ML20211B074

Leak-Before-Break Justification Rwcu,Dresden Station Units 2 & 3,Quad Cities Station Units 1 & 2

ML20211B074
Person / Time
Site:	Dresden, Quad Cities, 05000000
Issue date:	05/24/1984
From:	James Gavula, Kluge M NUTECH ENGINEERS, INC.
To:
Shared Package
ML20211B064	List: ML20211B059 ML20211B074
References
NUDOCS 8610170121
Download: ML20211B074 (15)

Text

-

4 LEAK BEFORE BREAK JUSTIFICATION REACTOR WATER CLEANUP SYSTEM DRESDEN STATION UNITS 2& 3 QUAD CITIES STATION UNITS 1 & 2 Prepared for Commonwealth Edison Company Prepared by NUTECH Engineers, Inc.

Cnicago, Illinois May 1984 Appro e /by: Issue by:

Ao4/

7.. me,e - ..

,g' ! Jt. e7.

.. . du6

'8610170121861bo6

._{DR ADOCK 05000237

)

P DR .. .I nutggh

"~

o .

1.0 INTRODUCTION

The LBB concept is based on the plasticity of austenitic stainless steel. Whereas brittle materials can be described by linear elastic fracture mechanics (LEFM),

ductile materials such as austenitic stainless steel cannot be described by LEFM because of their gross yielding and subsequent plastic instability. It has been found that the failure behavior of materials like stainless steel can be described by the net section collapse criterion.1,2 This model assumes that failure will occur when the net section associated with a circumferential defect in a pipe reaches a critical stress value, og, which is the flow stress of the

~

material. This LBB model is acceptable to the U. S.

Nuclear Regulatory Commission (NRC) in describing the

, behavior of the above mentioned piping with respect to intergranular stress corrosion cracking (IGSCC).

1 Machanical Fracture Predictions for Sensitized Stainless Steel Piping with Circumferential Cracks", Final Report, September 1976 (EPRI Report NP-192).

2 "Raview and Assessment on Research Relevant to Design Aspects of Nuclear Power Plant Piping Systems", Nuclear t

Regulatory Commission, July 1977 (NUREG-0307).

LBB-1 110 tech

o O 2.0 APPROACH The approach used consists of two steps:

1) Using membrane and bending stresses that are considered conservative (i.e., high), calculate the through-wall circumferential crack length at which the membrane and membrane stresses will cause plastic collapse.* Calculate the crack opening areas and leak rates through the cracks. Evaluate the leak rates to determine if they will set off the proposed RWCU room temperature monitors.

2) Calculating whether unstable fracture is possible in the pipe with the largest length-to-radius (L/R) ratio.

Axial pipe flaws are not considered because 1) the seamless pipe has no axial length weld that can be sensitized and become susceptible to intergranular stress corrosion (IGSCC),

so axial-length cracks cannot occur; and 2) the only place that pipe is susceptible to axial cracks due to IGSCC is at the circumferential welds. These axial cracks will probably not exceed the pipe weld width plus twice the heat affected zone ( HAZ) width. For the pipe sizes considered herein, i.e., up to 10-in. diameters, that figure is not more that about 1.4 in., which will result in a 0.6 gpm leak.

LBB-1 nu eM

~

3.0- METHODOLOGY The steps in calculating the circumferential through wall crack size necessary for plastic collapse are described below.

1) Determine the critical crack length for 4 ,6 ,8 ,

and 10-in. diameter piping base material and weld material. Pipes with cracks longer than these are postulated to f ail catastrophically.

2) Determine the crack opening area (COA) associated with these cracks and the leak rate associated with the COAs.

The steps are discussed in turn below, a) Determine the critical crack length. Consider a pipe with wall thickness, t, with a defect of depth, d, that is subjected to a primary membrane stress in the uncracked section of the pipe, P,, and a primary bending stress, P b. To determine the point at which collapse occurs, it is assumed that the cracked section behaves like a hinge with the stresses at og, the flow stress. The angle, 8, is the place at which stress inversion occurs. Normally 6

= 90' for uncracked pipes, but the presence of the crack causes the neutral axis to shif t away from the pipe's geometric center.

Equilibrium of the longitudinal forces gives o g(waf) - P,w 6= 2a Cl}

. f

(

i tes-1 nutech

+

Equilibrium of moments about the pipe axis yields (2 sin 8-fsina) =P 3 (2)

It should be noted that equations 1 and 2 are valid for the range 0 < a < w - 8.3 For conservatism it was assumed that the RWCU primary membrane stress due to normal operation was the same as the main recirculation system, i . e . , P, = 60 0 0 ps i . ,

The primary bending stress due to operating basis earthquake loads was assumed to be the

, highest value observed for straight pipe in the Quad Cities 2 RWCU system, i.e.,

Pb = 9500 psi.4 The flow stress, og, was estimated at 48,000 and 55,000 pai for the base and weld materials, respectively.5

~

Substituting these values into equations 1 and 2 above and solving them for 8 gives 8 = 77.5' for the base metal and 82.12' for the weld metal.

