (RCL) piping of the Trojan Nuclear Plant. The approach is based on demonstrating that rather than suddenly breaking in a double-ended guillotine fashion, as has been previously postulated, the RCL piping would develop stable, detectable cracks prior to rupture. The approach follows a defense-in-depth philosophy by showing substantial margin in each of the analytical steps.

The approach uses a three-step evaluation based on deterministic fracture mechanics to show that a double-ended rupture of the RCL piping at Trojan is not a credible event. First, the detectability of a large through-wall flaw under nomal operating conditions is established by comparing calculations of the fluid leakage through the crack with the detection capabilities of five existing leak detection systems. Second, the local stability of the crack under faulted loading conditions is detemined using elastic-plastic fracture mechanics techniques and lower-bound estimates of the fracture properties of the RCL piping and weld materials.

Third, the global stability of the RCL is evaluated by showing that the reduced section at the postulated crack location is capable of carrying the piping loads for faulted conditions.

i

Portland General Electric Report No. 01-0300-1395 Rsvision 0 >

Page 5 of 38 2.0 ETHODOLOGY AND NUMERICAL RESULTS 2.1 Criteria The leak-before-break evaluation was performed in accordance with the NRC criterie outlined in Generic Letter 84-04. The five specific criteria used are: .

1) The loading considered in the analysis should include the static loads (pressure, deadweight and thermal) and the loads associated with safe shutdown earthquake (SSE) conditions.

2) The evaluation is performed considering that a large circumferential throughwall flaw is postulated to exist in the pipe wall. The length of the flaw is to be the larger of either (a) twice the wall thickness or (b) the flaw length that corresponds to a calculated leak rate of 10 gallons per minute (gpm) at normal operating conditions.

(3) The material resistance to fracture should be '

based on a reasonable estimate of lower bound properties as measured by the materials resistance (J-R) curve.

(4) The applicability of J-integral and tearing modulus computations should be addressed to ensure that the results are acceptable for engineering application.

(5) The flawed piping should be evaluated to demonstrate that the code limits for faulted conditions ere not exceeded for the uncracked section.

2.2 Description of Piping The Reactor Coolant System (RCS) consists of four

. System similar heat transfer loops connected in parallel to the reactor pressure vessel (RPV) and other compo-nents such as the pressurizer and connecting piping.

Figure 1 shows a typical loop; the locations considered in the fracture mechanics analysis are indicated. Each loop contains a reactor coolant pump, steam generator and three segments of large-diameter piping. The three segments of loop piping are the hot leg, which connects the RPV to the steam generator; the crossover leg, which connects the steam generator to the reactor coolant pump; and the cold leg, which connects the reactor coolant pump to the RPV. The diameters and wall thicknesses of the three segments of piping are given in Table 1.

l

1 Portland General Electric Report No. 01-0300-1395 Ravision 0

Page 6 of 38 F

2.0 ETHODOLOGY AND NUMERICAL RESULTS i* The reactor coolant loop piping and fittings are austenitic stainless steel and the standards j governing the materials used in the piping, fittings, and welds are given in Table 2. Smaller piping which comprises part of the RCS boundary, such as the pressurizer surge line, spray and relief line, loop

~

< drains and connecting lines to other systems, is also austenitic stainless steel.

The design pressure of the RCL is 2485 psig and the nomal operating pressure is 2235 psig. The design l temperature is 650*F and the nomal full power

! operating temperature is 616.8'F for the hot leg, i 552.3*F for the crossover leg, and 552.5'F for the i

cold leg.

2.3 Crack Detectability The size of the postulated flaw evaluated in the Under Nomal Operating leak-before-break analysis is defined by comparing Conditions the leak rates from postulated cracks under normal operating conditions with the sensitivity of existing systems available to detect leakage from the primary system. The analyses perfomed to detemine the detectable crack size are described in this section.

2.3.1 Crack Leakage Analysis  ;

Two calculations are perfomed to detemine the leak rate through a postulated crack. First, the crack opening area of a postulated crack under the applied nomal operating loads is detemined. Second, the fluid flow through the crack opening area is calculated. These calculations were perfomed using the computer codes CRACK and IMLEAK, respectively.

CRACK is based on Reference 1 and is described in Appendix A. CRACK calculates the crack opening area using linear elastic fracture mechanics, including corrections for plastic zones at the crack tips, and considers internal pressure, axial force and bending moment contributions to crack opening.

