Text

Rev 0 to SIR-99-032, Flaw Tolerance & Leakage Evaluation Spent Fuel Pool Heat Exchanger E-53B Nozzle Palisades Nuclear Plant.
	ML18066A467
Person / Time
Site:	Palisades
Issue date:	03/31/1999
From:	Cofie N, Deardorff A, Hirschberg P; STRUCTURAL INTEGRITY ASSOCIATES, INC.
To:	;
Shared Package
ML18066A466	List: ML18066A465; ML18066A467;
References
	SIR-99-032, SIR-99-032-R00, SIR-99-32, SIR-99-32-R, NUDOCS 9905250080
	Download: ML18066A467 (36)
	v • d • e

ATTACHMENT 2 CONSUMERS ENERGY COMPANY PALISADES PLANT DOCKET 50-255 INSERVICE INSPECTION PROGRAM -SUBMITTAL OF RELIEF REQUEST NO. 14 FOR NRC APPROVAL Structural Integrity Associates Report SIR-99-032, Rev 0 Flaw Tolerance and Leakage Evaluation Spent Fuel Pool Heat Exchanger #E-538 Nozzle Palisades Nuclear Plant 35 Pages 9905250080 990517 l PDR ADOCK 05000255...

G PDR ..-

~~

I*

I Report No.: SIR-99-032 Revision No.: 0

-1 ' ProjectNo.: CPC-lOQ File No.: CPC-lOQ-401 March 1999 I

I I Flaw Tolerance and Leakage Evaluation Spent Fuel Pool Heat Exchanger E-53B Nozzle I Palisades Nuclear Plant I

I Preparedfor:

I Consumers Energy (Contract No. C0030239)

I Prepared by:

I Structural Integrity Assodates, Inc.

San Jose, California I

1* Prepared by: Date: 3

2 ~-9' P. Hirschberg, P .E.

I Reviewed by:

I I Approved by:

N. G. Cofie, Ph.D.

Date: 3/L-o/? 7 I

I

~*

1* e.

I REVISION CONTROL SHEET

_I Document Number:

SIR.-99-032, Rev. 0

Title:

Flaw Tolerance and Leakage Evaluation, Spent Fuel Pool Heat Exchanger E-53B I Nozzle Palisades Nuclear Plant I Client: Consumers Energy SI Project Number: CPC-100 1* Section Pages Revision Date Comments 0 3/29/99 Initial Issue I

I I

I

1*

I Table of Contents Section I_

1.0 IN'TRODUCTION ................................................................................ ~ ............................... 1-1 I

2.0 DESCRIPTION

OF HEAT EXCHANGER 53B GEOMETRY, MATERIALS AND LOADS ................................................ -................................................................................. 2-1 1-. 3.0 STRESS ANALYSIS OF HEAT EXCHANGER NOZZLE .................. ~ ............................ 3-1 I 4.0 ESTIMATION OF LEAKAGE AND SIZE OF LEAKAGE CRACKS ............................ .4-1 4.1 Leak.age Rate Determination .......................................................................... :.............-.- ... 4-1 4.2 Determination of Potential Through-wall Crack Length ................................................ .4-3 I 4.2.1 Leak.age from a Crack Parallel to the Weld ............................................................. 4-3 4.2.2 Leak.age from a Crack Perpendicular to the Weld ................................................... 4-4 I 5.0 DETERMIN'ATION OF CRITICAL THROUGH-WALL CRACK LENGTH .................. 5-1 6.0 EVALUATION OF POTENTIAL CRACK GROWTH ..................................................... 6:-1 I 7.0 STRESS ANALYSIS OF THE CRACKED NOZZLE ............................ :.......................... 7-1 8.0 EVALUATION OF MARGIN'S ............................................................. :............................ 8-1 I

9.0 CONCLUSION

S ........................*........................................ ~ ............................................... ~. 9-1 I

10.0 REFERENCES

............................................................................................................... 10-1 I

I I

SIR-99-032, Rev. 0 iii © Structural Integrity Associates, Inc.

I

1*

L I List of Tables I Table 2-1 Equipment Nozzle Load Summary .............................................................................. 2-2

I Table 3-1 Membrane Stresses Transverse to Shell-to-Nozzle Weld ............................................ 3-2 I

I I

I SIR-99-032, Rev. 0 IV © Structural Integrity Associates, Inc.

I

1* *-

I List of Figures Figure I Figure 2-1. Geometry at Heat Exchanger 53B Nozzle ................................................................ 2-3 I Figure 3-1. 3-D Finite Element Model of Heat Exchanger E-53B Head and Nozzle ................. 3-3 Figure 3-2 Nozzle-to-Shell Region Overall Stress Intensities ................................................... 3-4 I Figure 5-1. Appendix C Allowable Flaw Approach .................................................................... 5-3 Figure 7-1. Overall Stress Intensity with Through-Wall, All Around Crack ............................... 7-3 I Figure 7-2,_Nozzle Pad Stress Intensity, All Around Crack ........................................................ 7-4 Figure 7-3. Nozzle Neck Stress Intensity, All Around Crack ...................................................... 7-5 I

I I

I SIR-99-032, Rev. 0 v © Structural Integrity Associates, Inc.

I

1*

I

1.0 INTRODUCTION

I At Consumers Energy (CE) Palisades Nuclear plant, evidence of leakage has been observed at the reinforcing pad of the spent fuel pool cooling water heat exchanger E-53B inlet nozzles. At I the lower of the two vent holes (6 o'clock position), there is a residue of light-colored materials such as would exist after evaporation of water with mineral content. There is currently no I evidence of any moisture at the vent hole. Based on this, it is concluded that there may have been some leakage from a small flaw in the weld in the past, but that it has now stopped leaking, I becoming plugged due to foreign matter collecting in the leakage path.

