ML20106E325

Nonproprietary Once-Through Steam Generator Mechanical Sleeve Qualification

ML20106E325
Person / Time
Site:	Arkansas Nuclear
Issue date:	10/31/1984
From:	BABCOCK & WILCOX CO.
To:
References
BAW-1823, NUDOCS 8410290165
Download: ML20106E325 (38)

Text

,

I . .

BAW-1823 October 1984 i

l l 1

I i

I ~

l ONCE-THROUGH STEAM GENERATOR MECHANICAL SLEEVE QUALIFICATION

!I ll

'I

'I I

I I

I I

I I

I puTe!PJ468l*ha Babcock &WHcox P a McDermott company

BAW-1823 October 1984 I

I I

~

I ONCE-THROUGH STEAM GENERATOR MECHANICAL j SLEEVE QUALIFICATI0.N I

I .

I I

I

!5 I

I -

I ii BABC0CK & WILC0X Nuclear Power Division P. O. Box 10935 I Lynchburg, Virginia 24506-0935 Babcock &Wilcon a McDermott company

B I

CONTENTS I Page

1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 1-1

2. PROGRAM DESCRIPTION ..................... 2-1 2.1. Sleeve Design .................... 2-1 8 2.2. Design Requirements ................. 2-2 2-4 2.3. Test Plan ......................

2.4. Supplementary Tests 2-5 I

2.5. An al y s e s . . . . . . . . . . . . . . . . . . . . . . . 2-7

3. RESULTS AND CONCLUSIONS ................... 3-1 3.1. Leakage Tests .................... 3-1 3.2. Joint Strength Tests . . . . . . . . . . . . . . . . . 3-5 3.3. Light Expansion Tests ................ 3-8 3.4. Supplementary Tests ................. 3-10 I 3.4.1. Corrosion Tests .............. 3-10 3.4.2. Flow-Induced Vibration Analysis ...... 3-12 3.4.3. Strai n Te s ts . . . . . . . . . . . . . . . . 3-14 I 3.4.4. Adjacent Tube Tests ............ 3-15 3.5. An al y s i s . . . . . . . . . . . . . . . . . . . . . . . 3-15 3.5.1. Performance 3-15 I

3.5.2. Struc tural . . . . . . . . . . . . . . . . . 3-16 3.5.3. Process Control .............. 3-17

4.

SUMMARY

AND RECOMMENDATIONS ................. 4-1

5. RtreReNCtS . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1 I

List of Tables Table Page 2-1. Load Condi tions . . . . . . . . . . . . . . . . . . . 2-8 2-2. 2-9 2-10 I ~

ESIDCOCIE & MlCOM a McDermott company I

I List of Tables (Cont'd)

Page Table 2-3. Steam Generator Design Parameters . . . . . . . . . . 2-12 2-12 2-4. 2-13 2-5. 2 3-1. 3-19 3 3-2. 3-20 3 3-3. 3-20 3-4. 3-21 3-5. 3-22 3-6. 3-23 3-7. 3-23 3-8. 3-24 E, 3-25 5 3-9.

3-10. 3-26 ,

3-11. 3-26 5 3-12. 3-27 4-2 3

4-1. OTSG Sleeve Qualification Criteria .........

4-3 I

List of Figures Page Figure '

2-1. 2-15 2-2. 2-16 2-17 2-3. 2-18 2-4.

2-19 2-5. 2-20 g 2-6. g l

3-28 l 3-1. 3-29 l 3-2.

3-30 l

3-3.

3-4.

3-5.

3-31 3-32 l

3-33 3-6. 3-34 3-7. 3-35 3-8. 3-36 3-9. 3-37 g 3-10.

l l 3-10A.

3-38 g 3-39 i 3-11.