IPLEAK is based on Reference 2 and is described in Appendix A. IMLEAK evaluates flow through thin cracks in piping given the upstream themodynamic l

l

Portland G:neral Electric

! Report No. 01-0300-1395

< Revision 0 Page 7 of 38 2.0 PETHODOLOGY AND NUMERICAL RESULTS i

conditions and the geometric configuration of the crack. The flow model accounts for two-phase flow and includes pressure drop tems due to entrance <

geometry, friction, and acceleration.

i For this analysis, the leak rate calculations were performed for several postulated circumferential

crack lengths, ranging from two to nine inches. Two i possible loads due to nomal operating conditions i were considered. First, axial loads due to i self-balancing pressure stress were used to detemine lower bound leak rates. Second, bending moments due to themal loads were added to the equivalent 4 pressure loading to yield an estimate of the upper

bou'nd leak rate for nomal operating loads.

l The leak rate analyses were performed using the pipe i section geometry and operating conditions of the hot

leg. The bending moment applied to detemine the

leak rate in the second analysis described above was j the moment at the junction of the hot leg and the RPV.

I The leak rates calculated for the postulated circumferential cracks are tabulated in Table 3. The results considering the pressure only provide a lower bound estimate of the leak rate. The results which include the themal bending moment provide an estimate of the leakage that would result from a crack at a location of high stress. The leak rate will lie between these bounds depending upon the loads at the location of the postulated crack. For locations of high stress, where a crack would be more likely to develop, the leakage would be high. For -

locations of lower stress, which have less potential for crack development, the leakage would be closer to the lower bound leak rates.

2.3.2 Leak Detection Systems There are five methods of detecting leakage from the reactor coolant loop piping. The sensitivity and response time required to detect a one gpm leak rate have been calculated previously and are discussed in

'l Portland Gensral Electric Rzport No. 01-0300-1295

, Revision 0 Page 8 of 38 i 2.0 ETHODOLOGY AND NUMERICAL RESULTS I Section 5.2.5.4 of the Updated Final Safety Analysis i Report, Reference 3. The results are presented in

UFSAR Table 5.2-9 which is reproduced in Appendix B I

of this report. The results are summarized in i

Table 4.

2.3.3 Detectable Crack Size i The leak rates calculated for several lengths of 4

postulated circumferential cracks are presented in Figure 2. The leak rates calculated using pressure stress only and using pressure and themal bending F are graphed separately. Also shown in the figure are the sensitivities of the five existing leak detection systems.

i The lower bound leak rate curve shows that a crack j_ slightly longer than seven inches can be detected by 4

all five systems. The leak rate calculation which i includes bending shows that leakage from a 3.5-inch j crack can be detected by all five systems and that j the leak rate from a seven-inch crack is in excess of

eight times the flow rate detectable by the least

{ sensitive system.

j As discussed previously, the actual leak rate will depend upon the loads at the location of the crack.

t The lower bound leak rate results show that a seven-inch crack can be detected. The upper bound j results show that a margin of two on detectable crack i- 1ength exists for a seven-inch crack and that the '

leakage from a seven-inch crack would be three times the 10 gpm required by the criteria.

l Based upon these results, a seven-inch-long crack is sufficiently large such that it has a very high probability of detection under nomal operating conditions and is selected as the crack length for evaluation under faulted conditions.

2.4 Fracture Mechanics The postulated through-wall crack which was shown to Analysis be detectable under nomal operatir; conditions is evaluated to detemine the potential for unstable growth under faulted loading conditions. The

. Portland Gen 2ral Electric Report No. 01-0300-1395

Revision 0

- Page 9 of 38 l

2.0 METHODOLOGY AND NUMERICAL RESULTS

! fracture mechanics analysis perfomed to detemine the stability of the postulated crack is described in this section.

I 2.4.1 Loading Conditions and Critical Locations The stress analysis of the RCL piping was performed previously and is described in Reference 4. The analysis was perfomed for Loop 2 which includes the 14-inch-diameter piping that connects the pressurizer to the RCL. Six load cases were analyzed in the stress report: deadweight of piping, wet weight of primary equipment, thermal, pressure, operating basis earthquake (OBE) X+Y, and OBE 2+Y.

The separate load case results were combined to provide upper bounds on the faulted condition loads.

Deadweight, equipment wet weight, thermal, pressure, and safe shutdown earthquake (SSE) results were conservatively combined to give upper bound estimates of axial force and bending moment for the RCL piping. (The SSE load is 1.67 times the maximum of OBE X+Y and OBE Z+Y load.)

The upper bounds on faulted loads were determined for the teminal ends of each of the three piping segments. In addition, upper bounds on the loads at the intemediate locations were calculated by enveloping the load case results for several intermediate locations on each segment. The enveloping procedure results in intemediate loads which are higher than the teminal end loads and which ensure that conservative upper bound faulted loads are considered in the analysis.

The faulted condition loads used in the fracture mechanics analysis of the postulated seven-inch crack are listed in Table 5.