I Structural Integrity Associates (Sn was contracted to perform an assessment of the nozzle to justify continued operation without a repair. Sin~e the operating pressure is very low and the I nozzle moments are very low, it was expected that the nozzle should be very flaw tolerant. The nozzle also has a welded reinforcing pad that is significantly thicker than the heat exchanger I shell, adding to the overall structural strength of the junction. The geometry, materials and

1 loadings are described in Section 2.0 Thus, several evaluations have been conducted as described in this report:

I A detaile.d finite element analysis of the shell-to-nozzle intersection has been conducted.

I Pressure, dead weight, and seismic loadings were considered in the analysis. Stresses were predicted for the uncracked nozzle which were used as input to subsequent I calculations. (See Section 3.0.)

I Based on the evidence of past leakage, the maximum possible leakage from the vent holes was calculated based on conservative assumptions of temperature and natural I convection. The size of the nozzle-to-shell weld cracks that would result in the*

calculated leakage rates were determined. (See Section 4.0.)

I I

I SIR-99-032, Rev. 0 ) 1-1 © Structural Integrity Associates, Inc.

I

I*

I The critical size of through-wall cracks in the nozzle-to-shell weld was determined using ASME Section XI, Appendix C net section collapse methods, modified to account for a I shear failure collapse mode. (See Section 5.0.)

I Based on the maximum loadings, fatigue crack growth analy~is was performed to*

I demonstrate that sub-criticcl cracks wuuid not grow due to application of seismic and pressure cycling. In, addition the potential for IGSCC crack growth was evaluated. (See Section 6.0.)

I An additional finite element analysis was done assuming that the through-wall flaw had I grown completely around the circumference of the pipe, to ev~uate the capability of the

-I reinforcing pad to maintain the structural integrity of the nozzle to shell junction. (See Section 7.0.)

I _,

Using the results above, a comparison of the leakage size cracks and critical size cracks I was made to demonstrate that there is ample margin between the two. (See Section 8.0.)

I Conclusions are provided in Section 9.0, showing that there is more than a factor of two between the crack size that meets ASME Section XI structural acceptance criteria and the sizes that would I be necessary to pass the conservatively-estimated leakage rates. This includes the safety margins inherent in Section XI for normal/upset conditions even when the SSE seismic loadings are I applied.

I I

I SIR-99-032, Rev. 0 1-2 © Structural Integrity Associates, Inc.

I

I*

I

2.0 DESCRIPTION

OF HEAT EXCHANGER 53B GEOMETRY, MATERIALS AND LOADS I *The spent fuel pool heat exchanger is a shell and tube heat exchanger [1]. The tube-side water I communicates with the spent fuel pool. At each end of the heat exchanger, there are heads with nozzles connecting to the spent fuel pool cooling piping.

I The heads are made of SA-240, Type 304 stainless steel [l]. The cylindrical head has an inside I diameter of 25 inches and a wall thickness of Yt6 inches. The horizontal nozzle, from the side of the cylindrical head, is SA-312, Type 304 made from Schedule lOS pipe (12.75-inch OD and I 0.18-inch thick wall). The reinforcing pad around the nozzle is SA-240, Type 304. The reinforcing pad has a thickness of 0.25 inches and extends 2.5 inches radially beyond the outer I radius of the nozzle, welded to the nozzle. and shell with a ~-inch structural fillet weld [2]. The .

nozzle, with a SA-182 F304 flange, extends 7 .5 inches beyond the ID of the cylindrical head.

I Figure 2-1 shows the local geometry at the nozzle.

I The reinforcing pad has two Ys-inch vent holes located at 6 and 12 o'clock on the nozzle [2, 3].

The evidence of leakage from the vent hole is evident only at the bottom vent hole.

I The heat exchanger has a design pressure of 125 psig and design temperature of 150°F. The I maximum expected inlet/outlet temperatures of the spent fuel cooling water are 120-130°F inlet and l 10°F outlet [4]. The nozzle loadings [5], are shown in Table 2-1. Thermal expansion I moments were considered to be insignificant in the piping analysis.

I I

I SIR-99-032, Rev. 0 2-1 © Structural Integrity Associates, Inc.

I

I*

I Table 2-1 Equipment Nozzle Load Summary I

Equipment:* Heat Exchanger E-53B I Analysis Node Pt.: 523 I Load Load or Forces (lbs.) Moments (ft.-lbs.)

Case Load Combination

Fx Fy Fz Mx My Mz I 10 Dead Weight (DW) -23 -314 -54 -336 143 -26 35 OBE SAM (OSAM) 0 0 0 0 0 0 I 30 , Seismic OBE (OBE) 103 171 61 378 255 208 45 SSE SAM (SSAM) . 0 0 0 0 0 0 I 40 Seismic SSE (SSE) 205 343 122 756 509 415 I X Y

= Horizontal

= Vertical (transverse to nozzle)

Z = Axial to nozzle I

I I

SIR-99-032, Rev. 0 2-2 © Structural Integrity Associates, Inc.

I

I I

R = 12.5" I

I

- -- -- -- -- -- -- -- -- -12.39" I

I Suspect Through Weld Leak.

I Evidence of Leakage from 3/e" vent hole I

I I NOTTO SCALE I

I Figure 2-1. Geometry at Heat Exchanger 53B Nozzle I

I I

SIR-99-032, Rev. 0 2-3 © Structural Integrity Associates, Inc.

