ML20147A710

Technical Ltr Rept, Pacific Northwest Natl Lab Review of Main Yankee In-Situ Pressure Test Assessment, Rev 2

ML20147A710
Person / Time
Site:	Maine Yankee
Issue date:	12/31/1996
From:	Chattapadhyay, Simonen F Battelle Memorial Institute, PACIFIC NORTHWEST NATION
To:	NRC (Affiliation Not Assigned)
Shared Package
ML20147A693	List: ML20147A686 ML20147A710
References
CON-FIN-E-2029 NUDOCS 9701280126
Download: ML20147A710 (70)

Text

. _ _ _ __ __ _ . . _ _ . - _ _ _ _ . - . . . . _ . _ _ . . . . _ . _ . . _ . . . _ _ _ _ . . .

l l l

I i

Technical Letter Report- '

on  !

l PACIFIC NORTHWEST NATIONAL LABORATORY REVIEW 0F l

MAINE YANKEE IN-SITU PRESSURE TEST ASSESSMENT (Revision 2)

L i

l i

i S. Chattapadhyay I l F. A. Simonen f

1 l

l l December 1996-I i

l Prepared for l Office of Reactor Regulation U.S. Nuclear Regulatory Commission under Contract DE-AC06-76RLO 1830

NRC JCN E2029 i

Pacific Northwest National Laboratory Richland, Washington 99352 9701280126 970124 PDR ADOCK 050003C-P PDR j

l i _ _

J

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

l l

I i

l l

l l

l Technical Letter Report f

l on PACIFIC NORTHWEST NATIONAL LABORATORY REVIEW 0F j l

MAINE YANKEE IN-SITU PRESSURE TEST ASSESSMENT '

l (Revision 2)  ;

I l'

l S. Chattapadhyay F. A. Simonen

! December 1996 l

Prepared for Office of Reactor Regulation U.S. Nuclear Regulatory Comission under Contract DE-AC06-76Ft.0 1830 NRC JCN E2029 l

l t

l Pacific Northwest National Laboratory Richland, Washington 99352 e

e t

k 9701280126 970124 PDR ADOCK 05000309 P PDR

PACIFIC NORTHWEST NATIONAL LABORATORY REVIEW 0F MAINE YANKEE IN-SITU PRESSURE TEST ASSESSMENT '

This report documents a review of technical issues associated with in-situ pressure testing of degraded steam generator tubes at the Maine Yankee Atomic Power Station. This review was performed under Task Order No. 7 of NRC Job Code Number E2029 and addressed the specific issues of:

1. whether the load induced by in-situ testing across a circumferential crack located at the top of the tubesheet (TTS) effectively simulates the load anticipated from the maximum primary-to-secondary differential pressure during a main steam line break event, and

2. the validity of Maine Yankee's conclusion that a 360' circumferential, 79 percent through-wall crack is the limiting flaw for a TTS circumferential crack in steam generator tubing.

This report sumarizes the background, objectives, and conclusions of the review. Attachments to this report provide details for the calculations and evaluations that were performed in support of the review.

BACKGROUND The background on the issues that were addressed in the current study was highlighted in the statement of work for Task Order No. 7 as follows:

On July 15, 1994, the Maine Yankee Atomic Power Company, the licensee for the Maine Yankee Atomic Power Station, shut down the plant due to high primary to secondary coolant leakage in excess of the administrative limits. The licensee identified four leaking tubes in the Steam Generator 2. A subsequent eddy current inspection of all the tubes in the steam generator. identified numerous defective tubes with several containing cracking with significant through-wall depths encompassing the entire circumference of the tube. The active-degradation of concern was primary water stress corrosion cracking circumferentially orientated in the roll transition region near the top of tubesheet.

In order to assess the residual strength of the cracked tubes, the licensee for Maine Yankee completed analytical calculations as well as in-situ pressure testing of ten of the most severely degraded steam generator tubes. The analyses determined that the limiting top of tubesheet (TTS) 360' circumferential steam generator tube crack is a 79 percent through-wall flaw. Calculations were completed in accordance 1

}

i with the criteria in NRC Regulatory Guide 1.121. Based on the results of in-situ pressure testing, the licensee concluded that all tubes were l capable of sustaining loads postulated during a main steam line break l (MSLB) without rupture. l The NRC staff raised concerns that the methods used for the in-situ testing may' not have adequately simulated tube loads from MSLB dry ,

conditions and requested that Pacific Northwest National Laboratory '

(PNNL) independently evaluate the licensee's conclusion that a 79 percent through-wall crack is the limiting flaw size for a TTS circumferential crack.

The documents that were reviewed during the course of this Pacific Northwest National Laboratory evaluation are listed in Attachment #1. 1 i

OBJECTIVE The objective of Task Order No. 7 was for Pacific Northwest National Laboratory to provide technical expertise to assist NRC staff in determining 1 whether the in-situ pressure testing conducted during the Summer 1994 outage at Maine Yankee demonstrated that all tested tubes had sufficient remaining strength to withstand a main steam line break without failure.

CONCLUSIONS The evaluations performed by Pacific Northwest National Laboratory and '

documented in this report support the following conclusions for the two specific issues described above.

Loads Induced by In-Situ Pressure Testina Pacific Northwest National Laboratory concludes that the loads induced during in-situ testing across a circumferential crack located at the top of the tubesheet (TTS) do effectively simulate loads anticipated from the maximum primary-to-secondary differential pressure during a main steam line break event. Pacific Northwest National Laboratory has independently performed structural analyses (see Attachment #2) to derive equations which support the conclusions and the technical bases for validity of the in-situ pressure tests.

Validity of Limitina 79 Percent Throuah-Wall Crack Pacific Northwest National Laboratory disagrees with Maine Yankee's conclusion that a 360* circumferential, 79 percent through-wall crack is always a valid bound as the limit for TTS circumferential cracking in steam generator tubing.

Maine Yankee has assumed, on the basis of a limited set of tube burst data, 2

i that burst pressures can be predicted by simple considerations of average l

crack depth and the ultimate strength of the tube material. Pacific Northwest I Wational Laboratory's evaluation of a larger set of burst tests (see Attachment #3) indicates that this assemption, although conservative in many cases, does not always provide a lower bound on burst pressures, i

Nevertheless, it has been noted that certain Maine Yankee tubes with cracks considerably deeper than 79% as measured by eddy current testing (ET) had adequate structural margins as indicated by in-situ pressure tests. This j trend occurs because ET cannot resolve the ligaments typically present in service degraded tubes and which provide additional strength to the tubes.

However, for the widest range of possible crack morphologies (i.e.

configuration of ligaments and the circumferential variation in the depth of cracking), the use of ultimate strength with circumferentially averaged crack depths can give unconservative predictions of burst pressures. It should be noted that this concern may be limited to data from test specimens with machined rather than natural defects. The use of flow strength for the governing stress would, however, provide a more appropriate empirical basis for truly bounding the available data from the burst tests including data from tubes with machined defects.

Effects of SSE and Other loads on Limitino Crack Deoths The Maine Yankee submittal (GDW-94-92) provided estimates of the stresses associated with safe shutdown earthquakes (SSE), pipe break impulse response, and flow induced vibrations. Table 4 on page 35 of GDW-94-92 cites the 3 following levels of calculated stress for a 79% average through-wall detect: 1 Loading Condition Stress, ksi Steam Line Break 32.8 ksi Pressure AP = 1750 psi Steam Line Break 47.E; Ksi Pressure AP = 2520 psi Pipe Break Impulse Response 2.3 ksi Safe Shutdown Earthquake (SSE) 2.3 ksi Flow Induced Vibration 0.91 ksi The estimated stress levels from other sources are less than 10% of the AP pressure stresses, and do not attain their maximum levels at the same times as the dominant pressure stresses. Accordingly, the Maine Yankee submittal based the governing flaw depth only on the maximum pressure stress. Pacific Northwest National Laboratory considers this approach of using only the pressure stress to be appropriate, given the cited stress levels and time  !

3 l

i sequences in the Maine Yankee report.

4

SUMMARY

OF PACIFIC NORTHWEST NATIONAL LABORATORY TECHNICAL EVALUATIONS A number cf calculations and other evaluations were performed by Pacific Northwest National Laboratory in support of the review. Details of these efforts are provided in the following series of attachments to this report.

Attachment'#1 - List of Reviewed Documents Pacific Northwest National Laboratory was provided a collection of documents related to the in-situ pressure testing and related issues at Maine Yankee.

Attachment #1 lists those documents that were of primary interest and also lists other background documents of interest but not directly related to the in-situ testing issues. A third category includes documents from Pacific Northwest National Laboratory's files that also supported the review with a particular focus on burst test data.

Attachment #2 - Evaluation of Loads Induced Durino In-Situ Pressure Testina Pacific Northwest National Laboratory performed independent structural analyses of the loads imposed on the degraded / cracked tubes during in-situ pressure testing. In each case the calculated loads for the test situation were compared to loads imposed during steam line break events. While Pacific Northwest National Laboratory's approach to the structural analyses differed somewhat from that in the Maine Yankee submittals, the PNL work supported the overall conclusions of the Maine Yankee submittals. It is shown in Attachment

1. 2 that the in-situ pressure test does impose loads essentially the same or greater than those of the MSLB event.

Attachment #3 - Review of Burst Test Data Burst test data cited in the Maine Yankee submittals (along with a much larger collection of data from other publications) were reviewed for consistency with the assertion that a 360' circumferential 79 percent through-wall crack always provides a valid bound on the limiting flaw for TTS circumferential cracking in steam generator tubing. Attachment #3 documents the test conditions and test results from the collection of data from several sources. The tests include many specimens having machined / planar defects that are not necessarily representative of service induced stress corrosion cracks with ligaments. As such, the test data provide a lower bound for mature states of service degradation having corroded ligaments. Accordingly, the data show many exceptions to the statement that average crack depths and ultimate strength can be used as the basis for predicting burst of tubes with circumferential cracking.

4

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

Attachment #4 '- Crack Openina Disolacement Considerations The Maine Yankee submittals propose the possible use of elastic plastic

! fracture mechanics (J-integral criteria) as a theoretical method for j

predicting the stability of circumferential cracks. Attachment #4 discusses-

, some difficulties with this approach. For steam generator; tubes with deep part-through cracks, it is concluded that the physical dimensions of uncracked j remaining ligaments are very small (on the order of 0.005 in.). This

invalidates underlying assumptions of the J-integral approach. As an

alternative fracture mechanics method, Attachment #4 proposes the use of a crack opening displacement approach. Estimates of critical values of crack i opening displacement are given in Attachment #4 as a function of the remaining ligament dimension.

l Attachment #5 - Comments on Elastic-Plastic Analysis of Flawed Steam Generator j Tube Includina Effects of Temoerature on Material Properties ,

i

, Thi, Attachment addresses the validity of the method used by Maine Yankee to {

adjust burst test data from room temperature tests to account for the effects 1

! of elevated temperatures. The Maine Yankee approach considers only the i i effects of temperature on ultimate strength; whereas, Pacific Northwest '

} National Laboratory's approach also addresses the effects of temperature on i

ductile tearing properties. Data from the literature on the effect of l temperature on material strength and ductile tearing resistance are reviewed.  !

This attachment also describes a scoping analysis to determine the values of j the J-integral associated with an internal pressure of 5000 psi for various  ;

through-wall crack depths. '

i

) Ductile tearing resistance curves at elevated temperatures up to 800 'F were

found to be equal to or greater than those at room temperature. This trend in 4 l the material's ductile tearing resistance further justifies Maine Yankee's use  !

j of a room temperature pressure test with a proposed 15% increase in pressure to account for differences in tube material strengths.

e The results of Pacific Northwest National Laboratory's J-integral evaluation

! are compared with results presented in the Maine Yankee submittals. While l differences in numerical results were noted, the most important point of the j evaluation (shown in Attachment #5) is that the size domain of the cracks in steam generator tubes is inconsistent with the existence a J dominance field, j

and that published J-resistance curves are not applicable to the Maine Yankee '

i evaluation.

! 1

] Attachment #6 - Effect of Induced Tube Bendina on Fracture of Tubes with Part {

{ Throuah Flaws t

The fracture mechanics evaluation of Attachment #6 addresses the scenario of

] ductile crack extension in a pressurized tube containing a part-through l 5 1

a -, - ,-e- , , -.

circumferential crack of variable depth. .These calculations were performed to help explain behaviors observed during some in-situ pressure tests. The evaluation included the effects of tube bending as induced by the variable crack depth. The predictions are based on a crack opening displacement criteria for ligament fracture. The pred.icted crack extension behavior is consistent with the deepest part of the crack first breaking through the wall at relatively low pressures, and with the resulting through wall crack sustaining a substantially increased pressure before burst. The predicted scenario is consistent with events observed during in-situ pressure testing.

6

9 AP >A - m,na n- -b,4- A a m... M,.4,. M 4A--- n2 A 6 A u.s---,,,-- nJL,4a--- ,,a-A a .4: - - 6n,4-l l

1 A

k 4

ATTACHMENT #1 LIST OF REVIEWED DOCUMENTS i

4

1 ATTACHMENT #1 LIST OF REVIEWED DOCUMENTS Primary Documents Reviewed -

Maine Yankee Letter MN-95-24, " Response to Steam Generator Issues Request for Additional Information: Elastic Plastic Finite Element Structural Analyses of a Degraded Steam Generator Tube During In-Situ Pressure Testing and Actual Operating Conditions", March 10, 1995 Yankee Atomic Letter MN-95-22, " Attachment A : Expansion of Maine Yankee November 14, 1994 Letter's Enclosure Attachment B (Point of Loading Effect on Validity of In-Situ Pressure Loading)", March 3, 1995 Maine Yankee Letter MN-94-104, " Maine Yankee 1994 S/G Inspection Report GWD-94-92", November 14, 1994 CENC 1930, " Test Report for Leak Rate and Pressure Tests of Flawed Steam Generator Tubes Simulating Circumferential ID Cracking of Tubes Found in Maine Yankee Steam Generators", January 1991 to March 3, 1995, Maine Yankee submittal to the NRC, ABB C-E Report CENC-1934, " Maine Yankee Steam Generator Analysis of Circumferentially Flaws at Tubesheet", January 1991. to March 3,1995, Maine Yankee submittal to the NRC, ABB C-E Design Analysis MY-SS-900, " Steam Generator Tube Loads for Tube Lock-Ups",

October 5, 1994 to March 3,1995, Maine Yankee submittal to the NRC, ABB C-E Design Analysis MY-SS-901, " Steam Generator Tube Loads Due to Main Steam Line Break", October 5, 1994 to March 3,1995, Maine Yankee submittal to the NRC, ABB C-E Design Analysis CENC-1965, " Amended Analysis of Maine Yankee Steam Generator Circumferentially Flawed Tubes at Tubesheet", dated October 10, 1994 Letter Report from R. Gamble (Novetech Corporation) to J.F. Ely (Northeast Utilities), " Review of the Safety Assessment Performed for Millstone Unit 2 Steam Generators", March 28, 1990 Northeast Nuclear Energy Company Millstone Unit No. 2 Docket No. 50-336i Steam Generator Examination Results and Safety Assessment, February 22, 1992 Meeting with NRC Staff 1-1

Secondary Documents Reviewed Enclosure 5 to March 3, 1995, Maine Yankee submittal to the NRC, ABB C-E Technical Report M-PENG-TR-002, " Maine Yankee Run Duration Limit Evaluation for Circumferential Cracking", October,1994 Enclosure 6 to March 3, 1995, Maine Yankee submittal to the NRC, Yankee Atomic Electric Co. Memo TAG-MY-94-051, " Maine Yankee S/G Tube Leakage Assessment -

Summary of Results", dated October 27, 1994 Enclosure 7 to March 3, 1995, Maine Yankee submittal to the NRC, Yankee Atomic Electric Company Memo REG 203/94, "Offsite Doses as a Function of Primary-to-Secondary Leakage due to MSLB with Induced Steam Generator Tube Leakage (Rev.

1)", dated October 26, 1994. to March 3,1995, Maine Yankee submittal to the NRC, ABB C-E Letter WO94243.RM dated 11/10/94 to C. Eames of Maine Yankee,

Subject:

" Transmittal of Documents: MRPC Sizing of Circumferential Cracks" to March 3,1995, Maine Yankee submittal to the NRC, Proceedings:

1992 EPRI Workshop on PWSCC of Alloy 600 in PWRs. "EPRI TR-103345 dated December 1993: Paper " Potential Benefits of Zinc Addition to PWR Coolant" by R.E. Gold 0 to March 3, 1995, Maine Yankee submittal to the NRC, Applicable Extract of ABB CE NPSD-957 (CE0G Task 729) "S/G Tube Degradation at the Support Plates" by G.C. Fink and S.M. Schloss; October 1994 1 to March 3, 1995, Maine Yankee submittal to the NRC, ABB C-E Letter M-PENG-94-013 dated November 10, 1994 to C. Eames of Maine Yankee;

Subject:

S/G Tube Rupture in PWR Plants Other Documents on Tube Burst Tests Electric Power Research Institute Report NP-6626-SD, " Belgian Approach to Steam Generator Tube Plugging for Primary Water Stress Corrosion Cracking",

March 1990 r

Electric Power .esearch Institute Report NP-6865-L, " Steam Generator Tube Integrity Volume 1: Burst Test Results and Validation of Rupture Criteria (FramatomeData)", June 1991 1-2

I i

4 l

1 l

ATTACHMENT #2 EVALUATION OF LOADS INDUCED DURING IN-SITU PRESSURE TEST

i ATTACHMENT #2 EVALUATION OF LOADS INDUCED DURING IN-SITU PRESSURE TEST I

i This attachment summarizes Pacific Northwest National Laboratory evaluations ,

of the loads applied during the in- situ pressure tests performed at Maine '

Yankee.