2.4.2 Local Stability of Postulated Crack The local stability of a seven-inch-long postulated crack was determined by calculating the applied J-integral at each of the nine locations listed in

Portland Gen 3ral Electric Report No. 01-0300-1395 Revision 0 Page 10 of 38 2.0 METHODOLOGY AND NUMERICAL RESULTS Table 5. In addition, the applied tearing modulus was calculated for the location of the highest value of the J-integral.

For cases of purely elastic or small-scale contained plasticity, the J-integral may be derived from the stress intensity factor using the relations: _

J = K2 /E (for plane stress)

J = K2 (j,p 2)/E (for plane strain)

(Definitions of the terms used in this report are provided in the List of Symbols on page 30.)

As discussed in the following section, the plastic zone at the tip of the postulated crack is small even under faulted load conditions. Therefore, J-integral values derived from the stress intensity factors accurately represent the potential for crack growth.

The stress intensity factors were calculated using the program CRACK which is based on Reference 1 and is described ir. ?ppendix A. CRACK includes the standard correction for effective crack length due to yielding at the crack tip. Minimum material properties from the ASME BPVC (yield stress, ultimate stress and elastic modulus) for the piping material at the operating temperature were used to ensure upper bound estimates of the effective crack length and the stress intensity factor.

Further, the conversion to the J-integral was perfonned using the formula for plane stress conditions which results in upper bound J-integral values.

The results of the J-integral calculations are presented in Table 6. The maximum value is 595 in-lbf/in2 which was calculated for the upper bound intermediate loads on the hot leg.

j

Portland General Electric Rgpart No. 01-0300-1395 Revision 0

Page 11 of 38

! 2.0 METHODOLOGY AND NUMERICAL RESULTS The applied tearing modulus was calculated by l detemining the variation in the J-integral value for small changes in the crack length. The tearing i 4

modulus was then calculated using the relation: ]

T = (dJ/da) (E/ a,2 )

The tearing modulus calculated for the intemediate i location on the hot leg is 2.5 (dimensionless).

2.4.3 Limits of Applicability 4

f When the yielding associated ~with a given crack and loading condition is small, the J-integral based on the elastic stress intensity factor can be used to

, assess stable crack growth using the resistance curve i concept. That is, for cases where small-scale

yielding occurs, the stress intensity factor and J-integral based on it uniquely measure the intensity

of the stress field at the crack tip. However, for i cases where crack growth occurs under conditions of 1arge-scale yielding, additional requirements must be i met for application of J-resistance curve approach.

The plastic yielding r!ade at the crack tips for the 4

analyses presented W e small relative to the

_ total section of he t?g Table 7 shows the size of the plastic zone as the cd

k tip and tabulates the i size of the zone as a percentage of the uncracked j portion of the pipe section. This comparison i demonstrates that the yielding is small relative to i the total section. Therefore the 'Jse of the J-integral based on the stress intensity factor using i plastic zone corrections to the crack length is

appropriete. To further demonstrate the adequacy of

? the approach the requirements for fully-yielded

sections are also applied. However, it is to be

expected that the margins of applicability will be large since the yielding is small.

l j

i

Portland Gensral Electric Report No. 01-0300-1395 Revision 0 Page 12 of 38 2.0 METHODOLOGY AND NUMERICAL RESULTS Several requirements for assessing the applicability

of J-integral controlled fracture have been proposed. (See, for instance, Reference 5.) Two of the more widely used are evaluated here. The first requirement applies to both initiation and growth phases of cracking. The requirement ensures that the crack opening displacement is small relative to the

, wall thickness and uncracked ligament section. The requirement is:

b > 25 J/ a, The uncracked ligament, in the case of our analysis, is in excess of 40 inches. The pipe wall thickness is conservatively used as the structural dimension.

The inequality then becomes, for the maximum value of the J-integral calculated in this analysis:

2.43 > 0.37 The first condition is met.

The second requirement is based on considerations for the applicability of the deformation theory upon which J is based. The requirement is based on the nondimensional parameter w:

b dJ w = 7 7g ?C> 1 In evaluating w, we again conservatively use the pipe wall thickness for b. The lowest value of dJ/da for the two weld materials considered is 7800. Then:

w = 32 1

Generally a value of 10 is considered adequate, and I therefore this condition is met. l

Portland General Electric Report No. 01-0300-1395 Ravision 0 Page 13 of 38 2.0 METHODOLOGY AND NUMERICAL RESULTS 1

2.5 Material Properties 2.5.1 J-Integral and Tearing Modulus Data for Base Metal i

i The austenitic stainless steels used in the

! manufacture of reactor coolant piping exhibit

extensive ductility. Quantitative characterization

) of the resistance to initial crack growth and the stability of crack growth is based on measurement of i the J-integral and the tearing modulus. Measurements

of these properties for stainless steel base metal i were reviewed in References 5 and 6 and are 1 1 summarized in Table 8. Only data from tests at temperatures near the operating temperature of the RCL are included. Comparison of the JIC data for i '

base metal with the J-R curves for weld metal

j. discussed in the following section, shows that the weld metal has lower fracture toughness and, j therefore, governs the crack extension stability j analysis.