I

I' I 3.0 STRESS ANALYSIS OF HEAT EXCHANGER NOZZLE I To evaluate the stresses in the vicinity of the nozzle, a three dimensional finite element stress analysis was conducted using the ANSYS computer program [6]. The 3-D model was used so I

that the non-axisymmetric loading effects of the cylindrical nozzle-to-shell joint could be evaluated and the effects of the nozzle moments could be assessed. Figure 3-1 shows the model.

I The stress analysis was conducted using ANSYS version 5.3, SOLID-45 type 3-D structural solid elements, having 8 nodes and 3 translational degrees of freedom at each node. 3-D elastic I beam (BEAM4) elements were used to transfer axial, shear, and moment loads to the face of the qozzle.

I The finite eleil1ent model included the heat exchanger shell, end cap, nozzle neck, nozzle I reinforcing pad, and the two structural fillet welds connecting the reinforcing pad to the nozzle, neck and shell. The model assumed a fully intact double-v weld joining the nozzle neck and I shell.

I The loads applied. to the heat exchanger vessel and nozzle were taken from [5] and [7]. The design pressure of 125 psi was applied, which represents normal operating conditions. The I mechanical and seismic loads from the connected piping system shown in Table 2-1 were used.

Material properties were taken from the ASME Code Appendices [8], evaluated at 150°F.

I An analysis was run to evaluate the stresses in the unflawed component. The purpose of this I evaluation was fo determine the stresses in the vicinity of the shell-to-nozzle weld location.

I These stresses ate important from the standpoint ofopening potential cracks that would result in leakage during normal operation. Table 3-1 shows the computed membrane stresses due to I pressure for a quarter section at the weld location. The stresses in the circumferential direction at the double-v weld are an order of magnitude higher than the radial direction due to the geometry I of the combined nozzle/pad/shell model. Figure 3-2 shows the overall stress contour results for all loads.

I The details of the finite element analysis are included in Reference [21].

I SIR-99-032, Rev. 0 3-1 © Structural Integrity Associates, Inc.

I

I*

I Table 3-1 Membrane Stresses in Nozzle-to-Shell Weld I

Stress in Nozzle-to.:.Shell Weld, ksi (across crac~)

I Angle on Nozzle (from horizontal) Crack Parallel to Weld Crack Perpendicular to Weld 0 0.896 18.42 I 15 1.084 17.33 30 1.393 14.42 I 45 1.367 10.48 6.38 I 60 75 0.908 0.420 3.11 I 90 0.217 2.02 I

I I

I SIR-99-032, Rev. 0 3-2 © Structural Integrity Associates, Inc.

I

I I

I CPC-lOQ I*

I I Figure 3-1. 3-D Finite Element Model of Heat Exchanger E-53B Head and Nozzle I

I I

SIR-99-032, Rev. 0 3-3 © Structural Integrity Associates, Inc.

I I

I AN SYS 5.3 MAR 10 1999 09:07:18 I NODAL SOLUT ION STEP=l SUB =l TIME=l I SINT OMX =.02447 2 (AVG )

SMN =480 . 14 9 SMX =44542 I SMXB=54267 480.149 5376 10272 I 15168 20063 24959 29855 I 4 li!l!!fll 34751 39647 44542 I

I I

I I Figure 3-2. Nozzle-to-Shell Region Overall Stress Intensities I

I I

SIR-99-032, Rev . 0 3-4 @ Structural Integrity Associates, Inc.

I

I 4.0 ESTIMATION OF LEAKAGE AND SIZE OF LEAKAGE CRACKS I To determine the potential size of the crack that produced the observed leakage at the vent hole, an estimate of the leakage was developed. Then using weld stresses developed from the finite I element analysis, crack opening areas and the associated leakage rates ~ere developed.

I 4.1 Leakage Rate Determination 1* The evidence of leakage at the reinforcing pad vent hole is from the presence of a white residue left from the evaporation of water sometime in the past. There is currently no evidence of I moisture in the vent hole. There is no evidence of leakage from the upper vent hole~

I To estimate the amount of leakage, a mass transfer calculation was conducted assuming that the bottom and sides of the vent hole were w~t (100 percent humidity at the surfaces) and that.the air I

j .

surrounding the heatexchanger was dry. To ma:Ximize the predicted evaporation rate, it was conservatively assumed that the heat exchanger shell was at 130°F. This high temperature I maximizes the vapor pressure and mass concentration of water at the wet surfaces. The diffusion coefficient for water vapor in air was taken at this same temperature and was J.1162 ft 2/hr, based I on correcting basic data for 32°F [9]. Performing a pure diffusion 'inass transfer analysis (analogous to a heat transfer conduction evaluation), a leakage. rate of 13,88 lb/year was I determined.

I . .

However, the effects of local air currents and density differences can affect the rates of heat transfer and mass transfer at a surface. To determine_ this effect, the natural convection he~t I transfer coefficient at the heat exchanger was calculated [ 10]:

I h = 0.28 (8T/L) 0*25 = 0.7446 Btu/hr-ft2-°F I where I 8T L

=

surface to ambient temperature difference = 50°F characteristic length= 1 foot (assumed)

I SIR-99-032, Rev. 0 4-1 © Structural Integrity Associates, Inc.