Therefore, a pipe whose behavior can be oescribed by the net section collapse criterion can sustain a 155' circumferential crack in its base metal before failing by 3

D. A. Hale, et.al., "The Growth and Stability of Stress Corrosion Cracks in Large-Diameter BWR Piping, Volume 2:

Appendices", Final Report (July 1982) , Appendix A - (EPRI Report NP-2472).

4 RWCU Stress Data, Quad Cities, EDS Nuclear, November 16, 1981.

6 J. Strosnider, U. S. Nuclear Regulatory Commission, Private Communication, January, 1984.

tes-1 nutgch

plastic collapse. For example, a 4-in.

outside diameter (OD) pipe would have a critical circumferential crack length of about 5.4 in, for the base metal.

b. Determining the crack opening area (COA) and leak rate. From Reference 4, the COA, including plasticity effects is given by:

2 A = 201 F (a/oy) p where o = Stress pulling the crack open.

(' Assumed to be P, + Pb" 15,500 psi) 1 = Crack length E = Young's Modulus (a 25 x 10 6

~

psi) a y

= Yield stress (~25,000 and 45,000 psi for the base and weld materials, respectively) i

F = Plasticity correction factor

(~1.31 and 1.08 for the base and weld materials, respectively) l For the base material in the 4-in. OD pipe, A p i

= 0.048 in.2, I

l i

The leak rate can be estimated using Moody's i data 6 for two-phase flow through a

! frictionless nozzle. At 1000 psi, Moody estimated flow at 55 lb/ sec.in.2, g t

l

'aa-t rititgt!;l)

reasonable estimate of flow through "real" cracks - i.e., cracks with some amount of friction - is about 27.5 lb/sec.in.2 at 1000 psi.7 Multiplying the crack area by the flow rate per unit area and converting the flow rate to gallons per minute (gpm) results in a calculated flow rate of 9.4 gpm at 1000 psi for the 4-in. pipe. Flow rates through critical circumferential cracks in base metal and weld metal are given in Table 1. The calculated flow rates in Table 1 are readily detected by the proposed resistance temperature detectors (RTD's) in the RWCU rooms.9 1

A second source of unstable propagation of circumferentially oriented, through-wall pipe cracks occurs when the pipe is subjected to internal pressure and fixed end rotations.

This analysis assumes that the material is an elastic-perfectly plastic material with large deformations, that the remaining uncracked pipe section is fully yielded to limit load, and that the pipe wall is thin (i.e., the pipe wall-to-radius ratio is much less than 1.0).9 4

6 F. J. Moody, " Maximum Two-Phase vessel Blowdown from Pipes", General Electric Company, April 1965 (APED-4827).

7 W. J. Shack, Argonne National Laboratory, Personal gommunication, January 1984.

J. C. Hink to E. R. Zebus letter, November 18, 1982.

9 M. Mayfield et.al., " Cold Leg Integrity Evaluation",

NU REG /CR-1319, U. S. Department of Commerce, Washington, D.C., February 1980.

LBB-1 riutgch

The analysis involves calculating a dimensionless quantity, the applied tearing modulus, T,ppy. This is compared to the material tearing modulus, T mat, which is calculated from fracture mechanics parameters and which is assumed to be a material property. If Tappi < Tmat, the material will not tear in an unstable manner. 'In the case of a through-wall circumferentially cracked pipe, this means the pipe will leak before breaking. If Tappi > Tmat, then the pipe may catastrophically f ail.

The approach requires calculating the applied J, an elastic-plastic fracture mechanics parameter. This in turn requires a stress

analysis of the piping system. However, it will be shown that the term that includes J can be neglected as a first approximation.

1 nutech

r TABLE 1 -

8' LENGTH, ARCA, AND PLOW RATE THROUGH CRITICAL CIRCUMFERENCE CRACKS IN 4 ,6 ,8- AND 10-INCH DIAMETER PIPE BASE AND WELD MATERIAL Pipe Crack Crack Flow j Diameter Length Area Rato l (in.) (in.) (in.2) ggg,)

! 4 5.4 (5.7) 0.048 (0.042) 9.4 (8.4)

~

l 6 8.1 (8.6) 0.107 (0.095) 21.2 (18.9)

)S

' 8 10.8 (11.5) 0.190 (0.169) 37.7 3

(33.6) 10 13.5 (14.3) 0.265 (0.265) 58.9 (52.5) 1

• Base material figures a e given first; weld material figures are in parenthesos.

l 3

C:

b?O r n

,

7"

~

~

The Tappl calculations from Tada 10 will be used.

T,ppi = F1x h+F 2o R

. O where = Fy = ( sin a + cos 0)

P 7 0+w2 o (2wRt)

I a=

F1= fF y .