l 2.5.2 J-Resistance Curve for Weld Material i

j Data on the fracture resistance of stainless steel weld material was recently published in NUREG-1061,

Report of USNRC Piping Review Committee, Reference i 7. Data was presented for submerged arc welds (SAW) 4 and inert gas tungsten arc welds (GTAW) in the form of applied J-integral versus measured crack l extension. From this data, J-integral versus tearing i modulus curves were derived for the two types of

! welds. The data from Reference 7 is reproduced in

! Figures 3, 4 and 5.

4 To determine which welding process was more similiar 1 to those used in the Trojan RCL, four typical weld j procedures used on large diameter RCS piping were

! reviewed. The procedures include shielded metal arc

! welding (SMAW) and gas - tungsten arc welding 1

(GTAW). These two welding procedures were used for  ;

i.

the entire weld, or, alternately, the intitial GTAW '

i root passes were completed using SMAW passes. No other welding procedures are allowed by the specifications. )

)

1 l .

4 i

l

Portland General Electric -

R: port No. 01-0300-1395 Revision 0 Page 14 of 38 2.0 METHODOLOGY AND NUMERICAL RESULTS Consequently, the weld procedures used on the RCL are (for GTAW) identical or (for SMAW) closer to the J-resistance curve for GTAW welds. The lower

. fracture resistance curve for SAW which produces l coarser grain stucture with lower ductility, provides l a definite lower bound for fracture properties of the welding procedures used on the RCL.

2.6 Net Section Evaluation A large flaw such as the one postulated to produce detectable leakage from the RCL will reduce the load carrying capacity of the piping. To determine whether this effect is significant for the seven-inch-long flaw postulated for this analysis of the Trojan RCL, the net section stresses were evaluated using the ASME Code procedures and limits for faulted conditions. The primary stress for faulted conditions, calculated in accordance with Subsection NB-3656, is 27.9 ksi which is below the allowable of 37.6 ksi. Therefore, there is margin against failure of the flawed piping due to net section plasticity.

2.7 Summary of Results The results of the fracture mechanics evaluation are summarized in Figure 5. The values of the applied J-integral and tearing modulus calculated for the highest loads on the RCL piping are plotted with the material J-T curves for two stainless steel weld material s. Based upon the upper curve there is a wide margin against crack extension. Based upon the lower curve, some crack extension will occur.

However there is a substantial margin on the tearing modulus which ensures that crack extension will occur in a stable nenner.

Portland General Electric Report No. 01-0300-1395 Revision 0 Page 15 of 38 1

3.0 CONCLUSION

S i

The analyses presented in this report demonstrate that large margins against unstable crack extension

- exist for RCL piping postulated to have a long j throughwall crack. Lower bound leak rate 4

calculations show that a seven-inch circumferential i crack is detectable by all five existing leak detection systems. Consideration of thermal . bending

results in higher leak rates and, at highly stressed 4 locations, the leakage from a seven-inch crack is 30 gpm. Therefore, it is concluded that substantial i ,

margin to detect a seven-inch throughwall crack under normal operating conditions exists.

F Local stability was shown using conservative load combinations of plant operating loads and safe shutdown earthquake loads and lower bound fracture i mechanics properties of stainless steel weld metal.

j Review of the weld procedures used on the RCL 4 confirmed that the data for submerged arc welding l provides a conservative lower bound on the fracture

resistance of the RCL welds. For this lower bound data, a margin in excess of 40 for the applied

~

tearing modulus versus material tearing modulus was shown. It is concluded that a wide margin against unstable crack extension exists under faulted i conditions for the RCL piping.

4 The margin against net section failure of the cracked section was evaluated using upper bound loads for the faulted condition. The allowable stress is 33%

! higher than the upper bound calculated stress and

!. therefore substantial margin exists against net section failure.

Therefore, based on analyses performed in accordance

with the criteria of Generic Letter 84-04, it is

concluded that pipe rupture of the reactor coolant loop at Trojan Nuclear Plant in a double-ended guillotine manner is not a credible event.

I i  !