BACKGROUND In order to demonstrate the residual strength of circumferentially cracked

, tubes, the licensee for Maine Yankee has performed in-situ pressure tests of a 4

number of severely degraded tubes. The test is performed with a specially l

designed hydro test tool which is used to pressurize a short length of an individual tube. This length includes the circumferential1y cracked part of the tube which is situated just above the tubesheet. Pressure testing is performed at room temperature, with the pressures increased by 15% to account for the estimated differences in strength of the tube material at the i operating temperature compared to its strength at room temperature. The hydro tests are intended to simulate the loading on the degraded tube for the conditions of a main steam line break (MSLB) event, and to demonstrate a margin of 3.0 against burst at the normal design operating pressure.

The re' view by Pacific Northwest National Laboratory was conducted to determine whether the in-situ pressure testing conducted at Maine Yankee during the Summer 1994 outage actually demonstrated that all tested steam generator tubes had sufficient remaining strength to withstand a main steam line break without failure. While the pressures in the in-situ tests were sufficient to simulate the MSLB condition, pressurization of only a short length of a single tube did not correspond to the actual loading for the MSLB for which pressurization would occur over the full length of all tubes in the steam generator.

DOCUMENTATION SUBMITTED BY MAINE YANKEE Pacific Northwest National Laboratory reviewed documents submitted by Maine lenkee that describe evaluations of the in situ pressure test. The review was performed to establish the' ability of the in-situ pressure test to simulate conditions of concern to tube integrity such as the MSLB condition.

Attachment B of a November 14, 1994 letter submittal from Maine Yankee was titled " Point of Loading Effect on Validity of In-Situ Pressure Test". An I analysis concluded that the pressure test exerts more load than the same pressure experienced for the conditions of normal operation. Reviews 06 analysis by NRC staff identified concerns. The analysis unrealistically' '(his assumed that only one tube (the degraded tube being pressure tested) was locked by corrosion products into the first horizontal support or " egg crate".

The analysis also failed to account for the change in tube length (Poisson effect) associated with pressure induced hoop stresses in the tube. The 2-1

November 14,-1994 submittal was disregarded by Pacific Northwest National Laboratory in the current review.

A subsequent submittal of March 3, 1995 addressed the NRC staff concerns with the November 14th analysis. Attachment A of the March 3 submittal was titled

" Expansion of Maine Yankee November 14, 1994 Letter's Enclosure Attachment B (Point of Loading Effect on Validity of In-Situ Pressure Test)". The revised analysis built on the prior evaluation by addressing the " Poisson effect" and by more realistically assuming that all tubes were locked into the first horizontal support. Pacific Northwest National Laboratory found the assumptions and general approach used in the derivations and equations in the March 3rd submittal to be correct. The March 3rd submittal presented only the derivation of equations with no numerical results or conclusions regarding the validity of the in-situ pressure test.

Another submittal on March 10th used a finite element analysis to quantify parameters in the March 3rd equations, and concluded that the "in-situ pressure test imposes the same load through the steam generator tube as experienced from normal and/or transient steam generator tube pressure differentials when many tubes are locked in the horizontal support plates."

Based on an independent structural mechanics evaluation, Pacific Northwest-National Laboratory concurs with the March 10th conclusion regarding loads on the degraded steam generator tubes. The equations and input parameters from the finite element analyses appear to be reasonable, but were not checked in detail by Pacific Northwest National Laboratory. Independent bounding calculations were been performed by Pacific Northwest National Laboratory, and  !

these calculations do support the conclusions of the March 10th submittal. i INDEPENDENT CALCULATIONS FOR LOADING 0F DEGRADED TUBE 1 In the independent calculations Pacific Northwest National Laboratory placed bounds on support plate stiffness and on the stiffness of the cracked length of the degraded tube, whereas the Maine Yankee submittal used finite element ,

analyses to quantify these stiffness parameters. The following assumptions '

were made in the. Pacific Northwest National Laboratory calculations:

The tube was assumed (as a limiting case) to be completely cracked with zero stiffness at the crack location). Rather than calculating the load imposed on the crack, the relative displacement (or crack opening) between the ends of the severed tube was calculated.

• The support plate stiffness was taken to be either infinite or zero.

-The case of infinite stiffness corresponded to tubes near a vertical tie rod, whereas the zero stiffness case addressed tubes located remote from a tie rod.

2-2

The Pacific Northwest NaConal Laboratory evaluation addressed a single cracked tube surrounded by a large number of uncracked tubes, with all tubes assumed to be locked (by virtue of corrosion) into the first horizontal support plate. In one case (zero stiffness of the support plate) the vertical motion of the support plate corresponded to the change in length of the uncracked tubes extending from the tubesheet to the first support plate. In the other limiting case (infinite stiffness of the support plate) there was zero vertical motion of the support plate due to the presence of a vertical tie rod. The calculations were based on the following nomenclature as indicated in Figure 1.

l 1

D = Inner Diameter of Tube i E = Elastic Modulus ,

P = Internal Pressure fi = Vertical Distance from Top of Tubesheet to Upper Seal of Hydrotest Device f, = Vertical Distance from to Upper Seal of Hydrotest Device to First Horizontal Support Plate t = Wall Thickness of Tube 4

y = Poisson's Ratio Case #1 Cracked Tube is located Remote from Tie Rod During the hydrotest of a single tube, the vertical motion of the support plate is zero since the plate is restrained in the vertical direction by the large number of unpressurized tubes. These unstressed tubes (being locked in place) maintain the nominal separation between the tubesheet and the support plate. The single pressurized tube will try to shorten due to the Poisson contraction effect over its pressurized length (f ) i by an amount (pD/2Et)(vf i) . The unpressurized length (f,) will shorten by an amount (pD/4Et)(f,) due to the vertical thrust of the pressure acting on the upper fitting of the hydrotest fixture. This gives a net opening across the circumferential crack in the tube 6,,mt,,, = (pD/2Et)(f,/2 + vf i) (1)

For the steam line break condition with the same pressure, the uncracked tubes will all be pressurized and this will cause the tube support plate to move upwards vertically by an amount 6= (pD/4Et)(fi + f,)(1 - 2v) (2) 2-3

due to the Poisson contraction effect for the pressurized length of the tubes and due to the end thrust on the tubes. The single cracked tube (being deeply cracked or in effect completely severed) will have zero axial stress and will by virtue of the Poisson effect in the limit shorten by the amount (pD/2Et)[v(f + f,)] giving a net opening across the cracked location of 1

6,=

3t (pD/4Et)(f + f 2) 1 (3)

This gives the relative displacement (or load) across the crack for the hydrotest relative to that for the steam line break as 6,,,,.,,,,, /6st, = (f, + 2vf )/(f, + f i )

3 (4)

Assuming a Poisson ratio of v = 0.3, the following ratios of crack opening were calculated as a function of the loading fixture location relative to the top of the tubesheet.

(f/f) 3 2 6 ,,,,,,,,/6st.

0.00 1.000 0.01 0.996 0.05 0.981 0.10 0.963 0.20 0.933 0.50 0.866 The Maine Yankee submittal indicates values ofi f = 5 inch and f, = 42 inch giving f /frt = 0.12. For this ratio the hydrotest load on the crack is about 95% of the load expected for the steam line break. The March 10, 1995 submittal from Maine Yankee (based on finite element calculations) reported load ratios of slightly greater than 1.0, which is in general agreement with the Pacific Northwest National Laboratory analysis. The small difference can be attributed to the simplifying assumption of zero support plate stiffness used in the Pacific Northwest National Laboratory calculations.

.. .j Case #2 Cracked Tube is Located Ad.iacent to a Tie Rod For the hydrotest of a single tube, the vertical motion of the support plate is again zero, since the plate is restrained in the vertical direction by the 2-4

large number of unpressurized tubes as well as by the adjacent tie rod. As for Case #1, the single pressurized tube will try to shorten due to the Poisson contraction effect over its pressurized length (f ) t by an amount (pD/2Et)(vfi ) . The unpressurized length (f,) will shorten by an amount (pD/4Et)(f,) due to the vertical thrust of the pressure acting on the upper '

fitting of the hydrotest fixture. This gives a net opening across the circumferential crack in the tube of 6 ,,,,,,,,, = (pD/2Et)(f2 /2 + vf i) (5)

For the steam line break condition with the same pressure, the uncracked tubes will all be pressurized but the tube support plate will be prevented from moving upwards vertically by the high stiffness of the adjacent tie rod.

The cracked tube by the assumption of being " deeply cracked" will have zero axial stress and will by virtue of the Poisson effect shorten by the amount (pD/2Et)[v(f + f,] giving a net opening across the crack location of 1

6,=

3t (pD/2Et)(vf + vf r) 1 (6)

This gives the relative displacement (or load) across the crack for the hydrotest relative to that for the steam line break as 1 6,,,,,,,,,/6 3t , = (1/2v) (f, + 2vf )/(fi + f,)

3 (7)

Assuming a Poisson ratio of y = 0.3 gives the following ratios of crack l opening.

(f/fr) i 6hy4 rote,,/63t ,

l 0.00 1.67 0.01 1.66 0.05 1.64 0.10 1.61 0.20 1.55 0.50 1.44 The values of f = 5 inch and f, = 42 inch gives f /f, = 0.12. For this length i 1 ratio the hydrotest load on the crack is about 1.59 times the load expected for the steam line break. The March 10, 1995 submittal from Maine Yankee (based on finite element calculations) reported corresponding load ratios of only slightly more than 1.0. The finite element model addressed the specific case of the hydrotest of a tube remote from a tie rod, and evidently predicted 2-5

significant motion of the support plate at the tube location.

Summarv and Conclusions Pacific Northwest National Laboratory performed independent structural analyses of the loads imposed on the degraded / cracked tubes during in-situ pressure testing. In each case the calculated loads for the test situation were compared to loads imposed during steam line break events. While Pacific Northwest Nationa's Laboratory's approach to the structural analyses differs

somewhat from that in the Maine Yankee submittals, the results support the i

overall conclusions of the Maine Yankee submittals since it is shown that the in-situ pressure test does impose loads essentially the same or greater than those of the MSLB event.

i  ;

i 3

i 1 l

1 i

4 i

i i

i 4

2-6 4

I

4 l

l 4 4 5 I J &

b Assumotions on Support Plate Motion Support Plate (Tubes Locked iato Case #1 - Cracked Tube is Located SupportPlate) Remote from Tie Rod (Plate Motion Governed by Length of Uncracked Tubes) f, Case #2 - Tube is Located Near Tie Rod.

p End Fixture of Hydrotest

• Tool Grips Tube U

]

• 414

1. 1

• Pressure 4 4  %

q -

4 - 6 = Crack Opening

/ /rj g g i Displacement

' a- 4 '

/ 4 + s Tubesheet / /

/ ,

/ /

/ /

\\\\ /-/

4 Figure 1 Loading of Cracked Tube During In-Situ Pressure Test 2-7

k 1

l i

i 4

4 4

i

! ATTACHMENT #3 4

REVIEW OF BURST TEST DATA

)

4 i

l 4

l ATTACHMENT #3 i REVIEW OF BURST TEST DATA This attachment reviews the basis for the 79% limiting flaw depth, and l compares the assumptions in the technical basis with data from burst tests of steam generator tubes with circumferential flaws. The Maine Yankee submittals (GDW-94-92 and CENC-1965) have stated that the limiting flaw is a part through circumferential flaw with a uniform a/t = 79% through-wall depth. With this flaw present, a degraded tube is considered to be capable of sustaining a pressure of 5,000 psi, which is equal to three times the normal operation pressure.

The submittals make the following assumptions regarding burst of tubes with circumferential flaws.

/

1. An average flaw depth can be used to address those cases for which the actual flaw has a nonuniform depth. That is, the burst pressure is governed only by the remaining uncracked area of the cross section, and is independent of other details of the crack geometry (i.e. uniform part-through crack, nonuniform part-through wall crack, or through-wall crack). No test results are presented in the Maine Yankee submittal to support this assumption.

2. It is assumed that the tube will rupture when the average pressure induced stress on the remaining cross section is equal to the ultimate strength of the tube material. The governing stress is calculated as the end load from internal pressure divided by the remaining uncracked cross section.

3. It is assumed that the critical stress level corresponds to the ultimate strength of the tube material. In contrast, other published approaches for predicting tube rupture have used the flow strength of the tube material, with the flow strength typically being taken as 0.5(o, + o,)

or 0. 6 (o, + o o) . The Maine Yankee submittals provide no test data or other justification for a critical stress equal to the ultimate strength as compared to a flow stress.

The assumptions of the Maine Yankee submittal would, for example, imply that a tube with a through-wall flaw extending 79% around the tube circumference would have the same burst pressure as a tube with a uniform depth flaw with a/t=79%.

Basis for 79% Limitina Crack Depth For the limiting flaw with a/t = 79%, the stress on the net cross section for a pressure of 5,000 psi was calculated. The tube dimensions of interest are Outer Diameter = D, = 0.750 inch 3-1

, Wall Thickness = t = 0.048 inch Inner Diameter = Di = 0.654 inch l Diameter to l Outer Crack Tip = D = 0.730 inch 3

Cross Sectional Area = (n/4)(D,2 - D tz) 2 2

= (n/4)(0.750 - 0.730 ) j i

= 0.0234 in' The axial loads and stresses are as follows Axial Load From Pressure = (nDr/4) pressure i

= (nx0.730 /4)5,000.

2 i = 2,093 lb.

l Axial Stress From Pressure = Load / Area = 2,093/0.0234

= 89,400 psi e

This applied stress can be compared with the 80,000 psi ultimate strength of SB-163 tubing at a 600* F operating temperature as given in CENC-1965. It should be noted that this value of ultimate strength is somewhat less than the ultimate strength of 96,000 psi as given by data tabulated in Mills (1987).

It should be noted that the 89,400 psi calculated stress as calculated here is !

5 somewhat greater than the stress level of 80,000 psi as calculated in CENC-  ;

1965. The higher stress is due to the fact that the present calculation '

l accounts for the effect of the internal pressure acting on the faces of the crack.

, Scope of Burst Test Data Review Pacific Northwest Nacional Laboratory has reviewed available data from burst tests on steam generator tubes with circumferential cracks to determine if the assumptions made in the Maine Yankee evaluations are appropriate. It was found that the most test data for circumferential cracks were for through-wall cracks as opposed to the part- through wall cracks of greater concern to the Maine Yankee application. Nevertheless, there were some data for part-through wall circumferential cracks, and for part-through wall cracks of variable depth.

In reviewing the burst test data it became evident that the following details of the testing needed to be considered:

1

1. In most cases the burst pressure did not corresponded to a full double ended circumferential break, but rather a tearing of the remaining ligament for the part-through crack followed by large opening of the 3-2

)

s

, resulting through-wall crack. Such opened cracks leaked at high rates in the tests. As such, the definition of burst pressure for a given test was dependent on the flow rate capability of the testing system

being used.

2. In many cases, bending the tube at the crack location and consequently the opening of the crack was restricted in the test facility by simulation of the tube support plates that are present in a steam generator. In other cases the tests were performed on a free standing tube with no restraint against bending and opening of the cracks. The resulting burst and leak behavior of tubes is quite different under the

., two test conditions.

3. None of the tests fully addressed the mitigating factor for Maine Yankee tubes whereby the tubes are locked into the first tube support plate.

The burst tests (even for the cases with the simulation of tube support plate restraint) conservatively permitted the tube to slide through the hole in the simulated plate, and thus provided less restraint to burst of the tube than is thought to have existed for the Maine Yankee 4

condition.

1 4. In some cases the burst tests were performed with the aid of bladders and sealing patches to prevent leakage from an opening crack from defeating the object ive of attaining a large scale crack opening associated with tube burst. With such bladders and other sealing i

methods, there was no need for high flow rate test facilities to j simulate the flow rates available with an operating steam generator. l j

5. Most burst data are from specimens with machined defects rather than from tube specimens with natural cracks. ,uch data gives a pessimistic indication of the pressure capability of tubes with service related degradation mechanisms. There is a lack of data to fully characterize the burst behavior of tubes with nonuniform crack depths, and of tubes with stress corrosion cracking which is characterized by a network of cracks separated by ligaments rather being in the form of single planer '

flaws is approximated by machined defects.

Having stated some of the important issues associated with burst testing, the remainder of this attachment presents and interprets burst test data that were available to Pacific Northwest National Laboratory from various documents.

ABB/CE Burst Tests for Maine Yankee Tubes Tests were performed by ABB/ Combustion Engineering during 1991 on tubes with laboratory induced (by stress corrosion cracking) circumferential cracks (CENC 3-3

1930). While the main objective of the test program was to measure leak rates, there were a few tests that indicated burst pressures. These burst test data were cited in the recent Maine Yankee submittal as validating the assumption that a circumferentially cracked tube can sustain a net section i l

stress (based on average crack depth) equal to the ultimate. strength of the I tube material . 1 The tests of CENC 1930 were performed in a facility that simulated the restraint of lateral motion provided by the first support plate as situated l

about 47 inches above the tubesheet. While the test plan was to pressurize l all specimens to 4800 psi to establish their ability to sustain a pressure of '

three times operating pressure, some tests did not achieve this pressure because the leak rates exceeded the flow rate capability of the test facility.

None of the specimens had bladders or patches to seal the crack as is often ,

l needed for purposes of burst testing. Only for a few specimens was the testing sequence continued to higher pressures beyond the objectives of the leak rate measurements. Thus only a few tests achieved the higher pressures needed to actually burst the specimens. The tests of CENC 1930 should be considered of only limited value for establishing tube burst capability, since l no bladders were used in the testing and since the facility had a very limited flow capability.

l The circumferential cracking was characterized by an average through-wall i depth as measured both by UT tests and metallographic examinations. The corrosion induced defects were not single continuous cracks but were networks of cracks connected by ligaments. Thus the characterization of each specimen in terms of an average crack depth would tend to discount the reinforcing effects of the ligaments. As such, predictions of burst pressures based on the average depths would tend to underestimate the pressure capabilities of the tubes. Unlike the corrosion induced cracks of this ABB/CE test program, most other burst test data for circumferential defects are from machined type cracks without the complicating effects of ligaments.