I magg, 4 MCDermott company

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

i I i I 1. INTRODUCTION I 'Some steam generator tubes have been found to have a varying amount of wall degradation after only a few years' service. If the degradation is exten-sive, the normal practice of plugging defective tubes may reduce the effec-tiveness of the steam generators and eventually reduce the perfomance of I the nuclear steam supply (NSS) system. An alternative to tube plugging is tube sleeving. A sleeve is installed as a new pressure boundary inside the original tube to bridge the degraded area, thus permi tting the tube to remain in service. Babcock & Wilcox (B&W) has developed and qualified a mechanical tube sleeve that can be installed in degraded tubes of once-through steam generators (OTSGs). This sleeve is more resistant to the expected types of corrosive attack than the original tubes. It is strong enough and sufficiently leak-free to be used as a pemanent reme(y to keep degraded tubes in service. This report demonstrates the technical I adequacy of the OTSG mechanical sleeve for use in degraded OTSG tubes for both nomal and accident load conditions.

The development of the sleeve design requirements and the actual sleeve de-sign are discussed in the program description. This section continues to describe the qualification program and test specimens designed to evaluate the effect of service conditions on the strength and leakage of the mechani-cal sl eeve. The results of the strength and leakage tests are presented and evaluated for both pre- and post-service conditions. Supplerantal tests for corrosion resistance and the effects of tube surface condition and sleeve bending and straightening are also described. The results of an American Society of Mechanical Engineers ( ASME) Code analysis of the sleeve i are discussed, as well as the effect of sleeves on plant performance. The character of flow-induced vibration of a sleeved tube is compared to that of an unsleeved tube. The qualification test results are summarized, and conclusions are drawn regarding the use of the sleeve design.

I I 1-1

=a-* ="-

a McDermott comparty

I I

I 2. PROGRAM DESCRIPTION I

2.1. Sleeve Design Tube degradations found in OTSGs have been located most frequently within a few inches above or below the secondary face of the upper tubesheet (UTS).

There is some concentration at the 15th tube support plate (TSP) and in the 16th tube span (i.e., between the 15th TSP and the UTS), and the remainder seem to be rather randomly distributed at elevations below the 15th TSP.

The radial location of degradations tends to be more frequent toward the outer periphery of the tube bundle', and there is some concentration near I the open tube lane. Therefore, a sleeve that extends through the thickness of the UTS and a few inches beyond would serve the most frequent need. Ex-

. tending the same sleeve through the 15th tube support plate would be most ef fective. The sleeve must also be installable in any tube in the steam generator.

The head clearance over the outennost tubes in the OTSG is only about 13 inches, whereas the sleeve length required to extend 6 inches beyond the UTS secondary face is 30 inches. Thus, installation requi res that the sleeves be pre-bent to a gentle radius in order to clear the head and then straightened as they are fed into the tubes. After the tooling was devel-oped to bend and straighten sleeves as they are installed, it could be modi-I fied to bend and straighten longer sleeves. Therefore, the qualified de-sign was extended from 30 to 80 inches in order to span the entire 16th tube span and 15th TSP. This sleeve can be installed in any tube in the genera tor, and tooli ng modifications could permi t installation from the lower as well as the upper tubesheet. Thus, the design which has been qualified is a 30- to 80-inch sleeve located in any tube in the OTSG.

The sleeve is made from tubing, according to standards set by ASME SB-163, that has been mill-annealed and heat treated The tube size is I 2-1 ---

a McDermott company

As shown in Fig-ure 2-1, there is expansion in the tube-sheet, and expansion near the free end of the sleeve. The length of the sleeve could range between 30 and 80 inches overall.

I The installation method quali fied is roller expanding both ends of the sleeve. This method is quick, gives adequate leak tightness and pullout strength, and is easily adapted to remote operation. Both hydraulic and ex-plosive expansion techniques were considered to obtain a mechanical seal , g but the rolling technique was prefe'rred due to the speed and simplicity of E the operation and the leak-tight nature of the joint. Both brazing and wel ding techniques would yield sealed sleeves, but the installation rate was judged to be much too slow cnd expensive for large installations.