I 1

Partland General Electric R: port No. 01-0300-1395 Rsvisicn 0 Page 16 of 38 TABLE 1 REACTOR COOLANT LOOP PIPING DIMENSIONS Piping Segment Inside Diameter (inch) Wall Thickness (inch)

Hot Leg 29.0 2.43

, Crossover Leg 31.0 2.57 Cold Leg 27.5 2.30

Portland General Electric R: port No. 01-0300-1395 Revision 0 Page 17 of 38 TABLE 2 REACTOR COOLANT PIPING MATERIALS Component Material Reactor Ccolant Pipe Code Case 1423 Gr F304N or 316N, or SA351 Gr CF8A or CF8M Centrifugal Castings Reactor Coolant Fittings SA351 Gr CF8A or CF8M Welding Materials SFA 5.4 and 5.9 Type 308 or 308L l

Portland General Electric R! port Ns. 01-0300-1395 R vision 0 Page 18 of 38 TABLE 3 LEAK RATE UNDER NORMAL OPERATING CONDITIONS Leak Rate (gpm)

Crack Length (inch) Pressure Stress Only l ' Pressure plus Themal Bending 2 2 0.11 0.98 3 0.34 2. 91 4 0.75 6.26 5 1.40 11.3 6 2.33 19.7 7 3.57 30.0 8 5.18 43.4 9 7.22 60.3 1 - Axial stress equal to pr/2t.

l 2 - Axial stress equal to pr/2t plus themal bending moment at RPY outlet nozzle / hot leg.

Portland Gen:ral Elsctric R: port No. 01-0300-1395 Rsvisien 0 Page 19 of 38 TABLE 4 SENSITIVITY AND RESPONSE TIME OF LEAK DETECTION SYSTEMS Sensitivity Response Time (leak rate detectable (Time required to detect Leakage Detection System in one hour) 1 gpm leak rate)

C;ntainment Air Particulate Monitor 0.6 gpm 45 min Cantainment Gaseous Activity Monitor 2.5 gpm 100 min Containment Humidity Detector System 2.0 gpm Not Sufficiently Sensitive Containment Condensate Monitoring System 0.8 gpm 50 min Containment Sump Level 3.7 gpm 220 min I

From Reference 3. See Appendix B for applicable notes concerning methodology and  ;

assumptions.

l l

1 l

1 I

Portland Gen ral Electric RIport No. 01-0300-1395 Revision 0 Page 20 of 38 TABLE 5 LOADS FOR FAULTED CONDITION EVALUATION Location and 1 Piping Segment Stress Analysis Node F (KIPS) M (KIP-IN)

RPV Outlet (Node 405) 2085 30,406 2

Hot Leg Intermediate 2586 29,264 SG Inlet (Node 420) 2187 30,014 SG Outlet (Node 438) 1778 10,953 2

Crossover Leg Intemediate 1938 22,069 RCP Inlet (Node 459) 1968 20,424 RCP Outlet (Node 468) 1542 12,307 Cold Leg Intemediate 2 1560 16,404 RPY Inlet (Node 483) 1551 15,404 1 Locations and stress analysis nodes are shown in Figure 1.

2 - Upper bound estimate of maximum loads on intermediate piping segment.

l l

l i

1

Portland Gen:ral Electric R! port No. 01-0300-1395 Rovision 0 Page 21 of 38 TABLE 6 J-INTEGRAL VALUES FOR FAULTED LOADING CONDITIONS Location and Piping Segment Stress Analysis Node l J (in-lbf/in 2)

RPV Outlet (Node 405) 495 Hot Leg Intemediate 595 SG Inlet (Node 420) 509 SG Outlet (Node 438) 74

- Crossover-Leg Intermediate 173 RCP Inlet (Node 459) 159 RCP Outlet (Node 468) 136 Cold Leg Intemediate 195 RPV Inlet (Node 483) 179 1 - Locations and stress analysis nodes are shown in Figure 1.

Portland General Elcctric Report No. 01-0300-1395 Rsvisicn 0 Page 22 of 38 TABLE 7 PLASTIC ZONE SIZE UNDER FAULTED LOADING CONDITIONSI Piping Segment Stress Analysis Node 2 r(3) (in)  % Plasticity I4)

RPV Outlet (Node 405) 1.19 2.60 Hot Leg Intermediate 1.42 3.10 SG Inlet (Node 420) 1.22 2.66 SG Outlet (Node 438) 0.172 0.35 Crossover Leg Intemediate 0.376 0.76 RCP Inlet (Node 459) 0.349 0.71 RCP Outlet (Node 468) 0.301 0.69 Cold Leg Intermediate 0.41 7 0.96 RPV Inlet (Node 483) 0.386 0.89 1 - R:sults are for 7-inch circumferential crack 2 - Locations and stress analysis nodes are shown in Figure 1.