Table 1 lists the measured burst pressures as given in Table-7-2 of CENC 1930 along with the average net section stress based on the average crack depth.

Our evaluation specifically includes specimen T-12. The CENC 1930 assessment describes this specimen as having " guillotined during the PMSB test", and was therefore excluded from the Maine Yankee assessment of burst pressure criteria. Footnotes to Table 7-2 imply that the tube burst event was associated with cracks within the tube sheet crevice region, rather than the cracking mode of interest in the roll transition region. Pacific Northwest National Laboratory believes exclusion of this test from the burst assessment has not been adequately justified. The relatively low pressure of this test may be evidence that the proposed burst criteria based on ultimate strength and average crack depth is not a bounding measure of tuce pressure capability.

Nevertheless, in most cases the proposed ultimate strength criteria does correlate with the CENC 1930 test data. This correlation should apply to 3-4

I service degraded tubes, particularly when the crack depth is essentially uniform around the tube circumference or when (as for the ABB/CE specimens) the presence of ligaments characteristic of stress corrosion cracking significantly enhances the burst capability of the tube.

Figure 1 shows the average stress levels at tube failure from the CENC 1930 tests, with Figure 2 indicating the legend used for this and other plots discussed in this attachment. None of the tests of Figure 1 achieved pressures sufficient to stress the remaining cracked cross-section to a level approaching the ultimate strength. In all but one case the maximum pressure was limited by leakage corresponding to the 0.5 gal / min flow capability of the test system, and the. actual burst capability for these specimens could not be established. The one specimen (T-12) which did actually experience a burst l mode of failure had a net section stress of 38,000 psi at the failure l pressure. This stress is less than 50% of the ultimate strength. As stated j

above, the relevance of this specimen to the Maine Yankee situation is  !

questionable. The stress corrosion cracking was produced under laboratory rather than field conditions. The cracking corresponded to the fully expanded l tube sheet crevice region rather than at the expansion transition region of- l interest to Maine Yankee. Additional information on the test specimen and l testing procedures is needed to judge the relevance of specimen T-12.  !

I l

l l

3-5

r i

Table 1 Maine Yankee Burst Tests as Reported in CENC-1930 (January 1991)

Specimen Crack Crack. Crack Depth, Restraint Configuration Full Burst Pressure Average Stress Length e/t Circumferential (stress corrosion (everage depth) Fracture onNetRemainig i cracks with Cross Section"  !

Ligaments)

R75L34 Part Through Flaw 360* 45-48% Yes No >4,800 psi >28,300 psi (ttbe removed from NY)

T-8 Part Through Flaw 360* 58% Yes i

No >6,000 psi >41,900 psi l T-10 Part Through Flaw 360* 75% Yes No >4,950 psi >59,800 psi  !

T-12 Part Through Flaw 360* 77% Yes Yes 2,900 psi 38,000 psi (specimen may t.a have burst due j to cracks within

,e tthesheet crevice) I T-14 Part Through Flaw 360* 16% Yes No >9,350 psi >35,900 psi (1) Axist Stress = Axial End Load / Net Remaining Cross Sectional Area (2) Outside Diameter = 0.75 inch; Watt Thickness = 0.048 inch '

(3) Stress corrosion cracks produced in laboratory (4) Maximum flow rate = 0.5 gal / min i t

f k

i b

-r f

t

.1 6 L

h F

t

. _ . _ _ _ . _ _ _ _ _ _ _ _ _ _ - . . _ . _ _ _ - _ -._____.____m__ _ _ _ _ _ - __.__________._._._._._._____.-_m_ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _____

I i

1 I

l I

.id .1: W.1:.i' N: '

. . ;; . :i :i- l T -l 70; j.lT ;r::.j j ::q:pp: y: C.

-'"j. i-1 . . . . .l.. ].

• G. E=.! ..i...-  :

.F.

, s. ....p._,

... .se- {:." i . :.- { "1"t.

at.

~

i, . . . .-.p..

.... . .; i iiI'-T r.h.;. ' t.].~.r: -f . j.: a:j : ' : ji.: :i].;  !.j .g/ ;.] iip.~p pj;j-

p.F- ,

=ih  ;':st .,__. ., .c=--,-- -e : , r .r---j-=.- ,c_ ._ . . ._ ; _. :=- y-

. n .. .

i_. l i-h.=li r.br .:dr_ :- ' : ) _: =

. . . . l 6

, p .. .,;p. . . ~.:j.; ..g.::j[.:r,_-j :l ggg,

. p ..~

y =r- . .

arr :: . . _;. J _. ;

_. . _~i-_-l== .-- _r:  ; ~.  !.. ~; a : g. .-ij.ip..).sj:iFjs.:j=gp; :(js,sj.::q=sf -

.. g

-a r=w :;-=.

-==u_

. _=.

= .: .=_. =a_.t=.. . . ;.:n H .. 4

-:=.p_ . u=jv;45:3j= gi:_j.. w_ r=r.ip.1 :.g...

.d.-.

.-.r. = f. . .._: _._: . _ _ . _2

. . . - . 1- l' : ..L. . ~ i u. -i .q-

= c=.%_.;.ir . l.= =>."d e.a .u. ,

~.] m. .c._4.1. l .J. : - .- ,

ce.

-._h=...__....__n__t.=.-

+E= .us=2 . :- ! =. . t. =i. .. _ . .. e .:.l- r_h; .4,4 ::.4. - rps M-Ep.=J:.=Gp5ML:=fici+iF:e.;.: o

=is:4W.="_:v itatiir';. .r:i ).a :-- p4 p; ..,:g=.4 =ffjtN: us; . .=;=g =q=;.p- -

f- i=55Es-;:l=M.=L i:1 :. t-7s ai,: .!/%sN =is-k======_ :;ss:.hr.;j=ge.- g

=- --r.=u.=;L=MuhT.; '

=:n. h.mi;hs.==$;~ .;- _=_:_-= _% ,_ ".--lI=== h

= . = --O g 1

=Eti .:=iM.siyt. . =.i.l6 i;se , ;p:+.  :.g --i=m. :i.. 5 _ = r. . i_; _..__w.___. . . _ . . _ . -

- . ;-m  !

g

_ . __.....v_._.-..._p=.-

..a__.-_.__

. =. =.

r-_._.. _. _

.___4 . _ _ _

s-m i~:!E=-ki;f=it di:li L:ltiEh!!i=il :"P:: ri- +:  : L- -iti-JE:td z.a; :lrEl===E.insig - g

.;;;~_t =l=:_t-ii = = = y=. : m.-i.=. =y. - - - .

. =i. =; =.: t.=. qv- =.. :. . _ _ - - w

. .aa:.: , a u;. . . .

. . . _ _  :._=.=.= , .m

- gi c -3,

;.=_xi.: 2 r---. :q-n.:. -

= - 7 t r.1 -.. .: . - . L =: .u-*-. -m- ::

. . . . - u =.: 3-m I m . . , .---_-.-----hn=-

=lp : .J.p gj= =.20-i

m.4  : .: n . .ggmip _3:-;p=gs- : reg o I W 9-E!: /JI;3.7: ..l.* Li~ .t  ;  :! Mu. M.bi_Fii.pj=1:.._= n $

_l:  :=.1 -

'=r- -

_: _::.gi:_p_i_m .L 3._ . . .  ;.:.. a._ _.: ;_=4 =g .iuy.,;;gm @ u M sit,-:iNisM:!- _~:E-ir.=a-;au . n; i !:E sil=-1 =t E-EEi:E =. t:;E. r- aj . -j' .

4 .!:

. i:.

p I Q-.ij _:

qa : .EshioM=:isW -r:-

=;- g . Epi =j:p:;r;E-

)

,Y 5

v d2 MrT._~5"tG- Ei.:= . ri:i u!. - -ci e  ! .;.  ;.:=Mdih:J5;ig.tiF%=j:=F 5 .c

'=="- ='"

Ei=:a=%ii.riEiM=E.-

Jt .:s. . . cp . .!: . vihj;rFE-f==i;g =E-ap E-p .a. y cr =:.ie=m==; -12E_JEiMIE:': '!F W; " ! .' rp v5:. . : ;;q=2_i; =;.=31.=-q;r.Tf. ci

• _ ' = -- Et.i.5i ._99=fr ==.:== n:.m a q i.i : i.- q=- :4;. 17-s=j;;=p; i::.atig .g

=.. .a .. . . . = . . =,._.i._

. _ . . . ,2- - - . .

< a z . . . - -4.

r -- - r_ l.--r- -

--l 9- .r--

- -l . -- -- i. - l r-- ,i. _ =.a.r:- =-.: --e - ,

N *r - it ~. fjU:; ; = j=~! --y :i'il'"

-l- i  ;- l .l.- . )::f: ~, E

., 5:2 - S i r. .

+j .: .- : .t. r-

. ) .~i ..; j p p. , -i j up :.'p:.j.: . :p M

~'i i m

l:l.T. .

n.. i. ; -. ._.:i..h. i + '[.::iE ' a. l ; ..i .

4

.l . i. . ._.m-::..].

t _. 4 . t:.:

i m

. 13 --.... . -

"..,: - . =:,p _

_t." . 3.. - f m i I

1. ..y r.

l......  :.

u

. 9 g: .l . g

._- -e . _

_ .' .-..- . 1. . --- u :.

2. .

._=..l..

..2 .: . s. . . . . 1 . o a1

, l: . ... ]. .. . . . .

. a. e__ p e

y.

'.[ t) : l cJ z.:.l.i ij ; .rdcl- -- 4 d  !. l...!.. Q l. + il...l...l L.~ g g

ni . ,  ;. 1 l -t:  :.l - . l i, i JR,p l i. C

. r., .

s 3.- . at  : -- 1  : --l * ! - l  ! -l .l l CO j .; . . 7. . .

,  ;  ;  ; i ..  ;..y.. . p;_ ,

g u (p.; . l . ; a { . .; 2.r_h_; :4.:; .l.. g . p. 3 i

1: -

t r -j - ~; . l r .r 1  ::%- --T7 -l i :i=.: .

5 i .

>=

.l : l-- i,_.s . _ik::l  ! l- *

- - l # ~~~~'- '* i . : l * *- 'Fr I  :! - l : i.i:.] :.rr 1 .- P.ii,, =~ir - l -i. G :. :: .; i -

1. : { i .: g ."!:hy . dh:J : q  : :- g h e e .

W o C E

!sa 'uopaas ssoa3 JaN uo ssaalg ahaaay g en b

1 3-7

Figure 2 Legend for Burst Test Plots Attribute Symbol Description 4

Stress Corrosion Crack X Cross Through Symbol i

' Machined Defect None None i l

Through-Wall Crack in e Solid Circle ,

Restrained Tube 1 l

Part-Through Crack in o Open Circle Restrained Tube Through-Wall Crack in a Solid Square Unrestrained Tube Part-Through Crack in o Open Square '

Unrestrained Tube Burst None No Arrow No Burst Test Terminated When t Short Arrow Leak Rate Exceeded Flow Rate of System No Burst Test Terminated at Low t Long Arrow Pressure Less Than l Expected Burst e.g. 3 x Operating Pressure 3-8

Maine Yankee 1994 In-Situ Pressure Tests Although the test procedure was not intended to produce a full tube burst event, the 1994 in-situ pressure tests provide some indication of tube burst capability. The full burst condition could not be achieved because there was no bladder or patch over the degraded length of tube, and because the flow rate capacity of the system could not produce pressures for tube burst without j excess leakage from the opening cracks.

i Table 2 lists the results of the in-situ pressure tests. The net section stress levels for these tests are indicated in Figure 3. The leak rate for  !

all but the final three tests of this table were all very low (< 0.004 I gal / min), whereas the leak rates for the remaining tests were greater than the 0.496 gal / min capacity of the pump. Thus, it is not possible to differentiate between a relatively large leak or the onset of a burst event. Accordingly, burst pressures are reported as greater than or equal to the indicated values. 1 The cracks in the tubes were characterized by an average through-wall cracking depth as measured by ET tests. The corrosion induced defects were not single continuous planer cracks, but were in the form of a network of cracks connected by ligaments. The simplistic characterization of the specimens in terms of an average crack d1pth would tend to discount the reinforcing effects of the ligaments, and thus predictions of burst pressures based on the average depths would tend to underestimate the pressure capabilities of the tubes. j Nevertheless, in many cases the proposed ultimate strength criteria correlated with test data, in particular when the crack depth is essentially uniform around the tube circumference or when (as for the Maine Yankee specimens) the presence of ligaments significantly enhances the burst capability of the tube.

l Figure 3 indicates a number of the in-situ pressure tests which apparently  !

subjected the tube to stress levels well in excess of the ultimate strength. 1 These unexpectedly high stress levels are interpreted as evidence of the contributing effects of ligaments associated with the stress corrosion cracking. On the other hand one of the tests (R89L46) developed a large leak at a calculated net section stress of only 26,400 psi, which is well below the strength for the material. In this case, the unusually low stress level could be due to uncertainties in the eddy current measurements of flaw depth.

3-9

l Table 2 Maine Yankee In-Situ Pressure Tests as Reported in GDW-94-92 i

Crack specimen Configuration Crack Crack Depth, a/t Restraint Full Burst Pressure Average Stress i

i (stress corrosion Length (MRPC avg depth). Circuurferential on Net Remaining cracks with {

Fracture Cross Section* i tigements) '

s S/G #1 Part Through Flaw 360* 971 Yes ~ No >4,800 psi >476,200 psi R62L27 (max = 97%)-  !

l SG #1 Part Through Flaw 360* 94% Yes No >4,800 psi >238,500 psi R66L33 (max = 94%) .

I SG #1 Part Through Flaw 360* 881 Yes No >4,800 psi >119,700 pai ,

R106L63 (max = 98%)

't S/G #2 Part Through Flaw 360* 94% Yes No >4,800 psi >238,500 psi R38L131 (max = 94%) f i;

c.a 4 S/G #2 Part Through Flow 360* 92% Yes No >4,800 psi >179,100 psi o R49L124 .(max = 921) ,

i S/G #2 Part Through Flaw 360* 35% Yes No >4,800 psi >22,900 psi R49L126 (max = 95%) '

S/G #2 Part Through Flow 360* 47% Yes No >4,800 psi >27.800 psi

• R81L72 (max = 92%) ,

S/G #2 Part Through Flaw 360* 94% Yes No a4,000 psi a198,800 psi 1 R49L122 (max = 94%) i S/G #2 Part Through Flow 360* 69% Yes No m2,700 psi =26,400 psi R89L46 (max = 94%)

t s/G #1 Part Through Flow 360* 89% Yes No m2,900 psi 275,900 psi R70L55 (max = 89%) , i o

(1) Axial Stress = Axial End Load / Net Remaining Cross Sectional Area (2) Outside Diameter = 0.75 inch; Wall Thickness = 0.048 inch (3) Stress corrosion cracks ,

i t

i h

r r

i

l l

idl~U.HTt .U C

- ,.L :  :.:. . . l .: i ~ j

• l' T'" i$ d. Mr ~ilf::: F!! d i y N , dE [. i."-l:~5,::]M: . :U. i l ;;] y .j" ' ji- .Q fj:l~{m F y:J;.:-l' j_

gi Qi ap.;. Will:::t;n --

.] ! : -p ::jg l . F:- i;.y ' :!:j .g _.li-tr :';is C K.3 . , J.- -4p...,. c-- - ,-- i -+ : l-- .}--+ -j- ;, -

3_ . l. _ , . g. . . . .j . . li- '; -

r?- " g L;; :f.al g _. p . . .;;.: .. j, :.;l:gj.,_ ::[ d .-i.: Q nf-D W-2-- p}@) 9 % J-l :@ '

"i .h k -tt.:i li =i!":.1 0i d:lF '#Ois.:.I ON r W;49. *2 :4555 @E l=-idl ' ,rln &  :-j: . u;[:1:Q== }--] M:1-3t=l~ i.i?Il c:-

rg :. l.:: .e_::-

4._.

2_i. : -l:  :.g...j. in ..p. ._ i.a g .: t_.:; _A

.l ~r: :.: 2. d .. .p:=m, .a. . l ~. g: . 00  !

. r L :.4.o:. j . .-.).f. -:;- ._. g.b.-

5 En}-=_:j; .:_ . .;._ _ t::= =:. . . .

. :r: .

L:=

.. =..p. ;3p;::j

-;,_;, ;c: =3 ,

=. _ .4 :.o-. . . .. . ..-. .h: =..- -

t i

5iEril-7.=F.::sf?j"~d) e .11:

. FG - lj:j ; jl i T.gEIQ:if:-[ - ;_._p: ;. QEiiiP:!j;Eh'  :

=_-

j ..,,;7:=a:p= p:

. = _ - :=._  :: .- l:=d. - -.:_:. =1~;.: .._ u_x2 - - _j a. .- --.._2...-

--- . . . ,2 ,

.c_ e =: .. ..: :.. l- -

= .=- :-5.:v:=-l_:!1.d:u::e =. I ..o :t :

.-- .. :::I.-.:ru.:

: - .:. = . = . = - -.=.e-: : ==e;- &- - - - -. :-

:g=--a 4t-  ::K= :-irid =. .i.l --l-,-jil : , . F-  :. l:p %: - 2:q d;=: p = M y @ =jf-U N

.-_.__.4

,g__ o . .. .

_ .- :. ~~-- r. ._ ,. . . . . _.

. . . . . . . . _ , _ . . . _ _ . _ _ _ _ _ _ r-" _ . . _ _ _

g t--e - - - & --t.... ._._ ._...w..__.. , _ W

. _:.:.:.1.

. . ; =_ . ... r.y..;.:.n.:_.-

. .- ..-s

. .... . ] " ~ ~"

M i.:

Cn

. . l Q. .::.._n.-.

r:.