F2= - ( cos a - 2 sin e)

J e = Half-angle of defect R = Pipe radius t = Pipe wall thickness ao = Flow stress (48,000 and 55,000 psi for base metal and weld mee.al, respectively)

P = Axial force on the pipe (equivalent to the membrane stress: P m. Thus P = P m u Rt, where P, = 6000 psi) 10 P. C. Paris and H. Tada, The Application of Fracture Proof Design Methods Using Tearing Instability Theory to Nuclear Piping Postulating Circumferential Through Wall Cracks, NUREG/CR-3464, Washington, D.C.: U. S. Nuclear Regulatory Commission, September 1980, Section 1-2 LBB-1 nutech

l 1

Substituting various values for e , ranging from 15' to the critical value of 0, results in the equations in Table 2. It should be noted that the coef ficients of J are quite small, so the second term can be neglected, and T, ppt has a simple relationship to the L/R ratio. To calculate the critical L/R ratios, the Tmat ( the highest value of T the material can withstand before unstable fracture is expected) must be known for the materials of interest. *imat data for austenitic stainless steel base material, and gas-tungsten arc and shielded metal arc welds (GTAW and SMAW, respectively) are given in Table 3.

Tmat sets a limit for the L/R ratio. If Tmat for the 4-inch pipe base metal is, say 150, then the maximum value that T app y can reach before _ instability occurs is 150. Using Table 2, the largest crack that a pipe with a membrane stress of 6000 psi and a bending stress of 9500 psi can withstand is 155*

(i.e., 2 x 77. 5*). For 0 = 77.5*,

T,ppi n 0.615 L/R.

For Tappl d 150, L/R f 244. This L/R value (244) is the largest L/R ratio the pipe system can have without failing under the conditions enumerated above.

In another example, consider a piping system l with L/R = 150 and base metal Tmat = 150.

Inspection of Table 2 indicates that for cracks with e = 15' through 45', Tappl >

150. That means that the crack will grow LBB-1 ilutech

I TABLE 2 T appl FOR VARIOUS VALUES OF 0 9 BASE METAL WELD METAL i

~4 * ~4 15' T, =1.055h+9.41x10 J T,ppy=1.016h+7.43x10 J 25* T, y=1.091h+1.51x10 ~4 J T,ppy=1.053h+1.33x10 ~4 J

~4 30' T,ppy=1.090h-2.14x10 J T,ppy=1.052h-3.97x10 ~4 J

.45* T,ppy=1.015h-1.30x10 J T,ppy=0.982h-2.45x10 -3 g

~

60 * - T,ppy=0.856h-2.39x10 J T,ppy=0.828f-1.82x10 -3 J

~

77.5* T,ppy=0.615h-3.76x10 J 82.l* T,ppy=0.530h-3.18x10-3,

• L is the pipe's length; R is the pipe radius.

LBB-1 gg

l unstably. However, it will stop after Tappy decreases below 150 - for e > 47*. This anamolous behavior occurs because the pipe

{

behavior is displacement-controlled, i.e., the stress on the crack tip depends on the pipe displacement and not the loads. Obviously, if l Tmat were greater than about 170, no unstable growth would occur between 0 = 15 and 45*.

For a 4-inch pipe with L/R;< 266, the T mat Of 280 will assure crack stability for all crack lengths up to e = 82.l* in the shielded metal arc weld. For crack lengths greater than 0 =

82.l', the membrane and bending stresses will cause rupture.

The run of high energy pipe with the largest L/R is approximately 38 feet of 4-in. diameter

. pipe in the Quad Cities RWCU System. The L/R

, ratio for this pipe is 228. The worst case pipe run in the Dresden RWCU system is approximately 46 ft. of 6-in. diameter pipe, with L/R of 184.

l l

l l

nutgch lab-1 - - - - - - - -

TABLE 3 Test Material Temperature Tmat (L/R) critical Base Metal 10 550*F 623 571 Gas Tungsten Arc Weld 550*F 432 410 (GTAW)l0 Shielded Metal Arc Weld 550*F 280 266 (SMAW)11 4

10 Mir.utes of ASME Section XI Task Group on Pipe Flaw Evaluation, Special Meeting, March 6, 1984, Attachment 3.

11 W. J. Mills, Hanford Engineering Development Laboratory, Personal Communication, April, 1984.

Lan-1 nut.ech

. o 4.0 DISCUSSION AND CONCLUSIONS The RWCU piping has been studied from a fracture mechanics standpoint to determine whether a circumferential through wall crack will leak before breaking. Two scenarios were considered: 1) failure by the membrane and bending stresses acting on a large circumferential through-wall defect and 2) failure by

• tearing instability due to the membrane stress and fixed-end rotations. It was seen that the calculated

leak rates for the critical crack sizes for the first failure mechanism are much larger than necessary to trigger the proposed room temperature sensors. '

Therefore, circumferential cracks will never get to critical size, as the temperature monitor will alert the reactor operator to isolate the RWCU system long befor.e the cracks attain critical length.

4 Failure by the second mechanism is not considered possible because the critical L/R ratios are larger than the largest L/R ratio for the piping under consideration.

Therefort, it is concluded chat the proposed leak detection system is adequate because catastrophic pipe failure is not expected and the temperature monitors will alert the reactor operator (s) to leaking pipes long before the leaks attain critical size.

LBB-1 nutgch

_ - . _ _ _ _ _ _ _ _ - - _ . -. - - - - . -- -.