I

I*

I The Nusselt number is art estimate of the natural convection effect compared to a pure- -

I conduction effect. Based on the evaluation above, I Nu = hUk =46.5 I where k = air thermal conductivity= 0.016 Btu/hr-ft-°F at 100°F I

Assuming the increase* in mass transfer could be analogous*tothe increase in heat transfer, as I

indicated by the Nusselt number, the diffusion mass transfer determined above was

. conservatively increased by a factor of 50 to arrive as an upper bound estimate of the leakage I rate of 694 lb/year (0.000165 gpm)~

I This is believed to be a very conservative upper bound because 1) the one foot dimension used in I the Nusselt Number* correlation is large compared to the size of the vent hole, 2) the L\T used is conservative, 3) the local geometry of the vent hole would tend to mitigate the natural convection I effects, and 4) it is assumed that there is no humidity in the room. In reality, it would be .

expected that leakage of this amount would lead to some evidence of wetness or drop formation I around the vent hole.

I Considering the amount of liquid leaked in one: minute, the hemispherical drop size for this leakage rate was estimated to be approximately 0.5 inches. Thus, this is a very conservative I estimate of the leakage rate. If this amount of w*ater were leaking, there would definitely _be clear evidence of water leakage from the vent hoie, since the water could not evaporate into the air this*

I rapidly. However, use of this leakage rate is conservative for estimating the size of potential flaws:

I I

I SIR-99-032, Rev. 0 4-2 @ Structural Integrity Associates, Inc.

I

I 4.2 Determination of Potential Through-wall Crack Length I To calculate the crack size in the heat exchanger shell-to-nozzle weld, the SI program pc-LEAK

[ 11] was used. This program is based on the concepts of linear elastic fracture mechanics for I calculating crack opening area [12, 13, 14]. Fundamental fluid mechaniCs methods are used to calculate leakage of water through the crack based upon the crack opening displacement, surface I roughness and discontinuity losses (with more complex methods being available for calculation of two-phase flow). This program has been qualified under Si's Quality Assurance Program.

I The fluid conditions were taken as 120°F (from the assumptions above} and 125 psig. For I computing leakage, the membrane stresses across the crack were used consistent with the assumption in the fracture mechanics models, The distribution of the membrane stresses around I the shell-to-nozzle weld is shown in Table 3-1.

I The calculation was conducted for leakage from a pipe the size of the heat exchanger shell. In the analysis, there was no correction for the plastic zone at t.he crack tip so that minimum leakage I would be calculated, to maximize the co'mputed crack size. A surface roughness of 0.02 inches.

was assumed, but this had no effect on the resulting calculations, since the flow was in the I laminar flow regime for the very small leakage rates. An entrance loss coefficient K of 2.7 was 2

assumed, simulating a sharp-edged entrance to an orifice (C = 0.6 and K = 1/C This also had I

).

very little effect on the results since the pressure drop was mainly.due to friction in very tight cracks ..

I Two analyses were conducted to evaluate the leakage as described in the following subsections.

I 4.2.1 Leakage from a Crack Parallel to the Weld For this analysis, the crack was assumed to be through-wall and extend along the weld. To I simulate the additional stiffness of the nozzle side of the crack, the modulus of elasticity was I doubled, having the effect of only one side of the crack opening. Cases were run for a crack length up to 8 inches assuming a range of stresses.

I

SJR.-99-032, Rev. 0 4-3 © Structural Integrity Associates, Inc.

I

I The cakulated maximum crack size (2a) to cause a leak of 0.000165 gpm was approximately4.5 I inches (28:::::: 45°) assuming that the crack was over the most lowly stressed region. For the case of a crack located at the most highly stressed region, the crack length was reduced to about 2 I inches.

I 4.2.2 Leakage from a Crack Perpendicular to the Weld.

I For this analysis, the crack was assumed to be transverse to the weld. This type of crack is assumed because of the relatively large membrane stresses in the weld region in the hoop I direction of the pipe penetration. The leakage calculation was conducted using a model for a longitudinal crack in a pipe.

I The predicted crack size to produce 0.000165 gpm ranges was about 0.18 inches (2a),

I approximately equal to the shell thickness,

. for the most highly stressed location.

Lower stresses were present at the other locations, which would require larger cracks for the same amount of I leakage. Larger cracks were not evaluated. since pre-existing cracks in the base material were not considered to be credible.

I Details of these calculations are given in Reference [22].

I I

I SIR-99-032, Rev. 0 4-4 @ Structural Integrity Associates, Inc.

I

I 5.0 DETERMINATION OF CRITICAL THROUGH-WALL CRACK LENGTH I An evaluation was done to determine the length of a through-wall crack circumferentially around the nozzle that would still be able to withstand the applied loads without becoming unstable.

I The calculation used the limit load approach based on net section plastic collapse. The ASME Boiler and Pressure Vessel Code,Section XI, Appendix C [15], provides rules for evaluating I circumferential flaws in austenitic piping. In Subsection C-3300, equations are given for determining the maximum allowable flaw length and depth for a given load, or vice versa. The I equations determine the maximum load carrying capability of the remaining ligament, which occurs when the cross section of concern reaches fully plastic action limit load. Using the design I conditions and external piping forces and moments provided [5, 7], the maximum allowable flaw size for the applied loads was determined.

I The following assumptions were used in this calculation, based on [3]:

I

Thermal expansion loads were considered to be negligible, per Palisades Specification M-I 195 [24], for systems at temperatures of 150°F or less. A review of the piping isometrics showed that there is sufficient flexibility in the piping routing and support system for this I assumption to be valid.

I ** .

The seismic loads on the nozzle of the attached piping system were considered, but the self-inertia loads of the heat exchanger were not. The heat exchanger is classified as Cla5s ill I equipment, for which seismic qualification is not required. Of interest here was* whether the I nozzle-to-shell weld joint would maintain its integrity if the attached piping were to be subjected to seismic excitation.