,;_ . . .

$. _

... ._..__ _. . . _.:. _= :.=h.r._.: _.

_ . . ... ..  :.._ =-

.. t

__:_.:i.:d; :_:: =j _ - ._ lc h Ej:5- . .s. - y

;-i .=h: i p=f- E ,. _:=# ~ - "~ ~']' _

-1~

f \c o

.:.~.y _-.=ju: m - .:t:-

e (.D

. a. .. . . p . . i =. -

. ~

. . . . . ~: T.:::-

. ...-t. 1: .:e :: .. .:-.-.

. .. . . . _ j,:_ = _. . t. :=..._._.___...(._.

--+- - ___.:=...p___.. ..

Wi-l .j_j : ; af1] : _-fp@ 2 T- :4 "l . .: i . f;.T"-F: f=ili:i:f:f F:i3f==-!E=T !W g

N .~is b ' '.! Eti.-h1.5

_E l-~ b M 'l .1: hI:l. - .4i:- A ri M :2}$-iiE Ee D

,.y. 3 g g.= t- . a.a - .

n . r g; ,, .. ,,g,, ,,g g_ .g: 3-0 --- .

a o g y Q !d.T..i'p_.i j:__i-J: _:....2. f.% _J"._--

. . .. . . . -j.. f

..[ _1,,,,,i i- l .e i ::-l ;.l::-

2 p:: hn ,3=91h.K-?,.~' 5 o

g =; ;2..[:i:-y =j'Ei_=;._. -E:p _, m h4 :i .  !. N:p .

_::j-::. =:-i- )iG(-i:)f=pi}lp:.itiE'  ?"* cu fi# rim:5E-EMT_#E T-!-5:L: -1. i.::l=! l. "

5-' E::]:E':-[-TiW"f:h:Cir i

i" :b  ::he id:la:!w i !s.4pw 5 m

': /

P+ <! : 5: '-ili E-'i!:-EE-tid !/I fEf:

• T.

.2.=_=_..

. = 4. _=_=_ . .i.. . .L. = :=_u.:... 2:i.n

. _ .. .. ' i..  :

I. -  :;: .g -if. i..r4:.__.Ei_: o._t=_ _3l.:. q:. __.

o m ig :iifii+.i.edEjizj; f2::f.. r ~ " y

= :/i; +.:p 4 - 1.:-f__3 .i --

l : g F--L. :7:p

- r. wr2m::: t : . _ g

.= . . ....l-.. w

--- l - l r - - ) . ;- - 4 i -l -m--

r-1.--, p.

. -r iii , -ti: :f:~- }=i 3.jsp j- -j l.; l 4 y- l-4.l,l. 1: 4 m

h  ?"F 5 MI*[?!lir  ::lN L d l Il .

• i 'ii l" Tl 3 m

.. u. . . - d. . . ..h.E._d.

.: Fi: f::d. :. H. . . a.. l. . - l .. i j .L ..

. . i .1 -!.. i .1; m

- ::r - ~ ::-  :--~ 3: 4 g

.. iw. :f.: . 4 .iii. i 1t -i -l i -l- , i:-ln.af. . !. l . q - L

+- i. .

• e. r. . . _ u

..: r a

. .=_ .: . l .

1 *1 ric. - -U -W ii - p y

.)?.} .;I j . .ti.! .:.j c.J.i .: ~i. "

.l .l. I:

, .l l 1

.1 I i .' l . e' ,#

11  ; -} 9 :-n l i l l -i l: i i

. ani .; :.l4 .: - .4 . I 1> l } .l. .I 9

l ,c,j j j T- -

i l i l l-- ! - lrl.  !.' ; ' l a

:e 1 .: : lie.=_: .c.p:i4. .

.lm I. t ..l .% _t :4 91.b. :_ L i:_ y

,q.; 4 . i : a. t ,:n. - j . I- l-  :- u p .-l m.I

,. i l - g

! el b- ,=.r: r_ . . _ w. i '

i l i 1 - l 7 .:i r -. l .

-i >'

' L : ( i:1.':. , ".::: :- l . i r. "..i r.n . {  :]

-i. : i  :.1 p g u - g.:s h4 H a q . , -p :.;i y y g

o e e m

- o_ m e e m

Isa 'uopaas ssoa:) laN uo ssaals 88caaAy

[

cn e

b 3-11

Gamble's Evaluation for Millstone Unit 2 A letter report dated March 28, 1990 from Mr. R. Gamble of the Novetech Corporation to Mr. J.F. Ely of Northeast Utilities was provided to Pacific Northwest National Laboratory as part of the material to be used for the Maine Yankee review. This letter was one of several related items dealing with circumferential cracking at Millstone Unit 2.

The Gamble letter does not report on any burst test data. However, Gamble on page 7 of the letter recommends the use of a conservative value of flow stress of 0.5(o + o support less ,c)o,nservative assumptions.with the statement that there is a l In making net section collapse predictions of burst pressures, the flow stress parameter is often defined and set in fracture mechanics calculations to an empirically based value needed to achieve agreement with the relevant burst data. Thus, in one case flow stress may be taken as 0.5(o, + o ), whereas in other cases another value of 0.6(o, + o,) may give better agreement with certain burst test data. However, the use of a flow stress as large as the ultimate strength would be viewed as unconservative for general applications of fracture mechanics.

In the present discussion of circumferentially cracked tubes, the calculated failure stresses can be interpreted as the apparent values of flow stress that would be correlated with each of the burst tests. The limited data reviewed by PNNL would support Gamble's statement that 0.5(o, + o,) can be used as a conservative value for flow stress to predict the burst of circumferentially cracked tubes. In most cases a flow stress equal to the ultimate scrength would also provide conservative predictions, although the data i Nicate possible exceptions to this rule.

Framatome Burst Test Data EPRI NP-6865-L (June 1991) presents an extensive set of Framatame burst test data, including some tests for circumferential cracks. The test specimens had only machined defects. The burst pressure data are summarized in Tables 3 and 4 with net section stresses for burst indicated in Figures 4 and 5. Most tests were for through-wall cracks, with only one test for a deep part-through crack. In general it was established that tubes with circumferential cracks with a/t < 50% did not burst at the location of the crack, but rather at another location in the mode of an axial split in accordance with the hoop stress in the tube. Nevertheless, there was one test (Table 5-5 on page 5-19) with a unifonn depth a/t = 77% circumferential crack. This specimen burst in the mode of a double end break at a pressure of 7,290 psi with a net section stress (axial load divided by net remaining section of the cracked tube) of 116,600 psi versus a stated ultimate strength (room temperature) for the tube material of 107,300 psi. The full circumferential nature of the machined defect (and lack of pressure induced bending) was a likely contributor to the 3-12

relatively high stress at tube burst. This one test would support the Maine 4

Yankee assumption of using ultimate strength to predict tube burst.

A number of other Framatome tests are cited in the EPRI document. Interest i

here is limited to tests which included the effects of lateral support to 4

1 limit the bending that could occur in the cracked specimen. Table 5-5 and 5-7 from the EPRI Document summarize test results from Framatome experiments.

These data are reported here in Table 4 and Figure 5.

1 4

il 3-13

Table 3 Framatome Burst Test Data as Reported in EPRI NP-6865-L (June 1991)

Crack Configuration Crack Crack Depth, Restraint Fult Burst Pressure Average Stress on Net Length a/t Ciretaferential Remaining Cross Fracture Sect ion'"

Part-through Flow of 360* 77% No Yes 7,290 psi Uniform Depth 107,300 psi Through-watt Flow 250* 100% Yes No 7,200 psi 82,400 psi Throup-watt Flaw 180* 100% Yes No 9,700 psi 67,800 psi Through-watt Flaw 250* 100% Yes No 7,440 psi 37,910 psi Through-wett Ftau 180* 100% Yes No 9,290 ps1 67,100 psi (1) Axial Stress = Axial End Load / Net Remaining Cross Sectional Area (2) Heat NX3332 0, = 50,200 psi o,= 105,100 psi (4) Outside Diameter = 0.875 inch; Watt Thickness = 0.050 inch (4) Machined Defects cas i

i i

w. __ __ _

.. -m

'iF .l':i:C T.:.. FT ., .  :. l' 'l i T. Tl: C NT l :.r: jlf,:. ; n.rl : :1 : g S:sl7 !hlll.51Mi li~ ;

• i :1 'i ~i - l 7. ! - T- - }; li!:sf Ej :T-l l - -

. f **.l'.' -

[. ' :' j * ' .".' * * '

il* .'l' *i - f '; l.:*. - i' .; :h:* ;j-I".h.l ,.) bf'  !-li(.

W: 6 .; _..

-?.--l---- 1 -- r j -- h c.h -M . _ . . , . . _ , . b Jht .5- ;;=.-

55 2 T4 Eft'i i h. dj: .

.=l  : i :. ! -:l .! 4. . l':;u.!- Ll :._ : = E _:-: .: -itflb t

-'i5i-5E::N:#i:il$:4 9 "i: . _ i l- ..l: .liv!. :-j-:i.:i . :=l FEris :srhjC4i#:i.i :.lm i .

i -

~. '- N: i E P-l ~ - I *  : *= f j,,i,

.Fr. - b.. .dE;.;f_8.:.T._:, ..9_L:.

.- -.:-. . r.- .. .]-a .I.+.

+

. .. +i., .n. . . :_=3_:.:!.

r .!.. H. r . ..

i..:i.r.j

.- f.l.!i.::.17.=.f. 'r.. .. i: - M a ..J

'T }':5 .M-1:lN_\ :NI Q :d 1-- --Jiha -5: MIN l _

l W=M2 f-' i ii'- ; d

--': lie 4'XJr.i= -b6.i:.f iitE.-j .T_ 1- -l , rja:4: S .- ii: . : .T .l-:-i.T ' I-~2 i-=Iis : =) it'ri=Uh-

-:E- =%F5_i(_"=:.::'=]jpiJgup ..l: a. . .:;.- : t:.

i

4 .i .}_Mi- j ==;;=j:Ej=.hsj: d:.r_

===p---r.::.r._..t_ i .. ..-

.e i]-: -_ ...m ..;  : t . . -:: s : .:.=_rw :E=._y:.. -- ~; . : 2 -u w- . .r -- n .:

-- ~. - :z . -s Ri.:ji4=. .Mrb i =:.l:-l.- p+ n ;:. . , en

. :.l:: - : .:j: - : :!=.i:j=r'.g _,,,,__-l::p=:- --d - -0

. . ~ . . . . . . . __ . ... ._ . 2,::. . . " - . L. :. E-i E. i:i:-E.=i- ~ . ~m::7..:..,-"._._.'._____.-..

5 o

N:IINJ.!, IiM.~d f M i}:"ET::l t'I F: :M ":*. fine.i.:-2 $ O 3--$ 5} N NlNs 6;fWit tea -- sleif 9::in:PE = 1 :. 5

-1: =f4 ?:t.isr- C:. .~C ---~T. = . == 9 3.

F Iiii- = iiht".iiF

n-"iH 9;' .1=t - i: ici:= .19:15P:r?:: := = ;.~ - ,=ET-; C

. i']-i; Uj.: 'nT;,2 li:p':. .:J:0h-l/ *:t :.:~ a

.: ! - n r a- 7: E:Eh'3~:i.9 3:i fiji:Wds-4ksI J.  ::i]Eilf:0 E a: f. ti4 - l 2.i. :i .  : .Ii E. :1.J.i:.PWd=E.A4-ils n 8

.:ir- j=;._. .q:.= r-954iiJn 4 ;_..pi... . ;... g : ..; .i. . _.a._.4__.;: - - :piuy.4_v= $ y fi-irid- Ed;=g..MiM. . . .

tE-i'ir"_2l .W '-l : F 8' "F-w Tji?!@:: SiiE. 'h=== *Eli:".il.: , .4 i : :! T.i::

_.:--i: i'~ 2:3i4} ET=- 4=- M cg

-#& E";-lh=iLi&~M; f+!=l :=i .4":1.

Dr =hP :i-. l-i:54.=fik :ufE' t C $ i TM'iA ::[d- $-IiM: 4" !4 ~

i n vi- 7 12.. :ita =wiEl _4:Il-W24Jid=: O e I- "h IIWM:53597 -5?i[Wi'#3- 4

' T..

TE$"~J' -E5-?:1-_.i:- I.-M-- 2~~_ " ' Ti! '!:  :'i T :l - M #s%5-i5':H-' .?asd'i$ d i::2i:.g :i:'.-3:jgE!y:": '-

M.-  : :'i- m m.v - .t- :J : . !F

! E

' :4} iq-33 :{ :- 9t;;! (

2 i.~'T.~.i l2:I.i:ii_.-.r.r =}ii.l.: ] - 9.. .. -; : - -.- i . l - i . -!. w ..:: - i$; u.n: E

H-Qy 't,j_ Mj: ' ]:~. -F.}:i--[- -

1-l ;

.j-

- . ).::f: *

-l% gg

_. . s -r j.=4-@,-[: ji - ;j.--i .l l  : .i. -i frl :7{ ; # g

! E- ~9:l2'_d hd li $ii ' 5 I.~: '. 6:dl -

i e! Y l2 2..dsi: "M

--! - '  :- j ; -!-[~ --t-i: - ---- J: :

-1=i -i -l -l- .].i: rl :, ,ySg 7-

'O

-: --

• 4 -i.. . + j.

.; a,.... . .

1.  ;

-J q 1: -t: : -p H:. g- .

g g 8.:  : ; ;.:.- i.y i .-

. -l -- -

.l . . l . , :l . .

. . l - . ; . - l . . !. .

t . .- o y

i l i:1 '-] : .] - .l i l l l i . . . l ." l  !.

J. ..a:

.). 4:  : 2j....l-  ! l i -l. i. l. l .i.

F-1- [ r -

l l l- l ; - l !:i:- l. i-

i. nl .l .: n.l .:n_;.

l  :.; -

ML ij. - l . ; :. l ,

. ) :..i. l: ; : l.. p.: .l.  :.64. p s ii.. [ . t : Ji- .l t  ; j- . j'l! .

' l 4:Hp -l :;:.  ::;-

• h .:l .l H _. :i i - .ir I. I  ! l h T-i ~ i'.i 4~ - h i 0)

E

' L i f i: iM : a:2! ; r :g r. :- j .- . :] :: i

. i. - l 'j~u ' {.: :hi:/dh f . i rj =3 '-; o e u to C C e W C

- C

- tt) o to L

u.

Isg 'uolpas sso.O PN uo ssa.as ah.raAy <

u 3

CD r-b 3-15

Table 4 Framatome Burst Test Data for 7/8 Inch Tube Diameter Test Crack Restraint Full Burst Pressure Average Stress 1 Length Ciretaferential (Degrees) on Net Remain {ry 1 Fracture Cross Section" I l

- 110' Yes Failure Mode 9290 psi 48,700 psi Not Reported 140' Yes Failure Mode' 7805 psi 46,500 pst-Not Reported 180' Yes Failure Mode 7188 pel 52,300 psi Not Reported 200' Yes Failure Mode 6629 psi 54,300 psi Not Reported 250' Yes Failure Mode 6085 pel 72,300 psi Not Reported 300* Yes Failure Mode 3395 psi 74,100 pel Not Reported

- 180' Yes Failure Mode 7438 psi 54,100 psi Not Reported 45' No Failure Mode' >12,090 psi >50,300 psi Not Reported 90' No Failure Mode >10,143 pel >49,200 psi Not Reported 180* No Failure Mode 4365 psi 31,800 pel Not Reported 250' No Failure Mode 1234 psi 14,700 psi Not Reported 300' No Failure Mode >426 psi >9,300 pel Not Reported (1) Axial Stress = Axial End Load / Net Remaining Cross Sectional Area (2) Heat Nx3332 a,= 50,200 psi a,= 105,500 psi (3) Heat WB 578 0 = 52,200 psi a,= 108,500 psi (4) Outside D,iameter = 0.875 inch; Wall Thickness = 0.050 inch (5) Machined Defects 3-16

N: :iC'"I'i.:!: 1 : i : , :l- H p . 9.. l .;.l; .;b. . r. . : -a ~ ; .r: --

C*

=: .

, _ :..,. .. . l .:-_

l.

m.-l..p: . :1

_n - .,.:. .. Q. .: . .;pp 7- p. g q . . , =,,,

. 2 -- ..

3: I

=~: I - = l . , e_

a.

7- r1. . .--

. 3. 5 "..

t.t- - :j .: . j. .g q,q ;..; 4g

q. q:. - ._

5

.. .g ,, ,

. _ p , _.._.

j ., :

i- . p --l --. eq a.. . . .

--~.. p ...4--. --

ic- .3-.- . . _ . _ z  : -

, :. .R . ,

~i'-t'id_ -E:i! _.t@l.9  : p . .a :. l .. ! . . H. j .. . .c..

=-i'. ~"

....-".l..f.

. _-- L::_. .... %._.r;;.,:; : l :; :

~

.:. l;u.y- m. ..gj. ..ag

3 j.;_7_;{. . . . _ _.

.=g.;:.1. -..

. ~. .. r .; . l ,. i . . ..

.. . ... ..=.u...;g_

_"............_;._.,_j.. . , . 3. .:j g,. ;g

. -i: j ,_ , . . .

y r. r ---  :.p;q.:+.  : g ,:=q.,,,.q :gg. . _ .g.m . .

--r _- p ._ . _y_. , ,

~W 'fiil'~:I UDD~l=..h

- - - I-  ;[ .T"- ~.L----  ! }.: -.k _i :T:. - m--.r= .h= . 2.) .= .=: l-[. _ .n . '--1_c;E on.