Roller expansion has been used to seal tubes into tubesheets since the 1930's. The amount of expansion is easily contolled by limiting the torque applied to the expansion tool, since there is a rapid increase in torque as the tube is squeezed against the tubesheet and begins to be extruded axial-ly. Roller expansion of the sleeve in the free-span tube without a tube-sheet back-up is an innovative application.

I of various amounts of expansion is evaluated.

2.2. Design Requirements l

The sleeve design loading requirements have been establish +) to be equiva-lent to those of the unsleeved tube on the premise that a sleeved tube g could be totally severed without affecting the function of the tube.1 The m sleeve is considered to be a structural member that meets all nonnal, up-set, emergency, and faulted conditions resulting from nonnal operation and accident transients.

The ASME loading condi-tions are tabulated in Table 2-1, and the design transients are tabulated l

! in Table 2-2. 3 l l 2-2 Batwock &Wilcom a uconmore company l

I The maximum required pullout strength of the roller-expanded joint I

Figure 2-2 plots acceptable combinations of axial load and joint slippage, which are dependent upon the yield strength of the parent tube material .

If there were no slip (e.g., an unsleeved tube), the required axial load capability would be the combined themal and mechanical load (3149 lb) of

'I the worst accident condition, a main steam line break. If there were no axial load (e.g., a severed tube), the sleeve joint must be able to slip at least 1.09 inches, the maximum relative displacement between the severed ends of the tube without separation of the joint.

I A sleeve leakage objective was established based on plant operating limits for primary-to-secondary leakage and an assumed quantity of sleeves which g could be instal led. Plant Technical Specifications nomally require that plant shutdown be initiated if steam generator primary-to-secondary leakage exceeds 1.0 gpm.

I I

The corrosion resistance of the installed sleeve must be equal to or better than the original Inconel 600 tubes as heat treated and stress-relieved for the OTSG. This is interpreted to include resistances to the following types of corrosion and mechanical damage:

1. Pitting in the presence of acid chlorides or caustic.

2. Sulfur- or caustic-induced intergranular attack.

3. Caustic stress corrosion cracking.

4. Erosion from micron-sized particle impingement.

5. Abrasive wear from platel ets of debris composed of magneti te (iron oxide).

6. Corrosion with or followed by low-stress high-cycle fatigue.

The sleeve design must comply. with the structural requirements of the ASME Code for the 40-year design l'ife of the OTSG. This requires that the tube

~

Babcock &WHcom a McDermott company

,cY fatigue analysis be reevalua ted for the effect of sleeves on the fl ow-induced vibration of the tubes. This analysis must consider both severed and unsevered tubes due to the worst-case assumption that the degraded area of the tube bridged by the sleeve may not be structurally sound.

The sleeve must be shown to have an adequate pressure boundary thickness according to methods outlined in the ASME Code, for installation in a severed or unsevered tube. The OTSG design conditions are listed in Table 2-3. An ASME stress analysis report is required to certify adequacy of the sleeve.

2.3. Test Plan qualification specimens were conditioned and tested to demon-strate the strength and leakage adequacy of the sleeve design in any 0TSG, both before and af ter service conditioning to represent 40 years of ser-vice. In order to simulate the 40-year service life, the functional specif-ications for all 177-fuel assembly (FA) NSS systems were evaluated with regard to the quantity and severity of design transients. The conditions that a tube sleeve joint would experience were then arranged as thermal cycles, axial load cycles, pressure cycles, and vibration cycles. The ex-pected axial loads and service cycles were factored to establish test axial loads and cycles as required by ASME Code Section III, Appendix II, "Experi-mental Stress Analysis," in order to account for the effects of statistical variations, cycle rate, and test temperature. The resulting service conditions which were applied to the specimens are presented in Table 2-4.