I

The Safe Shutdown Earthquake (SSE) loads were used with the normal operating condition I safety factor. This is conservative because the SSE loads are higher than the normal (upset) condition (QBE) loads, with a lower probability of occurrence.

I I

SIR-99-032, Rev. Q 5-1 © Structural Integrity Associates, Inc.

I

Shielded Metal Arc Welding (SMAW) was assumed to be the welding process used. This

results in more conservative Z factors (see' below) than if the weld process is known to be I Gas Tungsten Arc welding (GTAW).

I

The external piping applied lateral forces were given at the flange joint, approximately 7.5 inches from the shell. In order to apply them at the shell interface, additional moments, I equivalent to the force times the distance from the flange to the shell, were applied.

I

The load carrying capability of the reinforcing pad was conservatively neglected in this evaluation.

I I

The flow stress in shear is assumed to be half the flow stress in tension.

I

Piping moments were combined by SRSS per ANSIB31.1, as required by [24].

I In the limit load approach, the applied forces and moments are compared with the plastic load-carrying capacity of the cross-section containing the flaw. The assumption is that the material I will form a plastic hinge. i.e. the entire cross section will deform plastically before the flaw grows to an unstable size. This assumption is valid for austenitic materials and non-flux weld I metal. The remaining unflawed ligament must be able to accommodate the applied primary membrane and bending stresses in order to be acceptable. Safety factors are applied on .the

. l I combined membrane and bending stresses to arrive at allowable loads. Relationships are given in ASME Section XI Appendix C between allowable applied stress and unflawed ligament size.

I In this calculation, the loads were known but the flaw size was not. The maximum allowable I through-wall flaw length was determined for the applied loads. The equations given in Appendix C which relate applied ~oad to unflawed ligament at plastic collapse are the following:

I I

I SIR-99-032, Rev. 0 5-2 @ Structural Integrity Associates, Inc.

I

~-

I C-3320(a) Equation (3)

I P,,' =

6

~ m ( 2 sin f3 - ~ sin (})

I and I {3 = ! (Tr - !!:_ (} -

2 . t .

Tr pm )

3Sm I where:

Sm = material stress allowable I Pm = piping membrane stress a = flaw depth I t = section thickness

~=angle between vertical and neutral axis (see Figure 5-1)

I 8 = half angle of circumferential flaw length I

I I

I I tl11St1 Figure 5-1. Appendix C Allowable Flaw Approach I

I SIR.-99-032, Rev. 0 5-3 @ Structural Integrity Associates, Inc.

I

I*

I Although the equations shown above were developed to determine the allowable stresses in piping with through-wall flaws, they are applicable to evaluating flaws in a weld between a pipe I and a shell, if the allowable flow stress (3Sm ) is reduced by a factor of two.

I To determine the allowable flaw length, safety factors. were applied. In addition, for flux welds (SMAW), an additional Z 1 factor was applied, to /account for the fact that the weld metal is less I ductile and may not reach fully plastic action before the flaw becomes unstable. Since the welding process was not known, SMAW was conservatively assumed. The Code equation is the I following:

I C-3320(c) Equation (6)

I I

where:

SF= safety factor= 2.77 for normal operating conditions I Pb = piping bending stress Z1 = 1.15 [1 + .013 (D0 - 4)]

I D 0 = pipe outside diameter I Pb' = limit bending stress Pe*= piping expansion .stress I The evaluation used the highest seismic loads, the SSE case. Using a Code safety factor of 2.77 I was conservative because it is intended for normal operating conditions, while the SSE seismic loads are faulted conditions. A safety factor of 1.39 is allowed for faulted conditions.

I It should be noted that the quantity 6Sm in the first equation represents twice the flow stress (and I 3Sm in the second equation is the flow stress). In this calculation, the section with the flaw is being loaded in shear. The limiting stress of interest is the shear stress on the weld cross-section.

I An allowable flow stress of half the flow stress under tension, or 1.5 Sm , was therefore used.

I SIR-99-032, Rev. 0 5-4 © Structural Integrity Associates, Inc.

I

II I Thus, the quantity 6Sm in the first equation becomes 3Sm; the quantity 3Sm in the second equation becomes 1.5 Sm. J -

I Results for ultrasonic testing were available which showed that the. shell thickness was larger I than nominal [16], but the nominal values were conservatively used. Sm values for the base .

metal were conservatively used for the weld metal.

I The result of the evaluation was that angle e =73.9 degrees. This represents the allowable half I angle for a through-wall crack in the weld circumferentially around the pipe that meets normal Code allowable stresses. The total crack length is double this, about 148 degrees, or 16.4 inches.

I Thus, the result of this calculation was that if the nozzle to shell double-v weld were to contain a I through-wall circumferential flaw extending approximately 148 degrees around the*

circumference, it would stili'be able to withstand the applied loads. This result conservatively I takes no credit for the reinforcing pad, and includes the Section XI factor of safety for I normal/upset loads.

I Another case was run applying a factor of ..Y2 to the applied stresses on the flaw, consistent with the philosophy of the Leak Before Break methodology specified by the NRC [ 17]. The result I was that even with a factor of safety of ..Y2 on stress, a flaw length of about 119 degrees around the circumference, or 13.25 inches, would be acceptable.

I Details of these calculations are given in Reference [23].

I I

I SIR-99-032, Rev. 0 5-5 © Structural Integrity Associates, Inc.