.= ==. p::= a -._,3= : 1. =. . _. ..y ;k ej.p q.

y : . L_ t=.: ~ 4 __p';p. g_- .,=3 ._.:= n ; : : ; __ c.;.,; 13: .;. .

t i=r 2 2

m

- ["-]- . i. 'M-~2- :1=k+;.I. tT:  ;;:-" ' ' - ~~~:#'-~~"

!!I i- :b :n

= r.=4'ETdEEE4 :?l#F

_,_ ._J=. _~; J.

7.hil  :::h--l"T+t:--

=_- e ..4.. - .. . . .. J. . . j.p. -

..i. l 4. _.ta_q. g . _,_

-..u.. . __ ==. _=c. f e __,_m . . =.a . . _. ;. .. J. ,-n., ,_

n- er . .. '.y

== = - n r

_--t: _. -.e---,s._-..;-.--_._.g--,.:;:;

. _a:_ , -_;_-t ..  : =.=:==,2:i:-ft,_ .

hg pn.raq u:-  ::=_;

2 -- 2.- - ,;--

- ----= : =

n.-r-

.r

[.. [" E ~d'I r = ~.c r.ui --pi::7.

2* "!!# l - E:- _. Et=1.~

~~- -

+

M ' .";

~ - - -

E E.=i_ ri..:. m:-4.

- - - - .; .t

_2 : . : k.: .:

[. . ==]

m _. .. r.. .. p- +: 4.

-~

. ' *j g .

f

.j -

k=:::._

. - qe~ ra:_-t_ _-..=.-.}i . _;_._ =-. :-. .:-

..~._

]--4'7~:I~m-c:1 '=-= = ;t. .=. ,.:=c7 .rF F  :; ! m.-hh; @

- . W.. : . . :=. t =

1:p. . . . .:7:- r=.-

...:- p. .--. ..;p 4:. j r ..p

y..:.$~;

a::

~M- '---- - - ~~ .- --- - - n. =.. =_- g

j ... w -... -

.- -..., .._ __ _ _ - -=.:::._: ,,._.

._.- .~1... .. pj::

.. -.dj

,, . .. L..~

q 1;  ; g,. j A . .t .-r-- --r-: -

" .7.: . - 4=. .].  ;. . y ygn.. --~ 3. ,p. - -: 3. o

=,:  : .' E M :.i. . . . . , - -. - _= 33 y, r p_:. n;.; ._ r- .l a .. r .

- m, -b.-...- ..

.'p..,

, ..p _; . g .i.

.m.t -- my: :-- - , p o

U 0 9:? fii :- :h'i-NMy . ;.l . ,[ ,p ,,4. g=T. . ,--~, . ~.. j _1.

E h.. .

_ . _._-j 2. @= . ;E. ;, .

r-.

, ; .~.;.,.

i

=.a _ n._-.=1=..

... .b._;

.z,

. .r._ ._m

=.

.,r4

. . . :r;.

... n. . l... ,;  :.m n; ::.]8 _ p;..

.,_,.

;g:,_; g

.r--.

t.

w C

,E a

-4 -

-- ..= . _ - - - - : _.:; i. -. s_.-9.g=g.:t

lj :  ; p.  : ;.;

- I =.nd . . . ; . . ,; .n---

_ .d-- .

g v

__ - =

r._a;:-- :=-t=s . .=c.= .q :.;-d.r. : .e ;-.

g.y4..'. ..9  ; . i .. . .i. . . _

_.:; .__cm L_

y w= .3. p=? g _,_f . .;m ,-C -l_ .[ .. .--

~

97 3

. t- .- _..g

- - - . :r-t ==1 ==.-_==:3[_25 G.-p;_- .l -

q . ;p: 2 n.3.:

m =- u. =:- =.,.:- : -. u =

m==.--  ::.: -

-. z-=-- . .T- t: . . ,1-.t_. g :- ~. p n.:r 4. . .p ... . .x , . , , _;,. . . ; .

A

u =:, = .-1 -

M*

g

=.u.. - -. .. :- n r

..._p-. . . . .;I. .;.. _. m.c - ..- . m ; 7_ . _.,_~.

---

~ =r .: :: ,._-. % . . ] ..;;;.= . p.. . .

- v~ n~- -r-

i .rl.g

3 .  :. . -  ;,:. p;_- . .;. ,_ l p g p g: ,

_ ,g _.l.; ;

.. .i 3 E Itl 7;.;:: g.:r;=.

ra- ..-f; .. A: ..p  :

.. . ; .g g,t l.

,- _ .c,,o,,,

- - . ,  ;. .. .; . .._...p .

f'an

!.- - _f, .]h:. .i: J hi=lf: :{' .- j. . . l,,, ,m ,

t 9

. _,.. _ j._

,,.--  ;--- F. nr:

q--+ .;-_. p . .y . =q ..; . l . 3.. p, j-... -

- 4 g

+-

2: V:i =tWF - : s._t.. q- - + . :. .i.i . .j ..j . .

.s - 4. . . .;

. . ' : ..-M y

^._ a. l -u ;4.3 4.i :.a :.g . : q. - q : .; l ,l , .,y , . .

9[-l . .!l i. -q = p _3 , ..

p i l ,. I l i ..i:

3y O

e e

J-

!...-b i . l . J.

- - .:.T. .' . . . . .l : 4 L 6. [ l .l w

_; 1 a .:V ~.: .= j. .lm, .,. _

g

u . l .. . . l m . ! : p. .=. _ . . w .p. .:  ;.,

s

. i . l.. . t _. ...._. -. . .

. u. . . .t.  ;. .<1.

u F. ;. p ., --q..;..

-l . 3 . -

. ..,1 .

! .i l l 29 .. l . g,4 m b -

d. 4

.! ].4 l. [, . .. . ! _ . t $

q-.. . .I14:.. d .. F; - l . .:. .l .; _.._. a.... m

u a

= ,2.. ( = n.

pg 1 g g . ;; g_., ,_6; ,.i,,,,.,_,

e o C O d@

LO O C C ~

- m ro E

rc Isa 'uo! pas sso.o jaN uo ssaa;s a8eaaay i m

W 5.

3

.C.D r

b 3-17

The Maine Yankee ultimate strength approach, when compared to the Framatome data, appears to be valid only for the one case of the 360* uniform depth crack, and is unconservative for all cases of through-wall cracks. For the through-wall cracks, use of the ultimate strength would predict burst pressures much greater than the burst pressures of the tests. A flow stress of 0.5(a, + o,) = 0.5(53,500 + 107,300) = 80,400 psi would be an appropriate assumption. This flow stress would correlate well with the burst data for the 250' long cracks but would be somewhat unconservative for the 180' long cracks. It should be noted that the " tube burst" pressures as reported for the Framatome tests for the long through-wall cracks do not correspond to full double-end breaks, but rather correspond to large crack openings with large leak rates.

Millstone 2 Burst Test Data Pacific Northwest National Laboratory reviewed information provided to NRC staff during a February 22, 1990 meeting with Northeast Utilities that discussed tube burst tests on tubes for an evaluation of the Millstone Unit 2 l steam generators. It was assumed that the dimensions of the Millstone 2 tubes '

were the same as for Maine Yankee tubes (0.75 inch outside diameter x 0.048 inch wall thickness). These data are listed in Table 5 and the first test results are shown in Figure 6.

A-number of tubes with machined circumferential cracks were subject to burst testing. Some tested tubes did not actually burst (but did leak), and other '

tubes with flaw depths of a/t < 50% burst, with axially oriented failures.

Only two tubes burst with circumferentially oriented failures. One of these specimens with uniform depth EDM notch (a/t = 75%) failed with a net section stress of 120,000 psi, which indicates a burst pressure that is consistent with the Maine Yankee ultimate strength criteria.

i The other specimen failed at an average stress level that was clearly less than the estimated ultimate strength. This specimen had a flaw of variable depth with a maximum depth of 91% and an average depth of 47%. The failure mode was reported as a leak. However, details of the test procedures were not available for review. Therefore it was not possible to determine if the flow rate capacity of the test system was sufficient to produce a full rupture once a large leak had developed in the specimen.

3-18

Table 5 Millstone Unit 2 Burst Test Data as Reported at February 22, 1990 Meeting with NRC Staff Crack Configuration Crack Crack Depth, s/t Restraint Fult Burst Average Stress on Net Length Circumferentist Pressure Remaining Cross Fracture Section'"

Part-through Ftaw of 360* 75% None Faiiure Mode 10,190 psi 120,000 psi Uniform Depth Reported Not Reported (EDM Notch)

Part-through watt Flaw 360* 47% None Failure Mode 8,500 psi 82,400 psi with Nommiform Depth (Average Depth) Reported Not Reported (Reported as Leak)

Average Depth = 47%

Maximum Depth = 91%

(1) Axist Stress = Axial Erid Load / Net Remaining Cross Sectional Area (2) No data reported on a, or 0,.

w (3) Outside Diameter = 0.75 inch; Watt Thickness = 0.048 inch 8

(4) Machined Defects e

6 F

l

l l

1 I

a 1

I l

I to i w l Tib .:1. "1 ZII:: -j i *i : -1 :q; ^

l ?: l~l -l. 7- q .-- . :. 7. . :;.e n . ;l ; C.

.i? j"l:i-l:3 :T}Qi . :"' : , (( gl."l 1-..  :.l . Etic. ,:;i[ e;j;hl.-M g:$ r . ,,

N

- =l ---it.q,2c + n :

= ..

.: ..p .! -  ;.. gg .b.. J.--i g ..g...,..g z

@ :i M ., ,-  :

. l -- " : - r.- p- p- p- ;- .,46 . , . ~J.,.,., --.-.; H .;,- -h

. a. . . w3.:

. :.. ._u. . ar:::._:. g. --. . g . .. ..

l: ,l._. l .J..;l. - 3::9 . _:j.:,, ...:7;,4. _

.. l - J..-

3

.1...M 5"5I :# :: i: 5":iititi .W-;

i .[- :.l:: pi.jp: 2 ---pi , :,ij::j.~2:-ig.: ta.r-i .p:pfl,q.: , ,

m D: M -@i=:: 3i I4pl=.p.. l.: r l .::i .: l .1. -in.p_ . ..;;ji:. ;, .,1:1.=:0._i-. : . . .g. r._.j . .

t ..jg .;g. 5 I

-- ---rg-:jij. . . . . . ..

p

._gr .g : ;~l ) 6 ;-l ji r-7

'4jy;3gg:g_yg={,.; j. . ;;.jr._ CO o

CEf%g.15 :: =. -. dr. .!.:j - ,.dr L l. . : .. -l : .j .:

o

.,4. _:;Q: 6;E-E4."7gj;_--i2:. }.:;;i.gg .g_: c' E

5 }YT=%:iigQd:-p -Si [...r--l : , :j : -T4. :-Q E.lr.iy=-}:=.::.p: p-.q;.3 j =:q:- ' o m

r :.-4 r az.u=.:

.-- ,- j-- :- --- --_t .2 -3....

_ - lr_:.,;: :..

.. b - " li*. d  :" A -

?i.lli ..: ::Tff.'_::_-u--- " . . . . - J : :.::. , .7.J:;*d ".

O

= r:L- :--

.=: _--if::in 2".9:T- }-j . I- ..h.. .. .TJf. :h.f. .-. _ .=_3_:-.___-_.__._--_

=._=_ . .

--~--

-==+

. . ;2: _ ,

@r-j:Q :%-ti :i.l:i.h- - ::: : . l u t- q:;. .j . _ , .g ::Hj: := 4 _ g .%... g ..-v - ed i

==-E = :$$~E 6 ~.E.f5 "-" ::=.kr.:i. F.!:i9.Ei2r. - - " - ---- ~

(

M=-l="f=lMit21 Mii-Mi=1 #:: 1 :"r  ; in 5Z-3.iE_5jk.- {

5--iiS; 1:Pri -.- -liiiijiFr .. :p".F-i . .

.fi: . :ha. :.:.,.fi.:_::[..-._' . .---"_."_-__.-r*=_=.=- e  ::s y

?9.!.- ' - l i.- .i. .md- :i_d4:.E .-.j-9.:_u qs.j ; =.i.. d

. .ai . I ..;j . ...- y. 22 .; 2. =4 . ._ ._.;.=.7..--

n .:.:.. .r :_. .ca E6 :J j:i .:T,':: 2Sjaa ::.P-iri j a E -l . a 5  :]:),: .; 3, ;; ;.g p[g::t: -;j= =.:3=_ :_:- :p u if air

.L:: ir.-h. ..l: .l d. . :1 : - il.:il~ i;...).'_iF:Jiiit::3

- --- 1~ r.-Eii n W

. .. av

.J C.:.=. j _dj . .. ,. ._+.. i ..; c;;.; u J . ' ,,,l ..;;,..

, ; .. ,,p 3 i, . _,;,;;,;,,g;;. _, y_m.73-  :---. 2 y 0

.(=

- f: Jif 7.iih. . , 4b .d. _.P:E: 9 F.al- M . -j -

-i !!i 2

i l.p:r "7- ': .:.:i:.d -~ g.. 4--E es W

. . . m u

.=*@_~"_ ..._ . . _ _:q. .>. :. .::.::*;, .

. . . __;,=_ _: . . .' :!: ; .

., -- j , g ;i - -.M_.:

L4*.:,

.:.2 7 ... r1;;." . _ p. . ._ , ,.. . ._. .

%a L

.;;,3 -- . --.;..p...

. . .l_.,..

T

'---'-=r-i=1g-- .}bb$Y Y. -: .

-1"i$4r.j . ; ..

~

Ii:MN:bEdNNM O h 2.

. .! : :4 :.l .': '-h: ::E.-[d ;:-f.i:;.-EF:I--- 9 CE w.

= . =5 -.- __.

-.....y._=.;: u. .:,_ . : - . _ .:.,.:.  : e :- --,q

. p %. .t . , . . r =,., . : - ~;.p=- .: .: .. -l-e g r0

-3,---. ,---t :=r._ :p, P. .:-e  :,c -::l 7

=-

~:: v::".:r r- 4- :1 4 :q"::: l- : ,Ey j. .: ,,. r.pi--j

.. ---. r- _a :=- -- -- .:

.: a . l -. .'_.- --

~ ..

--r--- .-=,-

._:p

-.._. y r-.-m - -- . --!. , 9 ..

. i : L :. . .;.7 . . .. . . .. p=:y:n M # ES IIWMi ~:1 ~dii~[ 'l j- -l ,

JA j -- p:. }.::t. .q :. Q 2 C: E- L: lait!_t:-M-  :

l:--i: - l. .) l , . { - j j :np..p;g .

i &

_.

i.1.;.p% - g+ji. =g g.: p. 4 .p p ;. ;. p. . . ... q .t. . y. 1.:

t p o

-- i .- - --i - [5 j. :1.. ..i ::. . .-- j:q ml=q

-t J . q . .l. l ..l'..;. .;-l , ..,... y

-i-~.. -l - - i p.j :9-ii j a

-- - . J. ._ .u.in_.

. _ _ . [. - . [ ._, . [ .l . . p ..j . . g , ..:.r..

. g .. .. i . .; . . . . : a ., . . . . -

i

,.. g m

g

..):'.l 2 : [ .4 . . ;.;.j a.ai.T.j.-l : .u. l. : P.: . .;

.l  :

.g ..) .; 7 l ; :i.; 4:- f y

I1 -f' q l l i ..l.: } ; I ,

• '- 'd.  : .L ' ' li . . l  ! l ,

! [ [ l ; y

} l

l. .l' l j. p [. ,.. . ,. .:. ...

g

- L. l .i-'  ;

: ;b. 7 4;m . ; :. l _.i .. [ ; .  ;._; :.l.. .i ._ a.;jg. ; . D n!.. .i: 0- -

j - .1 .  ;.[.i ,

.l -

.. . p . [. , [ .g,l 0

h k i ..ei :j_:. _ r_ .:i. 1 i .> '

J ..- . ;m r I . g 3

.L.. , i.-iJ.  ; t u.m- o- i.n..r 4 -- L . , _ _ q. , d . g ;;ie sq : 9;; q ,e c

e e C m

e e z e

!sa 'uopaas ssoa3 jaN uo ssa.as shaaAy [w b

3-20

i d

Belaian Burst Test Data

Appendix A of EPRI NP-6626-SD (March 1990) presents an extensive set of data

- from burst tests performed for Belgatom by J. Mathonet and P. Hernasteen.

i' Much of the data were from tests on tubes with circumferential cracks, with all defects being simple through-wall machined flaws. Part-through defects were not addressed. Both 3/4 inch tubes (0.750 inch outer diameter x 0.043 i inch wall) and 7/8 inch tubes (0.875 inch outer diameter x 0.050 inch wall) j~ were tested. Most of the Belgian tests were performed with the presence of simulated lateral restraint corresponding to tube support plates and flow l distribution baffles. -

1 c

j Tables 6 and 7 and Figures 7 and 8 present results of'the Belgian tests, and i

indicate the level of average axial stress needed in each test to burst the tube. The Maine Yankee burst criteria would imply that this stress in each case would correspond to the ultimate strength. It can be seen that this burst stress can be as little as about 25% of the ultimate strength. On the

other hand, there are a few cases for which this stress at the time of tube

burst was greater than the ultimate strength.

4 The following trends can be noted in the results of the Belgian tests:

l 1. The very lowest levels of failure stress (as low as 23,400 psi) as

calculated without consideration of pressure induced bending

~

j corresponded to test conditions with a free _ span with no restraint of tube bending provided by the presence of simulated tube support plates.

{ Given the restraint present with all tubes for the in- situ condition at i Maine Yankee, the low levels of failure stress are not relevant to the present evaluation.

i I 2. The are a number of tests for which the failure stress (as large as 164,700 psi) significantly exceeded the ultimate strength of the tube material. This seemingly impossible situation occurred only .for tests i with restraint present, and could be explained in terms of friction at ,

! the patch that sealed the defect in the tube and friction at the i locations of the simulated support plates, j Considerable care was taken in the Belgian tests to maintain a seal at the locetion of the defects in order to test the tubes up to the  ;

pressures levels needed to achieve the substantial crack openings associated with tube burst. The Belgian report acknowledged the reinforcement of the tube due to the presence of sealing patches. The effects of sealing patches are also addressed in EPRI Report TR-105505  ;

(1995).