3 - Radius of plastic zone correction at crack tips 4 - Percentage of the uncracked piping section included in the plastic zones at the tips of the postulated crack I

l

Pertland General Electric Rep;rt No. 01-0300-1395 Revision 0 Page 23 of 38 TABLE 8 J-INTEGRAL AND TEARING MODULUS DATA FOR BASE METAL d

Temperature IC T Material Specimen l (*F) (in-lbf/in )

Reference Reference 5 304 Forged CT 600 2569 2 31.1 304 Forged 3PB 600 2737 547.0 304 Forged CT 600 2308 388.5 316 Cast CT 600 1933 225.1 316 Cast 3PB 600 1200 612.6 304 Forged CT 600 1500 333.8 Reference 6 304 3PB 550 6000 750 304 3PB 550 6000 10uo 31 6 CT 600 5260 609 1 - CT is compact tensile specimen, 3PB is three-point bend specimen.

Portland General Eltetric Report No. 01-0300-1395 Revision 0 Page 24 of 38 FIGURE 1 REACTOR COOLANT LOOP GE0 METRY (TYPICAL) f

/

q

.4N l

~'

y, x 4 sf( / .>-

~

j

' .;pss*

/

\ 'i' 3 &' A

- B i C ,

95~ I i +

v

= )

m a p y,::&y v

(f' g

< -a. 'm N

A Ny.' i E k, N N '

~

~.,

Y(

% \s k

Od: notes break location analyzed:

A - Reactor Pressure Vessel Outlet Nozzle (Node 405)

B - Steam Generator Inlet Nozzle (Node 420)

C - Steam Generator Outlet Nozzle (Node 438)

D - R; actor Coolant Pump Inlet Nozzle (Node 459)

E - Reactor Coolant Pump Outlet Nozzle (Node 468)

F - Reactor Pressure Vessel Inlet Nozzle (Node 483)

(Bounding loads for intemediate locations on each leg were also analyzed.)

This figure is based on Figure 3.6-2 of the Trojan UFSAR (Reference 3).

Partland Gen:ral Elcctric R;p3rt N3. 01-0300-1395 R;visicn 0 Page 25 of 38 FIGURE 2 CALCULATED LEAK RATES FOR NORMAL OPERATING CONDITIONS AND LEAK DETECTION SYSTEM SENSITIVITIES 25 T 8ending 20 15 -

s g 10 Pressure Only Senst vte

~*

u '

/ '

/

-c W. d-'

1 2 3 4 5 6 7 8 9 10 CRACK $1ZI (Inch)

Leak Detection Systems:

a - Sump Level b - Gaseous Activity Monitor c - Humidity Detector System d - Condensate Monitoring e - Air Particulate Monitor

Portland Gen 3ral Electric R: port No. 01-0300-1395 Rzvision 0 Page 26 of 38 FIGURE 3 J-RESISTANCE CURVE FOR TUNGSTEN INERT GAS WELD (550*F)

!4000- 202 side croov .

A IT COMPACT SPECIMEN D 1T COMPACT SPECIMEN C C 2T COMPACT SPECIMEN Q D 2T COMPACT SPEC 0 '

0 I O D 3

E  ! O O

O

$sese.

a

' o* 8 a

4eaa. / $o,a

, 0 a m

D au MH. &

O O

/ On a 2000. goo 6

3 m . . .

I 8.000 "A.830 8.968 8.800 8.128 8.150 8.188 8.210 8.248 i CItACK EXTENSION in INCHES From NUREG 1061, Volume 1 Report of the USNRC Piping Review Committee, Reference 7.

Psrtland Gensral Electric Report No. 01-0300-1395 Revisien 0 Page 27 of 38 FIGURE 4 J-RESISTANCE CURVE FOR SUBMERGED ARC WELD (550*F)

E as e m' s s e n e e e a j, j D 4,g. - 202 side Crooves r

y<.ee -

j j n

(3.se .

j j -

1 ls.ee -

l l -

d 2.se -

/ / ,,

Ie2.se -

/ /

L s 1.se -

/ /

" se,r,e.

ye , + _..r- 4.vsA-s2 I.

1 -

/

• EN G . - AH store - 7 ell .e7

} y-zu7. -

2 ,,g. / alc _ si.s.s.v.i ss . _

,, / i i / . . . . . .

8.08 S.20 9.48 0.08 S.88 1.80 8.28 1.44 1.84 1.80 2.88 CRACK 8810l(TH CIN) E-Si From NUREG 1061, Volume 1, Report of the USNRC Piping Rev%w Comittee, Reference 7.

Partland Gen 2ral Electric Rep:rt No. 01-0300-1395 R: vision 0 Page 28 of 38 FIGURE 5 J/T PLOTS FOR SUBMERGED ARC AND TUNGSTEN INERT GAS WELDS (550*F) 7 , , , , ,

6- -

5 - ~

TUNGSTEN INERT

, GAS WELD 4 -

.5 h3*

s 2 -

SUBMERGED ARC WELD

~ ~

J = 595

• T = 2.5 0

O 50 100 150 200 250 300 TEARING MODULUS. T From NUREG 1061, Volume 1, Report of the USNRC Piping Review Comittee, Reference 7.