I

I*

I 6.0 EVALUATION OF POTENTIAL CRACK GROWTH I An evaluation was done to determine how much the existing flaw could potentially grow over time under the applied loads. ASME Section XI, Appendix C, Subsection C-3200 [15] gives I rules for determining flaw growth in austenitic materials. There are two mechanisms for growing flaws in austenitic piping - stress corrosion cracking, and fatigue cycling. Reference I [ 18] indicates that IGSCC is not a concern in lines where the normal operating temperature is less than 200° F. As the spent fuel pool heat exchanger inlet lines operate at 150°F or less, I IGSCC is not considered to have the potential to cause the flaw in the nozzle weld to grow.

I As for fatigue cycling, there are no significant cyclic loading conditions at the nozzle. The spent fuel pool pumps are judged to be too far away to cause vibration at the nozzle. Over the next I

several years, the number of pressure cycles from system starts will not be significant. The main cyclic loading' would be the seismic loading from the attached piping, should a seismic event I occu.r. It will be conservatively assumed that the seismic loads will produce the equivalent of 50 stress cycles.

I The EPRI Ductile Fracture Handbook [19].gives an equation for calculating the ~tress intensity I factor for a through-wall flaw in a cylinder under tensile loading:

I I where:

I O"t = axial stress Ft= 1 +A [5.3303(8/7t)r. 5 + 18.773(8/7t)4 *24]

I A= [0.4(R/t) - 3.0] 0.25 8 = defined in Figure 5-1 for 10 ~ R/t ~ 20 I R = outside radius of cylinder I This model is intended for a flaw in the nozzle pipe; however, there is no model available that exactly matches the geometry being evaluated, and it is judged that this model provides a good I

SIR-99-032, Rev. 0 6-1 @ Structural Integrity Associates, Inc.

I

I*

I approximation. Although this formulation is not strictly applicable for R/t =33 of this case, it is sufficiently- accurate for the purposes -of demonstrating that crack growth would be small. Since I this equation assumes that the loading is tensile, the shear stresses calculated above were conservatively doubled. The pressure and dead load stresses were separated out from the Pm and I Pb terms and treated as mean stresses, and the seismic (SSE) stress was considered the varying stress.

I Figure C-3210-1 of ASME Section XI gives reference curves that determine the rate of crack I growth per load cycle, da/dN, as a function of Af( and Kmin1Kmax. where K is the crack tip stress intensity. This subparagraph also gives equations which describe the curves, including I accounting for temperature effects. Since these curves are for air environments, the rates were multiplied by a factor of 2 to account for the difference between water and air environments [20] ..

I These equations were used to calculate the amount of flaw growth that would take place in an SSE event.

I From Section 4.2. l above, tl;ie estimated maximum length of a circumferenti~ crack to produce I the observed leakage was determined to be 4.5 inches. Thi~ very conservative length was used in I calculating the*crack tip stress intensity factor. The result was that for the maximum possible crack length, da/dN is 1.03 E-5 in/cycle. For 50 seismic cycles, the crack would grow 5.15 E-4 I inches, which is an insignificant amount.

I A load case was also run assuming the crack would not be repaired. It was conservatively assumed that there would be 50 pressure startup cycles along with the 50 seismic cycles.

I For the case of pressure plus seismic cycling, daJdN was 1.43 E-4 in./cycle. For 50 cycles, the crack would grow .007 inches from an initial length of 4.50 inches, or about 0.1 %, which is also .

I considered to be insignificant. Thus, it was concluded that the length of the flaw will essentially not change at all.

I Details of these calculations are given in Reference [23].

I I

SIR-99-032, Rev. 0 6-2 © Structural Integrity Associates, Inc.

I

I 7.0 STRESS ANALYSIS OF THE CRACKED NOZZLE I As a very conservative way to evaluate the effect of a flaw on the stress distribution in the nozzle attachment, a finite element analysis model was developed assuming the through-wall flaw has I grown completely around the pipe. It was desired to determine whether the reinforcing pad and its structural fillet welds would be able to maintain the structural integrity of the heat exchanger I vessel and the attached piping.

I To evaluate the acceptability of the stresses, the stress allowables given in ASME Section ill, Subsection ND-3300 [8] were used. Table ND 3321-1 indicates that the allowable stresses for I membrane and membrane plus bending are 2.0 Sand 2.4 S, respectively, where Sis the allowable stress value of the material at temperature. The Level D allowables were used for this I case because we are only interested in demonstrating that the reinforcing pad welds will allow the system to maintain structural integrity, under the assumption that the nozzle to shell weld I were to have a crack that extends all the way around the pipe. Although the vessel is designed to

" - ' I Section ill, 1966, this edition does not provide allowables that are applicable for this case. As I this is not a true design condition, the 1989 rules were used as a reasonable acceptance criteria.

I For the nozzle neck, three through-wall stress sections were evaluated near the reinforcing pad.

The first was taken at the location of maximum stress intensity in the nozzle reinforcing pad and I is designated as Path A. This location corresponds to the toe of the weld on the nozzle neck.

The stress allowables considered were membrane plus bending. The second was taken at the I maximum stress intensity location of the nozzle neck and is designated as Path B. This is the through-wall path b~fore the toe of the weld. The third through-wall path for the nozzle neck I was taken from the root of the weld. It is designated as Path C.

I Figure 7-1 shows the overall stress intensity values for the case of the through-wall, all around I crack. The stress intensity results for the nozzle reinforcing pad and nozzle neck are shown in Figures 7-2 and 7-3, respectively.

I The results are summarized in the table below:

I SIR-99-032, Rev. 0 7-1 @ Structural Integrity Associates, Inc.