The report describing the Belgian tests also indicated significant  !

interactions and contact forces at the simulated support plates, as the

3-21 j

, .. , , - ~ ,, _

support' plates restrained the bending of the tubes during the tests.

During the tests the contact and associated friction forces prevented free sliding.of the tubes through the holes in the simulated support plates. Therefore, stresses greater than the ultimate strength could correspond in part to the axial friction forces acting on the tube.

3. The Belgian tests gave much attention to sealing of the circumferential i

cracks in order to attain the large crack openings and ductile crack extensions associated with a tube burst event. Without the presence of sealing patches, it is expected that the cracks would have leaked at substantial rates for pressures much lower than the reported burst 4

pressures. It should be noted that the tube burst pressures, reported by other organizations, may correspond to an event that would have been considered in the Belgian tests to be only a large leak. Therefore, the i burst pressures from the Belgian work could be greater that those reported by other organizations.

4. The relevant tests for the Maine Yankee evaluation should be limited to those tests with restraint of tube bending. In all cases of the Belgian work, the average stress for tube burst was greater than the yield strength of the tube material. Taking the material flow strength to be the average of the yield and ultimate strengths, all but a few tests sustained a net section stress greater than the flow stress before tube burst occurred. Therefore, the replacement of ultimate strength with flow strength in the Maine Yankee criterion would result in a criterion that would be consistent with essentially all of the Belgian test data.

I i

3-22

d l

Table 6 Belgian Burst Test Data for 3/4 Inch Tube Diameter 1

Te c Crack Restraint Full Burst Pressure Average Stress Length Circumferential (Degrees) on Net Remainin,g Fracture Cross Section" 1 300* Yes Yes 5542 pst 120,600 psi

2 300* Yes No 7570 psi 164,700 psi 3 300' Yes No >4704 psi >102,300 psi

, 4 165' No No 6292 psi 42,100 psi 5 165' No No 5850 psi 39,200 psi 6 270' Yes Yes >9143 psi >132,600 psi 7 270' Yes Yes 8673 psi 125,800 psi 8 2 70* Yes No 6o73 psi 125,800 psi 9 270' Yes Yes >8893 psi >129,000 psi 10 300' Yes No 6409 psi 139,400 psi 11 165' No No 6497 psi 43,500 psi

, 17 270' Yes No 8967 psi 130,000 psi i 18 270' Yes No 7291 psi 105,700 psi 19 300' Yes Yes 6218 psi 135,300 psi 20 270' Yes No 7232 psi 104,900 pal 21 300* Yes No 6482 psi 141,000 pel

, 22 300' Yes to 6570 psi 143,200 psi (1) Axial Stress a Axial End Load / Net Remaining Cross Sectional Area (2) Heat 70699 0 = 50,170 psi o,= 107,300 psi (3; Outside Ofame,ter = 0.75 inch; Wall Thickness = 0.043 inch j (4) Machined Through rstl Defects 4

d 3

i 4

3-23

~

ii- :q.

' :tfrj r"- ,.:.'

~

- i T .=-  :

0 _

i : .

its.i ' . 5M@ ..r 1,: :s" 1

'i j-.g i#

J '

~

.e, U~ . i. Y:

L' . l <! c;G .i pg .

g.U 2

["~

3 1: 4 d 2- . T.

ne

' . =..a. " :..

' e.,N hQj,y;.

y  ;

g" g!li  :

i 1

I; J ~1 l' ltp k:2: .

- ., i 1]

@ ;; N 3 'I;,,'l! a-~'R

". .V I^:~ .

.' [:L. i r :: : -

l,"

d.

h,:. u T.DJ' F ' i .i;i Ny' l; t

%j Ig{j.;jil I

! j li g s. $ ,L~ . - .

! ,  !",A -

.~ - . i. d i,l i, j';=:

ij,l #F ih'igrni ' j.ghp da i : . :. _

,.Y b t'.- i' . "; * . i i! li , _

l 5

g {l; _

I i

,,T

-tgr"lj4 . i

. fI' s

q*i (I yi, 8

--@p jM: LN ik'a 6 p :.1 i - . T3.il 't t.,H'ipg E.,E M i; iP, ym,g. 0ii

.-l '.

.. c.e:!n, ny 0 i1 d4(!i i'tif !

'y l

d, M l*

U!

!1 ('

i

,e T : ' -

i 7

.hE  ;'

i  :

• : El.ha. lim '~ }l.

.t li, Fl!  ! dd; i: ,

.;[. h g: t

!lQj]j.,,

l'a 9-; .;V hlig-j _

a.I.- ' . .

i [ :.

j gg f 1  ;' .,.

i$  :-l.

-1

.

i .

li1

, '! d;!'dl'b i'

!;;p . ; h'i

. 4 '

o

.lj lj A  ;'t  !.' . .

.t,:

~ nII -

-.."l :  !

k13j  ;

"lI g d j d

i,

$gl !.jdir -;

ri:

t' i.iI:-si. j, l T .. :: '

fi '

a Tl.f.i i,.'i;;fL !j i i 7 l,

1 l g ll ;jrl'dI l l

i',

I

hij ; g i ir f;:"'

- Fi

'{.

i.[ :Ii , i s'f'fEiq.ilng{Vd

.-l; i ,

- ;;;d'd*:D$j r ;;i r'2' :37' i- .

_a  :- a i: tII .i i

jg L-

]jl:

i g, '

O i

i b ];[, , .  ! .: .

li d]  ;-j1 .

.?!!']r i' 4:. jJ i' .E  :. :

. I I . i3f i

' -a ' l; i!:'

ii;l!

b 6 g 0 j,

.lf@9pil; n dk

..j ir. lI i i'! _

a.[

a . h. 'i!

j trp i.l,  ::.:;;;lTli '

js 0

i . : l 1 -

n . :i Nay ll i 'd y i _

li' l

!., h a:'

t

,: i,2$

g ei ':: .

4l ..

i

.a .

O('

r .T1aYyI .

j .

l:

ij i b

u l! g; sw.pd T

l .' . - _

~. .

' .pm'd  : :

~ .N:.y

r. e t' ( g

- i]# gp _

i

' j g

g j.l i : - E!  !

q:dM; t t,i -

, b'. i1L L, .

a r

r _

Ii. ,

a * . :

1i -

, e e _

tk, . .

v t bu

.,'::';i.i eF: p 3ET.yQf

.i d"d3b'. 1F ? g r ~.

i4  : . :

[! j e

)

a e _

m i!l

~.

~

[ ,.;J .

d f h rll i

j::

.I lk ,-

qh i c' a a . - p" a. t= f !i i.

t

/ a _

QglIi a

!: i

!!e U . :.

!i .

a  : ,' p' - .  : ;

t.k iiqe$ N iM I { III I h [l D

!.  : i' (

f i:

pgyL!

-!i s A siEy' p j- .'

i'gd

' -; i ip l.-I il i

2'  : . ' s,- j J; 4 h i.l.kNd,j l

li j- jl jb  !

l "E qSc8c9' ' !-[ . .

- :.k u. ~.i.  !, . c l f'l l i

!8

[

I cl  :

i i

0 n I

'ap h I }p:Jt!I l l i :"

ajfr 2 - . _; .

-_ .,fc.. ':iB }M3j. ip9 i -

i

. i !l  !

- .i

. .~

a.

-d '1h y yy"

, , - . 4

!OM'-

l

) ] . - l'  : . .

/

+y 3

.@ jI ~ i ):.* - i1 1 ;

n '

2s ~

-  :.f

' .* '- hl' ::.3:

-.. ;T:i r

i

- i

' i

.1 c'

+- . .

f o

!. .9* ,- .i f:@i a.u 1: : .

. " :: c. r h3+ : - ". _ ..; , 5e _

!; i f',!;

g.d ,'l =;-  !

! i 8 d;!  :

a _

t p

a 6! b. ? jt}tr.;L. 8'1d .. . :. : a -

' l D

8 .

' .1l .

L 1 q

2 _

t.:i L l. ,.2t h.: _

..F r -

t _

.(R -

. {. 0 L.

' - s _

- i j.' e

_. - .t _

.. - . T _

. . - i . - . 4 f'.2. ._ ' . - - , -

~

t _

9 s T.

n pJ, r

1' .

Iy - .

-t. i

.i y 7 . .

u B

e: - ~ . i

. J .-  : . .

~a7 'f n

r *

i..7" 't Li" .' 7 .

i.. ..--. ,

T zl- .

.y -

a

_ " ..

;  ! i i

r s 'c *

• ht !a- aZ y :E g

! ,i j s[

' :' 7a!

6 ;.~  !

43 0 l 0 e 0 $ B 5 0 $ 0 1 1 7

D e

q r

• ] coWq.gm myyG wOZ gO mfgL~W A oCe y4 u g

i F

w 'N *=

. Table 7 Belgian Burst Test Data for 7/8 Inch Tube Diameter l Test Crack Restraint Full Burst Pressure Average Stress Length Circunferential (Degrees) on Net Remainin,g Fracture Cross Section" 1 300* Yes Yes 5145 psi 112,300 psi 2 300* Yes Yes 5468 psi 119,400 psi 3 270' Yes No >6027 psi >87,700 psi 4 180' Yes 8011 psi No 58,300 psi 5 210' Yes No >6791 psi >59,800 psi 6 210' Yes No 6615 psi 57,800 psi 4 7 240* Yes Yes 6791 psi 74,100 psi 8 210" Yes No 6497 psi 56,700 psi 9 240* Yes No 6615 psi 72,200 psi 10 270' Yes No 6100 psi 88,800 psi 11 270' Yes No 5703 psi 83,000 psi 12 180' Yes No 7982 psi 58,100 pei 13 300* Yes No > 4424 psi >96,600 psi 14 300* Yes Yes $174 psi 113,000 psi 15 210' Yes No 7026 psi 61,400 psi 22 180* No No 4748 psi 34,500 psi 23 240' No No 2263 psi 24,700 psi

. 24 240' No No 2146 psi 23,400 psi 25 180* No No 3454 psi 25,100 psi J

! 26 150' No No 6159 psi 38,400 psi 27 210' No No - -

28 210* No No 3601 psi 31,400 psi l 29 120' No No 7629 psi 29,400 psi

, 31 150' No No 5012 psi 31,300 psi 32 120' No No 8540 psi 46,600 psi (1) Axist stress = Axial End Load / Net Remaining Cross Sectional Area (2) Heat 71383 0,= 40,200 psi o,= 94,975 psi (3) Heat 71692 0,= 42,340 psi o, = 101,640 psi (4) Outside Diameter = 0.875 inch; Wall Thickness = 0.050 inch

, (5) Machined Through wall Defects 3-25 I

l t

?

l l

t l

l l

l i

tiH . C i.l..?~T

• 1-i"1" 1 :

p= .- ,. : q :.:

. . ::: .  : L:

'iJ l' T:~l ~iti"

...-l.*

7 ".U""!. ~1  !:: il y f ..Tg- ;:. .n t .i -::

. ~ ;m, . .. .

y,,.;... .

.,. ,.,l.  ; . .. .

l. g l rl.... . j ._ ,

. :.: g :-

d., .- .e. .j ,.,,;. p .- o.

,g ii? ~lcf.+ jii31 :t:= , -j p:.l ' Q .j:.4a.:)7:l ( 4 : 17 ]1 : ii l .W .1. tr [h!!

"=i L

_.::f" ~U

.i~

'. .,,4

"!rg??j.':i :f"I';l :" ' rJ l:'.

. 3 4-j ---H -- Pet--h- .l- ;--'j- ,_ ..

i

... _ . . ._ ; . . .. : ., . r"~ ' --

. .f 9 . . ..t j , *.  :. _ k -*. . .' !.N j""_

. I. . . . _. . ..Ef r.._l ' 3.

., .4. : .:  : . I. . l.:...  !

[""b[k

.. i .

!!5hPh.k I"j:,,.

l*  !.. "j i. l."4 :f;; *'-":j.!(i.M #:5(* d}"i7fMT.N,k ;.).

• ld .

l

~4;;W:::::: .

.e. :.

- .t..j p l . . : l .4:-- . : . f. _ a f -.... . 1 ....._ W.:::. f _J,p; :' cj. _..

hi-El'i?[=2 MESS:..C:m w .l uCl ;\.- s.T: ::t;E.[=------- $.lf.=} ::#:e ad?: en

'r-E!:I=T=%-7-*:22I. lh d j~L0 ~.!+'l 0 + li._4.-.

d P:-R - E E4: M":--!5:iM~N=~f-i-MIitf ' d ii:15:-:~-T=f;.::lfi :dEfrrU2 ..j- ^ t:h: a:E!i- : "F' f:ii i;9{::.r.= tie ~==:1=:t :!P'U~'- -

-'I:=:-

- - ~ ~ - .5:5~^_;iE5f=255idl@d : DM V : .it - -l~ 5:i:@rl~;'*" :._-5_:_:% N ~-N.'

=se: -,~=r.=r=-_ -*T.

-~ mu *

~ ~ ' ' ~ ^ ~ ~ ~

n n 3' *:' -d d' .m- u..  ; A. e

? 'w: ..:._--  : .

W4 d *.*r =-Pr'".".*"_:"-::t ": '_~_i .:r -- 4:i:_T'*% .ir

.=.._i.l :p. :1- .p:; : q.:-- r:p._j=_ ;_].;-i-- .r-

Mi e=-l.: -_-it 4_ . _@_j. ;,. -b .- ~t":~-E--M

- . . . _ . _ __.:-,:.----*=_._ _ .. . L- - . ..- ._ :- - ._._._2 _ -#-_

_.;_- . g

.__.,__._m. _ _ . .

E. -T-dr f55t'!! :::M4:Cii -1::9=i l "r R .* tWrE :iit .R---#::=-Pfi9# s

";;xt.ygr.ap_:r.c. ._l:s i=.",=.' .r-4 t:

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

1 - :- wu-gga _;: :C ;----~ =_:==

q p j

.: ? " r. I . f, y' N. .L: . . ., . .~.:W f~f05 ' ' "'" l %'ii:

_i..h.-.-

. f;,f-(:---P--3 . g

- *---k: ".J:- .m i

'El-i: c% gj.- Eg+. .-y 4 .e .'rl '

p.ir;."--p gw.y -c:i=_p=;. . =;=z

] l

. )._ l , I a i

y l'

/ ].:.. . . ' : .1. . .  ::.. -<

ir ry:. &

4-  : l: - - .n Mi;.J .. .i;;;-.. ..,: .a. ,. . .

• - l, g l

.?d.':.7; j.1 g

' ,I.[ l . ;. '-l .4. !. . . .j.; E. l

[ .f'. f _~g -I M.: :.f.j:

=+:n A-h :i:i E9W:4- Mie-4Nu- l ai .I :4 .Ia. *- I-0; :-% E i]=;iE mii~=-

y

= ,?

=j=Tp:2.!=_-1 :id.s_=__rg--E.Wiik ri414 :- l " i da :.; : - -H - - =9=i: si:i:F=fE' u-M t 5 j

bbkNbbENd dd.N.i.-!d.S * '-!' k - "l I ! .~ .' d b .' ."N

= = -

"MZ. ; -=## -FiE4-@= 1- "4 <!, ~!, - !b__m ap;ii_=g =;=ay=q

. . .i._.:-=.=.:a.=::..i n. _ _. _ j.. :c. : -

T. . .:

!  : i. i :

,4ir-r 1;"

+. ._. .g=_;._= y.=2 . .-

=r.r.y':: o 5c w

i_FZi@'Et.ii:LEjji.3j ;i';==!: =n.fi:l 7 % u i" .l-

.j -l' . yf. i_

q:::ffni!%,:j-%. #t; ; g lif.I'Q.M: :"'M Q ~ Q. 2~ # , \ .:p., .. .j :. ..l.l,j. . 15.

. r; hkyd 747.; . j. :--

i:a --W.Hi..= g Eri=r"y:n- .!- 1 i r -lru . kl, c g a i3 + -r i.pl ih -- f;- : :ilr-jf , .l :i .

.; r. Rp j.y .i.+_ii. M j.::j g 4l.

l l l.i

.t i-

-f l ~:r #ti 15 l .: . .ii. p

. ct.: ,

a f- .! _ f .-  ; . I. : n

.l' W 1- . l . .t .. I . ..

I. . I i l ;4 ,

' I.. g

. U. --.-j - JL: l. .q" j

': ". ca" : * -- ;" '.a-

.. } : *jf- j 8-Ti ': . - { t'!:-" O

!!. - N

. .}i! l: 4; . :. .; l._l 4..T. j rl . . .l?,

y l 1

..M  ; l . . . l. . . b- d j

.!  ::l -1 i H l l i l l 1 "l - o

,. 4 M: _ 1 l l I 4 1- l1 ,

.e

.) : l-l 1 i: .

l l l- l 6. i - x

.i :. -a s . c-g41. C

.c - l 4

..p.a- ,-; :.l.< ; l. .A . : . -

up .i: F. :t : e.-i sr-

. ; :. l .;.

. j !

. l-l ; I - r.- H H 5

P l i4=- l e h  : '

.,-* ..i.. f i;.x l .:- _ r_ :. i. : +: .I l l ~~~~~

. ' .g.- .f.:..I**"E*.._ l:

.f.-

.. n ". 'l 4 . h ".,l:" . . ]. - *. l .:'.I r.:r

• " ; : p I. a i 1. ' . l f'W I ."; 2]p.".%7:

.q g w p.,,

o o W e o =

- C m C m

W

!sa 'uo! pas sso.q aaN uo ssaa;g a8caaay b e

a l

l 3-26

( Discussion 1

The review of available burst test data identified many cases for which l

circumferential1y cracked tubes sustained net section stresses that equaled or exceeded the ultimate strength of the tube material. On the other hand, there were also a large number of cases where the failure stress was significantly i less than the ultimate strength and was rather more consistent with a flow l

stress or yield strength criteria. The following factors appear to contribute to burst at higher levels of net section stress at tube burst:

l 1) The reduction of bending stresses at the cracked section from the restraint of the lateral motion provided by tube support plates.