1 i

I l

l

- Portland Gen:ral Elsctric R port No. 01-0300-1395 Rsvision 0 Page 29 of 38 REFERENCES 1 - H. Tada and P. Paris, " Estimation of Stress Intensity Factors and the Crack Opening Area of a Circumferential and a Longitudinal Through-Crack in a Pipe," Del Research Corporation, St. Louis, Missouri.

2 - R.P. Collier, et. al . , " Study of Critical Two-Phase Flow Through Simulated Cracks," Interim Report BCL-EPRI-80-1, Battelle Columbus Laboratories,1980.

3 - Portland General Electric Company, " Trojan Nuclear Plant - Updated Final Safety Analysis Report," July 1982 plus Amendments 1 and 2.

4 - Westinghouse Electric Company, " Stress Analysis of Reactor Coolant Loop / Support Systems for Trojan Nuclear Plant, Unit No.1," WCAP 8438, Revision 1, December 1975.

5 - W.H. Bamford and A.J. Bush, " Fracture Behavior of Stainless Steel," in ASTM 668, Elastic-Plastic 4 Fracture, November 1977.

6 - K.H. Cotter, et. al., Application of Tearing Modulus Stability Concepts to Nuclear Piping,"

EPRI NP-2261, February 1982, t

7 - Pipe Crack Study Group, " Investigation and Evaluation of Stress Corrossion Cracking in Piping in Boiling Water Reactors," NUREG-1061, Nuclear Regulatory Commission, Washington, D.C.,

1984.

i 4

4 1

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

Portland Gen 2ral Electric R part No. 01-0300-1395 Revision 0 Page 30 of 38 LIST OF SYMBOLS a Half-crack length a,77 Effective half-crack length, accounting for yielding A

b Crack opening area due to bending A

total Total crack opening area A

p Crack opening area due to pressure A

t Crack opening area due to axial tension b Structural dimension: uncracked ligament or wall thickness E Young's Modulus J J-integral Jy Calculated J-integral J

IC Material J value Kg Stress intensity K

b Stress intensity due to bending K

p Stress intensity due to pressure K

t Stress intensity due to axial tension M Bending moment p Internal pressure of pipe r Inner radius of pipe or radius of plastic zone at crack tip R Mean radius of pipe

Portland General Electric R; port No. 01-0300-1395 Revisicn 0 Page 31 of 38 LIST OF SYMBOLS (Continued) t Average wall thickness of pipe T Tearing modulus T, Axial load w EPFM nondimensional parameter 9 Half the subtended angle of the crack ab Longitudinal stress due to bending ag Flow stress ap Longitudinal stress due to pressure at Longitudinal stress due to axial tension A Non-dimensional crack parameter, A= a//tii F Poisson's ratio

Portland Gen:ral Electric R;part No. 01-0300-1395 RIvision 0 Page 32 of 38 APPENDIX A Description of Computer Programs Descriptions of the two Impe11 computer programs, CRACK and IMLEAK, which were used in the leak-before-break evaluations are presented in this appendix.

Portland General Electric RIport No. 01-0300-1395 R2 Vision 0 Page 33 of 38 APPENDIX A DESCRIPTION OF COMPUTER PROGRAMS Description of CRACK The program CRACK is based on the method developed by Tada and Paris (Reference 1). The equations used to calculate KI and opening area for circumferential cracks are presented below.

The circumferential through-crack may be subjected to a tensile axial load, Ta; a bending moment, M; and internal pressure, p.

The total KI is composed of three components:

Kt otal

• Kt+Kb+Xp where K

t

• 1. t ( Fr RG)Ft!

p Kb*#b (frRG)1/2 K

p = ap (7rR 6 )l/ F p

T

1. t" M

b"g]

1. p "

F t

= 1 + 7.5 (j) - 15 (f) ' +33(j)

F b=1+6.8(f)' - 13.6 (f) ' +20(f) 2 1/2 Fp = (1 + 0.3225A ) , for 05A51

= (0.9 + 0.25A), for 1<As5 I

. _ ~ - . -

Portland General Elcctric R2 port N3. 01-0300-1395 Revision 0 Page 34 of 38 APPENDIX A DESCRIPTION OF COMPUTER PROGRAMS The crack opening area for a circumferential through-crack is a function of the crack length, internal pressure, axial tension, and bending moment. The total crack opening area is given by:

At otal " At+Ab+Ap where

+

At*Ab= nR It [I + ( 4

}]