I

I;, ... .,

I Path Membrane (ksi)

With Flaw No Flaw Mem + Bending (ksi) Allowable Stress (ksi)

With Flaw Without Flaw Membrane Mem + Bendinl?

A 23.9 20.0 37.4 25.3 36.6 43.92 I B c

23.4 25.7 18.8 21.7 37.1 28.l 30.0 24.1 36.6 36.6 43.92 43.92 I

For the structural fillet weld attaching the refoforcing pad, one path through the wall section was I taken along the base of the weld. The path begins at the toe of the weld on the nozzle neck and

. ends at the root of the weld. The stress allowable considered was membrane plus bending. This I section is designated as Path D.

I Path Membrane (ksi) Mem +.Bending (ksi) Allowable Stress (ksi)

With Flaw I No Flaw With Flaw I Without Flaw Membrane I Mem + Bending I D 23.3 I 19.6 28.9 I 24.1 36.6 I 43.92 I For the reinforcing pad, one linearized stress path was taken through the thickness of the pad.

I The path begins at the toe of the weld on the pad and proceeds through the pad thickness. The.

stress allowable considered was the membrane plus bending. This section is designated as Path I E.

I Path Membrane (ksi)

With Flaw I No Flaw Mem + Bendini (ksi)

With Flaw Allowable Stress (ksi)

Without Flaw Membrane I Mem + Bendinl?

E 19.42 I rs.8 26.65 20.29 36.6 I 43.92 I

I It was found that the membrane and membrane plus bending stress intensity results for the with-flaw case satisfy the Level D stress limit requirements of the ASME Code. Thus, even if the I nozzle to heat exchanger shell weld were to completely rupture, the reinforcing pad and the welds attaching it to the nozzle would be able to maintain the structural integrity of the nozzle I joint.

I Details of these calculations are given in Reference [21].

I

I I

ANSYS 5.3 I MAR 4 1999 08:56:52 NODAL SOLUTION STEP=l I SUB =l TIME =l SINT (AVG)

OMX = .024461 I SMN =50.878 SMX =4520 0 SMXB=55079 I

- 50.878 5067 10084 1510 0 20117 I 25134 3015 0 I

I 4

- 3516 7 4018 3 4520 0 I

I I

CPC-lOQ I

I I Figure 7-1 . Overall Stress Intensities, All-Around Crack Model I

I

I I

I ANSYS 5 .3 MAR 3 1999 16:35:21 I NODAL SOLUSrION STEP=l SUB =l I TIME=l SINT (AVG)

DMX = . 01 7 511 SMN =1056 I

SMX =35 774 SMXB=43357 1056 4913 I

8771 12628 16486 20344 I

24201 28059 1 31916 35774 I

I I

I I CPC -lOQ I

I I Figure 7-2. Nozzle Pad Stress Intensities I

SIR-99-032, Rev . 0 7-4 @ Structural Integrity Associates, Inc.

I

I*

I I

I I ANSYS 5. 3 MAR 10 1999 17 : 20 : 01 NODAL SOLUTION I STEP=l SUB =l TIME=l SINT (AVG)

I OMX =.018589 SMN =50 . 878 SMX =38164 SMXB =49272 I - 50 . 878 4286 8520 I 12755 16990 21225 25459 I 29694 33929 38164 I

I I

I I Figure 7-3 . Nozzle Neck Stress Intensities I

I

1 1* \ll' *.,

1' IJ I 8.0 EVALUATION OF MARGINS I A comparison was made between the largest flaw that could be present, and the smallest flaw that would still meet Section XI structural criteria.

I Using very conservative assumptions for the rate of evaporation, the maximum possible leak I flow rate was determined. Applying this flow rate at the location with the least stress available to pull the crack open, a maximum possible crack length along the nozzle to shell weld of I approximately 4.5 inches was determined.

I Using a limit load approach, it was determined that a through-wall crack in the nozzle to shell double-v weld could extend 16.5 inches around the circumference, and still withstand the applied I loads without rupture. This includes incorporation of the Section XI safety factors for normal I upset conditions, and does not take credit for the resistance offered by the reinforcing pad ..

I Applying an additional '\/2 safety factor on stress, the allowable circumferential flaw length,

.1 assuming the flaw is through-wall over its entirety, is reduced to approximately 13.25 inches .

This is a factor of 3 larger than the maximum possible flaw size.

I The calculated crack growth, assuming 50 cycles of both seismic SSE loads and pressure, was less than 1% .

I I The observed flow rate through the crack has been characterized as being sufficiently low such that any leakage evaporates before it can be detected. This leakage rate is probably characteristic I of a much smaller leak than has been conservatively determined by analysis. There has been no observable change in the flow rate over the past several years.

I Thus, it is concluded that there is significant margin between the probable through-wall crack I size, and the crack size that just meetsSection XI structural margins.

I I

SIR-99-032, Rev. 0 8-1 @ Structural Integrity Associates, Inc.

I

Ir **\11* .,

I

9.0 CONCLUSION

S I The flaw causing the leakage at the vent hole in the Palisades spent fuel pool heat exchanger E-53B inlet nozzle reinforcing pad was evaluated for acceptability for continued service. The I evaluations included the following analyses:

I

A finite element stre.ss analysis of the nozzle and vessel junction region determined the stresses in the unflawed nozzle for use in flaw evaluation calculations.

I

A very conservative analysis of the evaporation rate at the vent hole determined the.

I maximum leakage flow rate that could exist without showing water accumulation.