2) Cracks with a uniform circumferential depth can eliminate the bending stresses associated with circumferential cracks of nonuniform depth.

3) ' Stress corrosion cracks with multiple ligaments can enhance the strength of the cracked tube specimen compared to specimens having machined defects with simple planar crack geometrie: . The reinforcement associated with uncracked stress corrosion ligaments permits test specimens which are deeply cracked to sustain net section stresses two or more times the ultimate strength of the material.

4) Testing facilities with limited flow rate capabilities are often not capable of producing burst failures that would occur for the flow capacities available in operating steam generators. In some test programs this limitation has been addressed by use of sealing bladders I or patches which allow significant crack opening without premature pressure losses. However, for other tests any premature loss of pressure is reported as a leak without any means to know if the loss of pressure is evidence of the onset of a burst failure.

In summary, the review of the burst data has addressed the assumption (as in  ;

the Maine Yankee submittal) that circumferential1y cracked tubes consistently I burst at average net section stresses equal to the material ultimate strength.

There are many burst tests that support this assumption. However, there are other tests with burst pressures significantly less than the ultimate. strength limit, but most of these tests are of questionable relevance to the Maine Yankee submittal due to the use of machined flaws rather than service degraded specimens. Specific examples of lower pressure bursts are the Framatome and Belgian tests for through-wall circumferential cracks, for which failures occurred at net section stresses more representative of yield or flow strengths. These tests indicate that the Maine Yankee criteria based on ultimate strength would be unconservative for cases of unexpected cracking morphologies consisting of: 1) stress corrosion cracking of highly nonuniform depth which results in high levels of pressure induced bending, and 2) stress l corrosion cracking with minimal reinforcing effects from uncracked ligaments.

3-27 l

j

References Mills, W. J. 1987. Fracture Toughness of Two Ni-Fe-Cr Alloys. Engineering 4

Fracture Mechanics, Vol. 26, No. 2, pp. 223-238.

, Keating, R.F., P. Hernalsteen and J.A. Begley, 1995. " Burst Pressure Correlation for Steam Generator Tubes with Through-Wall Axial Cracks," EPRI Report TR-105505, February 1995.

I J

1 i

l 1

4 4

t i

i s

t i

i a

! 3-28 l

l

1 l

i i

ATTACHMENT #4 4

1 CRACK OPENING DISPLACEMENT CONSIDERATIONS i

i i

i e

4 1

a I

i

d ATTACHMENT #4 i

CRACK OPENING DISPLACEMENT CONSIDERATIONS l The consideration of part through-wall cracks in steam generator tubes involves dimensions of the remaining ligaments that can be.on the order of 0.010 in. or less. This presents a dimensional scale that is outside of 4

generally accepted validity limits of J-integral controlled fracture mechanics. However, the crack opening displacement parameter offers an 4

alternative approach that can be used to address the situation of small

] ligament dimensions. A crack opening displacement approach is also well i

suited to the displacement controlled situation that applies to the cracked tubes at Maine Yankee, whereby the locking of tubes into the first horizontal support represents the source of the displacement controlled loading situation. Therefore, the crack opening displacement approach has been selected as the most appropriate fracture mechanics criteria for predicting the growth of circumferential cracks in degraded steam generator tubes.

l Kanninen and Popelar (1985) relate the crack opening displacement (indicated j in Figure 1) to the applied value of the J integral as follows: l l

6, = d,(J/a,) i i

where 6, = Crack Tip Opening Displacement, d,, = Constant (function of strain hardening exponent n),

4 l n = Strain Hardening Exponent, 1 J = J Integral, and a, = Yield Stress.

A paper by Mills (1987) reports values of the material parameters for the Alloy 600 tube material at room temperature as n = 3 and a = 41,300 psi from

, which an estimated value of d, = 0.2 is obtained. MillsafsopresentsJ 1 resistance curves that indicate the following 4

Ja = 4,570 in-lb/in' J(Aa=0.2 inch) = 11,400 in-lb/in from which estimated crack opening displacements covering the range of interest for the ductile tearing process are 4-1

J Integral 6, = Crack Tip Opening Displacement J = Ju = 4,570 in-lb/in 0.020 inch J = J(Aa=0.2 inch)

= 11,400 in-lb/in: 0.050 inch These values of crack opening displacement significantly exceed the dimensions of the remaining ligaments for deep, part through-wall cracks in steam l generator tubes, indicating that the assumption of J controlled crack growth I is invalid for these cracks. For the Maine Yankee steam generator, the tube wall thickness is 0.048 inch. In this case the remaining ligament for the limiting crack depth of a/t = 79 % is only 0.010 inch. Clearly the crack tip will. blunt as indicated in Figure 1, and the ligament will fail in a mode of tensile instability rather than in a mode of J-integral controlled fracture.

The levels of crack opening displacement needed to fail remaining ligaments of various dimensions can be estimated. The effective gauge length of the remaining ligament that fails by tensile instability will be assumed to be two times the dimension of the remaining ligament. From Mills (1987) the tensile elongation of the Alloy 600 material is about 40 percent. For the 0.048 inch l

wall thickness of the Maine Yankee tubes, the following crack opening displacements were calculated as a function of crack depth. l Crack Depth Dimension of Limiting Crack Limiting Crack I (Fractionof Remaining Opening Opening Wall) Ligament Displacement Displacement (FractionofWall (inch)

Thickness) 50% 0.0240 inch 40% 0.0192 inch 60% 0.0192 inch 32% 0.0153 inch l 70% 0.0144 inch 24% 0.0115 inch 79% 0.0101 inch 17% 0.0081 inch i 90% 0.0048 inch 8% 0.0038 inch 95% 0.0024 inch 4% 0.0019 inch i

4-2

l l

l These estimates, while based on certain simplifying assumptions, are believed to correctly indicate some significant trends. It is seen that the crack opening displacement of 0.019 inch for the 50% wall crack is about the same as cited above for crack initiation (Jn) in larger scale fracture mechanics specimens. It is also noted that these numerical values of crack opening ,

displacement are similar to values cited by a fracture mechanics specialist on I l steam generator tube integrity during an informal discussion with Pacific Northwest National Laboratory.

P.eferences Xanninen, M.F. and C.H. Popelar 1985. Advanced Fracture Mechanics, Oxford University Press, New York.

Mills, W. J. ~1987. Fracture Toughness of Two Ni-Fe-Cr Alloys. Engineering l Fracture Mechanics, Vol. 26, No. 2, pp. 223-238.

I l

l i

i l

l 4-3

. - ~ - -- ___

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

l f

b 1 l

. \

l 4

\

' 4 5' [\ \

\

%^\

U2 d  !

N j s

6t .

Y

/

/ I n/

r

_c

.i il&

Figure 1 Definition of Crack Tip Opening Displacement 4-4 s

1 i

4 i

I I

i

! )

i ATTACHMENT #5 l

COMMENTS ON ELASTIC-PLASTIC ANALYSIS OF l FLAWED STEAM GENERATOR TUBES ~ INCLUDING

{ EFFECT OF TEMPERATURE ON MATERIAL PROPERTIES I

i i

i l

l

ATTACHMENT #5 COMMENTS ON ELASTIC-PLASTIC ANALYSIS OF i FLAWED STEAM GENERATOR TUBES INCLUDING EFFECT OF TEMPERATURE ON MATERIAL PROPERTIES i

The elastic plastic finite element structural analysis of the degraded steam generator tube of Maine Yankee (MN-95-24) has been reviewed. The steam generator tube was assumed to have a circumferential crack just above the tube ;

sheet. The analysis incorporates increased flexibility of the tube due to the-presence of the crack. The increased local flexibility is nonlinear and is dependent on the crack depth and the magnitude of the loading. Two structural models, one simulating the in-situ pressure testing and the other representing the operational situation (Figure 1) were compared for the ratio of the load transmitted from the tube to the tubesheet to the applied axial end load. In both situations, the tube vas assumed to be locked at the first support plate elevation. The pressure induced axial end loads were applied at the end plug location for the in-situ test case and above the tube support for the operating condition case. For the operating condition case, an additional vertical displacement was imposed on the tube support to model the stretching (bowing) of the support plate under main steam line break (MSLB) pressure.

Two degraded crack configurations (79% and 90% through-wall) and two values of MSLB differential pressure (2800 and 5000 psi) were used for load comparisons.

A ductile tearing analysis was also performed making use of the finite element generated J-integral and the material J-R curve. The J-integral values were calculated for crack depths ranging from about 70% to 90% through-wall for an internal pressure of 5000 psi. It was concluded that the steam generator tube could be susceptible to ductile tearing under 5000 psi pressure for an 85%

through-wall flaw. For smaller defects and/or lower loads, it was concluded that the tubes would fail in tension, and not in ductile tearing. l A scoping analysis was performed to determine the values of the J-integral associated with an internal pressure of 5000 psi for various through-wall crack depths. The tube was assumed to carry the entire axial load (end cap) due to the internal pressure of 5000 psi, which was calculated to be 1680 lbs.

The strength properties were obtained by Framatome and reported in EPRI (1991). An average of five sets of data on alloy 600 gave a 0.2% offset yield strength of 40 ksi and an ultimate strength of 96 ksi at 650'F for the material. The 0.2% offset yield was used as the reference stress, on , and the reference strain, s o= go/E = .0014 with E = 28 x 10' psi (the modulus of elasticity). The hardening exponent was reported to be 0.38 ( EPRI (1991) ),

which is the reciprocal of n, the exponent used in the Ramberg Osgood relationship. The true strain of 0.38 indicates the onset of tensile instability (Mendelson, 1968). The engineering strain corresponding to the true strain of 0.38 is 0.462 (cm, = in (1 + c,y ). Using the ultimate strength value of 96 ksi (a) at a strain of 46.3)% (c),o a = 40 ksi, 5-1

to = .0014; the parameter a in the Ramberg Osgood relationship was determined to be 32.76.

"= +ax( ")"

For the material, the Ramberg-Osgood exponent, n is 2.632 (1/0.38). This value is close to 3, a value used in the elastic-plastic analysis of flawed cylinders by Kumar et al. (1983). The results from that analysis were used to estimate J-integrals for 12.5%, 25%, 50%, and 75% crack depths with the exponent of 3, and a b/R, = 0.15 (halfway point between reported values of 1/5 and 1/10), b and R, being the wall thickness (0.048 in.) and the inside radius (0.327 in.), respectively. The J-integrals were subsequently extrapolated for larger depths. A Lagrangian form employing the set of four calculated values were used to develop a cubic polynomial which was made to go through the four data points. The calculated values of the J-integrals using the analysis of Kumar et al. (1983) and the extrapolations are superimposed in Figure 2 on the J-Applied curve. The procedures used to obtain the values are tabulated in Table 1. In Table 1, P is the axial load of 1680 lbs. R, = R, + a , a n d E' = E/ (1-v')

The extrapolated value of the J-integral at 90% depth of 250 in-lb/in is considerably less than the Maine Yankee finite element analysis (JRH-95-67) value of over 800 in-lb/in'. The J-applied curve is also seen to suffer a steep rise as opposed to a gradual rise in the Pacific Northwest National Laboratory calculations. These differer.ces could conceivably result from the additional stiffness parameters and more complex loading in the Maine Yankee finite element analysis.

Apart from the accuracy issues of the J-integral calculations, there is a fundamental point that needs to be emphasized. Before it can be assumed that J can be used to correlate the initiation of ductile tearing, J dominance needs to be demonstrated. The regions in which the finite strain effects are important must be contained well within the region of the small-strain solution dominated by the singularity fields. For the case of a 90% deep crack the ligament available is only 0.005 in. whereas the required characteristic size of the region, 20 J/aois 0.4 in. (for an assumed J of 800 i n-l bs /i n') . This characteristic size ensures that the singularity fields provide a good approximation to the complete small-strain theory solution.

For the cases considered (over 85% deep cracks) such a region is not available. It was therefore concluded that the J dominance field does not exist, and the fracture resistance curves are not applicable in this case.

Considerations based on plastic instability should govern the failure mode as i was concluded for shallower cracks (less than 85% deep) in the Maine Yankee analysis.

5-2

4 a

Table 1 Elastic-Plastic J-Integrals for n=3, b/R, =0.15 i 1

4 a/b (3)

Depth Limit I:1 P/P o h3 F f,[r] J, in-Fraction Load, Po lbs/in* j j

(lbs) 0.125 4315 0.39 6.02 1.18 2.34 1.55 ,

1 0.250 3730 0.45 5.87 1.29 5.60 4.48

0.500 2528 0.66 5.11 1.72 19.90 23.16 0.750 1285 1.31 2.40 2.32 54.32 121.64 i

Notes:

2

[1] Po= 2- c on (Ro*-R )

4 I

8

[2] f 1=

x(Ro-Rj) 2 1

j i

[3] J=f1 Pa/E'+aceoi c h c(a/b) (P/Po )'

E i

i a

5-3

EFFECT OF TEMPERATURE ON MATERIAL PROPERTIES The J-resistance curve obtained by Mills (1987) is the most recent test data on Alloy 600. For a 3 in. thick plate specimen, the J, curves are shown at room temperature and at 800'F and 1000'F in Figure 3. This figure shows that the elevated temperature J,, values (3300 in-lb/in z) are 50% higher than the room temperature value (2200 in-lb/in'). In Figure 4 [ Mills (1987)], the J, curves for 2 in. and 3 in. plate specimens at 800'F have been compared. The 4

Jr, values are not found to be significantly different from each other. The applicability of these data to the Maine Yankee degraded tube is questionable, because of the lack of ligament thickness needed for crack tip blunting for a remaining ligament of only about 0.005 in. However, the favorable trend of d

fracture toughness at elevated temperature should be a conservative feature to account for results obtained from an in-situ test performed at room temperature. The decrease in the yield and ultimate strengths for the material at operating temperature are typically less than 15% of the room temperature values (EPRI 1991). The yield strength drops from 46 ksi to 40

' ksi, and the ultimate strength from 102 ksi to 96 ksi (based on averaging five sets of test data reported for Alloy 600 in EPRI (1991) ). This justifies the room temperature pressure test with a proposed 15% increase in pressure to

, account for the differences in the tube strengths.

REFERENCES Maine Yankee Letter NN-95-24, " Response to Steam Generator Issues Request for Additional Information: Elastic Plastic Finite Element Structural Analyses of a Degraded Steam Generator Tube During In-Situ Pressure Testing and Actual Operating Conditions", March 10, 1995.

Kumar, V., M. D. German, and C. F. Shih. 1983 Elastic-Plastic and Fully Plastic Analysis of Crack Initiation, Stable Growth, and Instability in Flawed Cylinders. Elastic-Plastic Fracture: Second Symposium; Volume I-Inelastic i

Crack Analysis, ASTM STP 803; C.F. Shih and J.P. Gudas, Eds., American Society of Testing and Materials, pp. I-306 - I-353.

Mendelson, A. 1968. Plasticity: Theory and Application, Macmillan, New York.

Mills, W. J. 1987. Fracture Toughness of Two Ni-Fe-Cr Alloys. Engineering Fracture Mechanics, Vol. 26, No. 2, pp. 223-238.

EPRI. 1991. Steam generator Tube Integrity, Volume I: Burst Test Results and Validation of Rupture Criteria (Framatome Data). EPRI NP-6865-L, Electric Power Research Institute, Palo Alto, California.

5-4

d 1

1 J

Locked Horizon:21 F I suppen _b ,

Il 1

EINE.SD -

l

/ ~ b y

2 V~rh- ,{/"-k u

9 ) -  %

h M L = 42" or more - 1 D = 47" or A. greater * *

+-- --+

L = 5" o less e e n ~

- ---- - l

.  ; . . e

///////// \\ \\\\\\ /////// // NNNNNNNN l In Situ PressureTest Loadings Normal Ops (locked horizontal supports) (locked horizontal supports)

ELASTIC PLASTIC FINITE ELEMENT MODELS Figure 1 h

5-5

4 4

i i

a 1

a i

a i

i  !

t 4

3 i

i i

! i i

} [ Figure 2. J-Applied vs. J-R Curve for 5000 psl Applied Pressure l 1

1000 s

h e

90 %

l 800 "

F 4

A 600 l f

/

& J

3 i /  ;

400 I8 '

kNL I

[

• f 200 e 1 + J-R p # [ 6 J-Appled e -

LOO < 7gg i

o__ - f sts 0.035 0.037 0.039 0.041 .0.043 0.045 0.047 0.040 0.051 Crack Extension (in) e i

r i

4 4

I t

4 i 5-6 1

1

i I

4 l

l 3000 i- i  : i i INCONEL 600 G 24*C Y 427'C ~

A 538'C g

2000 -

I 427,538'C Je = 581 2 31 kJ/m2 gh

.E dJ/da = 340 215 MPs - jA 7

T = 395 Y g -

-I YA 1000 -

/y 24*C s AY Je = 382 2 89 kJ/m2 9 dJ/de = 338 g 39 MPa T = 428 J

$/ = 40.(aa) 0 ' ' ' ' '

O 1.0 2.0 3.0 4.0 5.0 6.0 CRACX EXTENSION. mm Fig. 3.1, curves for the 63.5 mm thick plate of Inconel 600. The open symbol denotes stretch zone formation. Symbols with a vertical slash fell outside the 0.25 and 2.5 mm exclusion limits; bence, were not included in the regression analysis to determine /,

5-7

i l

l l

l i

l 1

W. J. MILLS i i i 1500 -

Y -

i

. O l

~

25.4.mm PLATE t0.5777 & 0.525T CT) 'O Je = 575 g 42 kJ/m2 .

dJ/da - 420 g 34 MPa / 0 T = 522 0, -

1000 -

63.5 mm PLATE (2T CT) -

O Je - 581 2 31 kJ/m2 ,

O y dJida = 340 = 15 MPs j "k - O T = 395 _

l a

4 -

o .