A p = h (2nRt) pG a t= T,/2rrRt Ob = M/wR t o a pR/2t p

1.5 2 I = 262 gj ,( ) {8.6 - 13.3 (j) + 24 (j) f t

+(f)3 f22.5-75(f)+205.7(f)2 -247.5(f)3

+242(f) ), for 0<G<100*

4 G = A2 + 0.16 A , for W1 p

Gp = 0.02 + 0.81 A2 + 0.30A3+ 0.03 A4 , for 1< A55 The above fomulae are based on linear elastic frac-ture mechanics. Usually the material near the crack tip yields, and this is accounted for by a plastic-zone correction. An effective half-crack length which accounts for the plastic yielding is computed by:

• Partland General Electric L R2psrt No. 01-0300-1395 Rsvision 0 Page 35 of 38 APPENDIX A DESCRIPTION OF COMPUTER PROGRAMS 2

K aeff = a + 2ra z o

The ae ff value thus computed is used to recalcu-late K I, and the procedure is repeated iteratively until the a ffe values between two consecutive computations are within 2%.

l 4

F 1

• Portland General Electric

• rip 3rt No. 01-0300-1395 Revision 0 i Page 36 of 38 APPENDIX A DESCRIPTION OF COMPUTER PROGRAMS Description of IMLEAK IMLEAK is a computer code developed to evaluate the leak rate of pressurized fluids through narrow cracks. The code is based on the approach outlined in Reference 2. The code is used in leak-before-break analyses to estimate leak rates from postulated cracks to detennine whether the crack is detectable.

The analytical model in IMLEAK is a modified verision of Henry's non-equilibrium two-phase flow model. The model accounts for non-equilibrium effects in the flow due to flashing within the flow path. The model also includes pressure drops due to entrance losses, friction, and fluid acceleration.

The model handles complex crack geometry, including turns in the flow path, variable flow area, and crack surface roughness. The program also contains its own steam properties subroutines. The program has been benchmarked against test results.

Partland General Electric 0 R;ptrt N3. 01-0300-1395 R;visien 0 Page 37 of 38 APPENDIX B Sensitivity of Leak Detection Systems Table 5.2-9 from the Updated Final Safety Analysis Report, Reference 3, is reproduced in this appendix.

i l

i

- - - - ---w

h 4 Portland General Electric R;p3rt No. 01-0300-1395 Revision 0 Page 38 of 38 APPENDIX B SENSITIVITY OF LEAK DETECTION SYSTEMS l

TABt.E 5.2-9 SENSITIVITY AND RESPONSE TIME OF RFACTOR COOLANT PRESSURE BOUNDARY LEAKAGE DrtECTION SYSTDI Sensitivity Response Time Leakage Detection (leak rate detectable (Time required to detect Sveten in one hour) 1 ass leak este)

Containment Air Particulate Monitor

• 0.6 spa 45 min Containment Gaseous Activity Monitor (b) 2.5 spa 100 min Containment Numidity Detector SystemI *I 2.0 spa Not Sufficiently Sensitive Contai e Condensate Monitoring System 0.3 spa 50 min containment Sump Leve1I *I 3.7 spa 220 min (a) (1) Sensitivity threshold f or identifying RCF3 leakage is 7.0 x 10-8 pC1/cc.

(2) Reactor coolant corrosion product activity level is 0.005 bci/cc.

(3) All other assumptions are as described in Section 5.2.4.4 for the Containment air particulate monitor evaluation.

(4) Detector delay time is 20 min.

[b] (1) Sensitivity threshold for identifying RCFB leakage is 1.0 m 10~6 pCi/cc.

(2) Reactor coolant gaseous activit 1evel is 0.15 bci/ce.

(3) Containment volume is 5.60 x 10{0cc.

(4) Containment leakage is sero. Normal recirculation flow is 655.500 cfm ,

and is sufficient to ensure complate minias.

(5) The time of activity buildup to detectable levels is much shorter than the half-life of elements of concern.

[c] (1) 30 percent of RCFB leakage flashes to vapor at the point of leakage.

(2) Containment leakage is sero. Normal recirculation flow of Containment air coolers is sufficient to ensure complete mining.

(3) Initial Containment conditions are 120*F and 53 percent RN.

(4) Numidity detector sensitivity threshold for identifying RCFB leakage is e 3 percent RN.

[d] (1) 30 percent of RCF3 leakage flashes to water vapor at the point of leakage.

(2) Containment leakage is sero. Normal recirculation flow of Containment air coolers is 355.500 cfm and to sufficient to ensure complete mixing.

(3) Air coolers condense all leakage which has flashed to vapor.

, (4) Sensitivity threshold for identifying RCFB leakage is three-valve drainage cycles within one hour.

[e] (1) 50 percent of RCFB leakage drains as liquid directly to Containment Building sump with least sensitive leakage detection espability.

(2) Sump is initially empty and has 10-ss11on/ inch capacity.

(3) Sump level indicators are provided as described in Table 11.2-9.

(4) Sensitivity threshold for identifying RCFB leakage is receipt of second sump level light.

i k

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