I

A calculation was done using the stresses determined in the finite element model to open a

)

through-wall crack to obtain the calculated leakage flow rate.

I

A net section plastic collapse analysis was done which determined the maximum allowable I circumferential length of a through-wall crack under the applied loads.

I

A flaw growth analysis was performed which showed that the crack would not grow significantly under both seismic and pressure cycling.

I

A finite element stress analysis was run assuming the flaw had extended completely around I the pipe. This evaluated the ability of the reinforcing pad and its welds to maintain the integrity of the nozzle to shell junction.

I The result was that the maximum potential leakage-size crack that could be present in the nozzle I weld was about one-third the allowable crack size, including the factor of safety for normal/upset loading conditions. This margin existed even when an additional .../2 factor was applied to the I stresses. The more probable existing flaw size is a pinhole sized crack that is plugged with I residue from the leakage. The analysis of the nozzle weld that was assumed to be cracked completely around showed that the reinforcing pad and its structural fillet welds were capable of I

SIR-99-032, Rev. 0 9-1 @ Structural Integrity Associates, Inc.

I

1t~n,.11 I

maintaining the structural integrity of the joint, with all stresses meeting ASME Level D

-1 allowables.

Thus, since there is no current observation of leakage and there is significant margin between the I allowable crack size and the potential size that might exist, it is concluded that continued service is acceptable.

I I

I

10.0 REFERENCES

-1* I. Industrial. Process Engineers Drawing F 5623-3 (M9-Sheet 2), "Fuel Pool Cooling Heatup Exchanger Item 53-B," Rev. H.

I 2. FAX from G. Schrader (CE) to N. Cofie (SI), "Spent Fuel Pool Heat Exchanger Inlet Nozzle Leak Repai~ Meeting Agenda," 11/30/98.

I 3. Letter from G. Shrader(CE) to P. Hirschberg (SI), "Palisades Plant Spent Fuel Heat Exchanger Nozzle," 2119/98.

I 4. Palisades FSAR Table 9-8, "Spent Fuel Pool Cooling System Design Ratings and Construction of Components," Rev. 20.

I 5. Palisades Nuclear Plant, Engineering Analysis EA-SP-03325-01, Table 4.4, "Equipment Nozzle Load Summary, Heat Exchanger E-53B, Analysis Point 523," Rev. 0.

I 6.

ANSYS Linear Plus Thermal, Version 5.3, ANSYS Inc., October 1996.

I 7. Transmittal from G. Schrader (CPC) to N. Cofie (SI), -"Spent Fuel Pool Heat Exchanger Nozzle Leakage Input Information", December4, 1998. *

I 8. ASME Boiler and Pressure Vessel Code, Section ill, 1989 Edition.

9. Rohsenow, W. M., and Choi, H, "Heat Mass and Momentum Transfer," Prentice Hall, I. Er:iglewood Cliffs, NJ, 1961.

10. Cheremisinoff, N. P., "Heat Transfer Pocket Handbook," Gulf Publishing, Houston, I Texas, 1989.

11. SI Program pc-LEAK, "<;alculation of Leakage Rates From Through-wall Cracks,"

I Version 2.0,

12. Paris & Tada, "Estimation of Stress Intensity Factors and the Crack Opening Area of a I Circumferential and Longitudinal Flaw in a Pipe," Attachment to Letter - U.S. NRC to Rochester Gas & Electric - Docket 50-244, LS05-82-091, dated February 22, 1982.

I 13. NUREG/CR-4572, "NRC Leak-Before-Break (LBB.NRC) Analysis Method for Circumferentially Through-wall Cracked Pipes under Axial Plus Bending Loads," Battelle Columbus Division for U.S. NRC, March 1986.

I 14. NUREG/CR-3464, "The Application of Fracture Proof Design Methods Using Tearing Instability Theory to Nuclear Piping Postulating Circumferential Through-wall Cracks, I Del Research for U.S. NRC," September 1983.

15. ASME Boiler and Pressure Vessel Code,Section XI, 1989 Edition.

I

I 16.

\\

Fax from G. Schrader .to P. Hirschberg, "Ultrasonic Testing Results," March 1, 1999.

. . . . . . . .. . (* . . .

17. NUREG-1061 Vol. 3, USNRC, "Evaluation of Potential for Pipe Breaks," April 1985.

I 18. Generic Letter 88-01, "NRC Position on IGSCCin Austenitic Stainless Steel Piping,"

January 25, 1988.

I 19. EPRI Ductile Fracture Handbook, NP-6301-D, June 1989.

20. ASME Section XI Task Group for Piping Flaw Evaluation, "Evaluation of Flaws in 1 I. . Austenitic Steel Piping," Journal .of Pressure Vessel Technology, Vol. 108, August 1986.

21. SI Calculation CPC-lOQ-302, '.'Finite Element Model Stress Evaluation", Rev. 0, I 3115199.

22. SI Calculation CPC-lOQ-301, "Evaluation of Leakage From Reinforcing Pad Vent I Hole", Rev. 0, 3/17/99.

23. SI Calculation CPC-lOQ-303, "Limit Load Evaluation of Nozzle Flaw, Rev. 0, 3110199.

I 24. Palisades Technical Specification M-195, Revision 5, "Technical Requirements for the Design and Analysis of SafetyRelated Piping and Instrument Tubing", 1/5/98.

I I

I

ML18066A467

Text

Title:

2.0 DESCRIPTION

9.0 CONCLUSION

10.0 REFERENCES

1.0 INTRODUCTION

2.0 DESCRIPTION

9.0 CONCLUSION

10.0 REFERENCES

Navigation menu

Search