500 -

@ INCONEL 600 _

J = d o,ta al TESTED AT 427'C

' I ' I '

0 '

l 0 1.0 2.0 3.0 '

CRACK EXTENSION, mm Fig. 5. Comparison of Ja curves for the 25.4 and 63.5 mm plates of Inconel 600.

5-8

4 4

6 1

1 ATTACHMENT #6

EFFECT OF INDUCED BENDING ON

FRACTURE OF TUBES WITH PART THROUGH-WALL FLAWS 4

1 l

d

ATTACHMENT #6 EFFECT OF INDUCED BENDING ON 1 FRACTURE OF TUBES WITH PART THROUGH-WALL FLAWS The service induced degradation of concern to Maine Yankee.is part through-4 wall circumferential cracking that extends 360' around the tube, but with the depth greater on one side than on the opposite side of the tube. The test data !ndicated that under increasing pressure, the deeper portion of such

, cracks will " pop through" the wall resulting in a significant leak. This " pop through" pressure can be considerably less than the pressure needed to burst a tube with an equivalent uniform depth circumferential crack of the same average crack depth.

} The discussion below presents fracture mechanics calculations that compare the fracture of steam generator tubes for cracks with non-uniform depth, with the corresponding fracture behavior for cracks with uniform depth. Two key factors addressed in these calculations are 1) tube bending due to the nonsymmetric crack geometries, and 2) the effect of uncracked remaining ligament dimension on the limiting crack tip opening displacement.

Tube Failure Scenario Figure 1 shows the tube configuration of interest. The length of 47 inches corresponds to the distance between the top of the tube sheet and the first horizontal support plate for the Maine Yankee steam generator. The tube outer diameter is 0.75 inch with a tube wall thickness of 0.048 inch. In the

, present calculations, a part through-wall circumferential crack that extends i

360' around the tube is considered. The circumferential crack, as addressed by the present calculations, extends nearly (e.g. a/t = 90%) through the entire wall over half (180') of the tube circumference, and to a lessor depth for the remainder of the circumference.

j The effect of the non-uniform crack depth of the crack, and the resulting failure scenario is as follows:

1. Pressurization results in an imposed end load on the tube. In the case of tubes that are locked into the tube support plate, this end load can be represented as an equivalent imposed remote displacement.

2. With the circumferential crack being non-symmetric, the end load induces bending stresses at the location of the crack.

3. The bending action enhances the opening for the deeper portion of the crack, and thereby imposes a greater level of tensile strain on the relatively thin remaining ligament for the deepest part of the crack, which is also the part of the ligament that has the smallest limiting 6-1

value of crack opening displacement.

4. Under increasing pressure, the imposed crack opening displacements can exceed the ability of the thin ligaments of the deepest part of the i cracks to accommodate the imposed high levels of tensile strain.

5. The thin ligament fails by tensile tearing, a through-wall crack is

!. i produced, and a leak in the tube results.

6. The remainder of the tube circumference, which has circumferential cracking that is not as deep, is capable of sustaining the increased pressures before a complete rupture of the tube occurs.

Results of Fracture Mechanics Calculations Figures 2-4 give the results of fracture mechanics calculations that address the bending action associated with the circumferential cracking of non-uniform depth. Results for a crack with a. uniform depth are also shown for comparison purposes. In both cases the average crack depth was taken to be a/t = 0.79. '

Calculations for the non-uniform flaw assumed a through-wall crack over 180' of the tube circumference, and neglected any structural contribution of the very deep, part through-wall crack over 180' of the cross section. This simplified calculation approximates a very deep flaw, for which the stress acting on the small remaining ligament has a relatively small effect on the opening of the crack. The approximate calculation also neglected any loss of strength associated with the less severe cracking over the other half of the circumference, and in effect assumed that compression from bending stresses will close this part of the crack.

Elastic-plastic finite element analyses were beyond the scope of this evaluation. The simplified, hand type calculations addressed the two bounding cases of linear elastic behavior and fully plastic behavior. Handbook solutions (Tada 1985 and Zahoor 1989) for circumferential cracked cylinders were used for the linear elastic cases. The fully plastic cases assumed a uniform stress level acting on the entire remaining ligament of the cracked tube, with the stress level equal to the flow strength of the tube material (taken to be 65 ksi).

Figure 2 addresses the case of a part-through crack of uniform depth, with no induced bending. The crack opening displacement (from linear elastic fracture mechanics as given in Tada 1985) remains very small until the net section stress on the cracked tube equals the 65 ksi flow strength. Additional increments of imposed remote displacement are then accommodated entirely by crack opening displacement at the site of the crack.

6-2

Figures 3 addresses the case of a part-through crack whose depth is nearly through-wall over 180* of tube. This crack configuration gives rise to

substantial bending. The " kink" angle associated with this bending was calculated from linear elastic fracture mechanics solutions given in the Tada Handbook (1985). Bending moments in the 47 inch length of tube needed satisfy the displacement and rotational restraints at the tube sheet and at the tube support plate were addressed by treating the tube as a beam. At higher levels of imposed remote displacement, the net section stress on the cracked tube was set equal the 65 ksi flow strength, with the bending action due to the eccentric loading being used to calculate the crack opening displacement. It is of interest to note that the slope of the curve in Figure 3 is only slightly greater in the plastic range than for the elastic range. This indicates that the imposed remote displacement even in the elastic range is largely accommodated by opening of the crack.

I Figure 4 shows the difference in crack opening displacement for the two bounding cases of crack configurations. The 180' crack has much greater crack opening displacement for remote displacements up to about 0.040 inch. The following levels of allowable crack opening displacement as a function of the l crack depth (or dimension of the remaining uncracked ligament) were estimated in Attachment #4. '

Crack Depth Dimension of Limiting Crack Limiting Crack (Fraction of Remaining Opening Opening Wall) Ligament Displacement Displacement '

(FractionofWall (inch)

Thickness) _ j 50% 0.0240 inch 40% 0.0192 inch 60% 0.0192 inch 32% 0.0153 inch 70% 0.0144 inch 24% 0.0115 inch 79% 0.0101 inch 17% 0.0081 inch 90% 0.0048 inch 8% 0.0038 inch 95% 0.0024 inch 4% 0.0019 inch From this table of allowable crack opening displacements, and from the calculated crack opening displacements in Figure 4, predictions have been made for the remote displacement (and equivalent level of applied nominal stress) needed to produce a through-wall crack in the tube. Two cases are addressed, both with an average crack depth of a/t = 0.79. In one case the crack depth is uniform around the entire circumference of the tube, whereas in the other case the crack depth is not uniform but has a depth of a/t = 0.90 over 180' of 6-3

4 i

] the tube circumference. The following failure predictions were made, i

1 Crack Configuration Allowable Crack Critical Remote

! Critical Noelnsl Opening Displacement Displacement Applied stress

Case #1 Average Crack Depth = 79% 0.0081 inch 0.031 Inch 18 ksi 360' F!av of Unifore Depth j I

j l Case #2 i

Average Crack Depth = 791

Flaw Depth Not Uniform 0.0019 inch 0.0199 inch 11 ksi Maximun Depth 1.

= 90% over 180* of 8

Tube Circunference

The calculations in the case of the crack with non-uniform depth should be 1 considered as a rough estimate of the failure stress. Nevertheless, the j

results support the scenario that a through-wall crack occurs at a  !

significantly lower load for the non-uniform crack depth than for the crack j with uniform depth. The relative loads have the ratio of (11/18)x100% = 61%.

The critical axial stress of 11,000 psi for the non-uniform crack corresponds j to an internal pressure of 3,460 psi, whereas the predicted the burst pressure

! for the crack of uniform depth is 5,670 psi (corresponding to the nominal j applied axial stress of 18 ksi).

4 j References

! Tada H., P.C. Paris'and G.R. Irwin, 1985. The Stress Analysis of Cracks l Handbook. Second Edition, Paris Productions Incorporated, St. Louis, Missouri.

Zahoor, A., 1989. Ductile Fracture Handbook, prepared by Novetech Corporation, Gaithersburg, Maryland, for Electric Power Research Institute, 1 i

Palo Alto California.

l

{

1 6-4

)

M i

f i

}ll////////// )(

, d i

, Remote Displacement (

or Load l l

i 1

)

i I

j f = 47 inch Deep Part Through Wall Flaw

_~.

V

// / /// / / / / / / / / ////////

Figure 1 Loading of Tube with Deep Part Through-Wall Circumferential Flaw 6-5

l Nominal Stress, ksi 0 20 40 60 80 0.15 -

1  ? A- 8 m ifa m  : O . = pp- e : . . . 6,., .

i.; e., :.c ;in g . ,i: y.-

)' : !'l '.ld. i[pl;. e 4

i.. .p

. :jii je ;t.

l!. h ' '

__.-Net Sectionk'

• ^

I' '

Y dh

• "" 1 * 'E

'" 0 Yield ' I e l'

h  ::!' ,u, c :ni :9 qg -

nt p:

hl :H .  ;;i' .I i ly !p! bl, ,:  ;!i'  ::! p. -i; p. t i: pq jp ;g. jp !4 >;l- ;p ig (;;

in ih ,

iC ;g  :;I ij!: 90 nh ;l -i-  :;i .

i ,p! O p;, %t t- !g  :;r it- 9 . yh A; q,

,,g I!i h4 -

lU' lbll hh Hij !!!! M! :le: !!. :!!' '

i:l ' ' jt! li:  !!. !.j: iif; ;g i M  :!li j '4!  ! .

il;! 'ii: 'i!i 'r 'i

!Pi till r!h i-i  :!M I: tl !! i; W iF . ,ll j.I ::h gi 0.10 N5 10 * '

"" # # 'F L' E N* JE J 41 in O I I BEi liii !ih  :  !!:i a:! si: !i- : .n :ic i:- * :fr 'ii; ,!ji ;ih  :

y ii; " cp th :p  :.: !;- t.. :yl. ::t .!F :]i .,  ::,- 3p g -

ip pg

!.y iin .; ; .

i.. .f:: ;p .

. jr .y ;m g in i;i

.i.v i- -

x .g; if

3. :  :-

e -

a .ln l :..o .i .

U.-l..! r  !- i.

f _

.jp C. I 360* Part Through Flaw h / in ', -

.50 Q

(Uniform Depth) / "

l k "i: ii' !H 1  !' i in

i .

f. ,

p .n

:!g .! ip u

!. .i;-

em .

. i i f- i. p. 49 .h; !H. :D ;;;. nh

.5 o'o"- 1 n-e y .

o o /r n .

,:or F- msnm ;ic ut. in a no n 5 *n "" * "" "

- - - " 'Ir * '" :c *": toricikiiduin ntno un a c., .m i;l- hp qii t u: vi t- 4 :i,; iid 4 f. - an @ ac .ip 9 :i O na uh mi nc n; un on n.; un ly. yn y nioin sh a e i (a/t) man = 0.79e H g III O

h di 'iU ip -ii! i!!' FE ii ! i

#: f!  :

14! illi lih til! 00  !  !! lh! lh ll ! [h -llh N} 4!i hI[

y yh i![ WI !!i: !!h IH di !4 @ Pfl OF 1h ti rd! # :l!i iib dd ;ji! i iin $! !L! ii Iti it Ith thi th L. :Hi Hl: lll; W; !!ii Hs WI ! ii !MI im i  ::n ii!! ;W !iH llli 14i !!h ilil ;i iiii Hi H!

U  ! f li:i i t tit li'i ilt lil! il # R ini ip ?! iMH 0ll li!I iih it!.! *i ud I!N .!R @ tik 4k S! ;il: !!# Hi Bt Pt l t.@ i+4r

.E Fi IFJ U!l UH li:i F#i! WhM 4!! !U: Si # eM i!h 4l!H@ !!!: N :i.5 04! -

Lid i li !.t # iH @

in t :: 'if ih! h- ii[i d- hh hd ,iH H- ti :!H U!I di! ii!! 5: I : il l!!i ..:ij !Ol ij IN! $l ;!i '.l@9i o m m .n we ce,3 _ .m un- e.mm m, a 0 0.05 0 10 0.15 Remote Displacement, inch Figure 2 Crack Opening Displacement for Part Through-Wall Flaw of Uniform Depth 6-6

R 1

4 l

l l

Nominal Stress, ksi 1

l 0 20 40 60 80 0.15 -

__ ?A _

s . m it. s a .a si e, gi. i ',.  ; t.. .

m g ra ,,.

l' I f: ' , ;: fl 'M 'i y l r

.- Net Section ,g "~ h M, 4i .;,. cp '

y- .!a  !  !. ;lp j i. ,

I

.p. L[

t ia

c. 'gj

', p ~,

'i'

" "~

..i.

..l Yield

'"' . h. . '"'

m

. ..i-

- n, ,a ig p..

4 ,  : n  :,.. ;.; ..

lg ig; g;

'; H . ,

n .1 ! Hi i p; e pi' ::i h- a lu  !: 4; .p; pq ,p ,,, pp { }; .p: qq y; i

li it: fp i ;'s l .il- Ul: .l" 'i :h !! id F: P iF' 'i i P' i

P:

li 'i!- :il' U:. !jl, .h. . ' . p  !

y O

}{ ii :

l:ji h! ti:i .

,! g.i '

-!.- ,. i;: , y:. i:- 'i: :!!. ;p .i;j g ,p j.i

!I 1 :d. !i.: 'ii- 9li t; ha ni dh in :!r D: :r! %

C' TIF ' . p.. .lH in Ph ni

/.,Y.-

n. p.

O'10 M ' $~"F'$

'" 'm.

" ci' ' '",,~ '"

..i

"' 'U! ,I, Pf!---r""

'l ' " ~

J ai ~i, i nt ik hi m d:: no vn .;p i. i:- x g. .jp .' :it.

C III qu E. ' , .

11 m;g'T

/ 4 Wi  :  ::c sp g L ..

d. Lil in: ii .! i j 1] 41 o . )f

.. -l., , e !. . an u < .<

3c,, ",.180* Through Wall Flaw

~

" .f .. .n,

.O

~ -./ T

.,s . ..t

.$ u f t.. .h- .;- i;;pa 14 pi;7 A

r.

A m 3p m

,:o :e , ,

f i . .

em ;n c . .  ! f: # t. / ' nn in is. a:- hh I!!

.5 0.05 1 * "" 9- ' *#" * *"# " "" * "U d4

• H ' H!-

C 'n ,' ei ut.

. ti! in. p. in, pl 0: ,  : f .lp n!v ' .in uh :Hi ilq glj gli y :th hll @

, i

$ !p 41 90 !hl :r it ni i: .t du f: fo . !M 'ln li :l:! ;14 o

y li; si nh

iti hj' :lh 9ec c=  ; n:.  :

a/v -

i!! ih! Ul! '$/:L,:!"ty ah 1H! !!!! !ih rip !it!

me n: iioni-$i ru (alt),,,,,,, = 0.79 g!!i

.!i- Ul:  ! h H!!  !!llll!! Ul!lii Bll[4110dii p :th hi~ @ lui :lii Ih: kl ! it liPf,# : :hi in in 'idi in ib i dit iin i ti IF: 'h! $! $ iti tittllt ihi l!i C Ifi !:l: lik  !!h :i!i !M ii! iWil e 'i!! 10 iii! I! ! !i!! ll0 @ !!h 11 i! 4h Hij HI 11 lLi tllt ijt hh !it U 4!! $ :!h in 94 Dd#hi UH i:!i h!! PE itE %! lN lM !;; i'k !ik i!k $! lil: # Il i Bi ttt! Ilh:!!b niUW

]sg i!u gu,,;pT: ;gi rin lii;e 91: @ ,u ;g q@ !!h r!' ggiiti !N g 4l Ip; y ;[jj imp is g.

4p44; ti :!U Uh w lii: s .hi k l!H :!d !!H iin iM i i k !hl m

, in ni il"  !!!, Un in 0 Fiwe * = >"' "" * "N"" i" * * "" = "" * " " * "* un "" "m ' la ' " + "

0 0.05 0.10 0.15 Remote Displacement, inch Figure 3 Crack Opening Displacement for Part Through-Wall Flaw of Non Uniform Depth 6-7

l l

l Nominal Stress, ksi 0 10 20 30 a ju- t.: h.. ;i t , ..i l.' id :F ,

~

" ~

i- i: H 0.06 -.. Net Section Yield T 'i:

5 l'  !. l v 1

.6 4- U+ .__

_i. - = _. . _ .

... _ . . .L_ y .  ;

- (a/t)..., = 0.79 .

a 0.05 .- -

c -" -- - - -

6 4

i.

l N 0.04 l'

- 1.__a,

),4 A

_180* Through Wall Flaw ,. jy em 0.03

-; :l,, ..g -

..)7 2

1

,g --

7 ., .

g l' f "'

"ti  :

h d: [ft!

r u! 4 .

.// r .

p; 0.02 1; .

, .it yy..

4g.. -,

7 y

r H , p.  :-

\\ .c a .f: /le.  :

a m :c :n: .. +: q f. .j ,. .g .i..,. .M*.g;iii' 6

g 3j ..  ;

.j O 0.01 ~.  !

m n.: .j. . . ; 360' Part Through Flaw; [

3 3 i f .:

j (Uniform Depth)  ;;

/  ;. i:' i:ii' :g  !!ii; O # ' '

- i n' ' nh h! ii' I' h' *i 0 0.01 0.02 0.03 0.04 0.05 0.06 Remote Displacement, inch Figure 4 Comparison of Crack Opening Displacements for Part Through-Wall Flaws of Uniform and Non Uniform Depths 6-8