In actual service, the sleeved tubes would simultaneously incur thermal, axial load, pressure, and vibration cycles, but the specimens were indepen-dently subjected to pressure and vibration cycles in order to assess their individual ef fects. Control specimens were also processed separately with no conditioning to represent as-installed pre-service sleeve performance. g The largest set of samples was subjected to both thermal and axial load 5 cycles as an expedient because the effect of the themal cycles was expected to be minor. The specimens are identified in Table 2-5, and the processing is diagrammed in Figure 2-3.3,4 I

I 2-4 Babcock &WHcom a ucoermore company

4 . ,

I I

I I  !

I I

I I

I I

~

I I

I I

2.4. Supplementary Tests Corrosion resistance was demonstrated by cold-worked samples and control samples which were maintained in a 10% sodium hydroxide solution at 550F in an autoclave with a 190 mV potential Each 24-hour day at these accelerated test conditions represents approximately 2-s - - = = - -

5 a McDermott company

one year of service in an OTSG.,

I I

A me t-allurgical examination of the samples after the autoclave exposure per-mitted the comparison of stress corrosion cracking found in the various ,

cold-worked samples with the unworked samples. The comparison permits a l conclusion regarding the effect on corrosion resistance due to the  !

cold-working by roller expansion or bending and straightening.

A group of free-span joint specimens was leak and pull tested in order 1 to evaluate the effect of the tube surface on the joint strength and leak-age.

I I

I

I l

Four complete samples were assembled to determine whether the sequence of roller expansion has any significant effect upon the quality of the rolled l joints or the strain in the sleeve and tube.

I I

I I

2-6 Babcock &Wilcom

. ucoermore company g

I sleeves were bent and straightened to represent the worst-case (i.e.,

most curved) sleeves which would be inserted into tube s. These sleeves were assembled in tubes with tubesheets and then leak and pull tested to determine whether the bending and straightening affects the quality of the rolled joint.

The effect of rolling sleeves in adjacent tubes was demonstrated by rolling plugs in a sample tubesheet test block.

I I

p. Analyses The effect of sleeve installation on the performance of the steam genera-tors was analyzed for flow restriction and heat transfer capacity. In or-der to scope the worst case, it was assumed that as many as 10,000 80- inch long sleeves are installed in one OTSG. Temperature distribution and I effect on steam outlet temperature were also analyzed.

The likelihood of flow-induced vibration of a sleeved tube was evaluated.

I It was assumed that an 80-inch or a 30-inch sleeve was installed in any tube in the generator, and that the tube'may be severed or unsevered. The sleeve adds mass which tends to reduce the natural frequency, but it also adds stiffness which tends to increase the damping. The net effect of the sleeve is not obvious, and all r,vbinations were considered to assure that l

the worst case was identified.

A sleeved severed tube was analyzed to confirm compliance with Nuclear Regu-latory Commission (NRC) and ASME Code requirements. This included the cal-culation of required thickness for normal operating loads, faulted condi-tions, and primary plus secondary (thennal) stresses according to ASME Section III Appendix F, and NRC Regulatory Guide 1.121. The resulting l minimum sleeve wall requirement establishes a degradation limit which in-dicates when a sleeved tube must be plugged or removed from service.

lI

)

l The potential for a sleeve to collapse from external pressure was evalu-ated. The bending and straightening operations may leave the sleeve with a slightly oval cross section. Under some accident conditions, there could

^

2-7 gatacock & Wilcox l a McDermott company L

be a secondary-to-primary pressure difference. If a sleeved tube is sev-ered, an oval sleeve could conceivably be subjected to this external pres-sure. If a leaking tube subsequently self-seals, water trapped between the sleeve and tube may also tend to collapse an oval sleeve. The consequences of a collapsed sleeve were also evaluated.

The variables of the roller expansion process were analyzed to assure that adequate roller expansions could be reliably reproduced.

The result was an engineer-ing requi rements document which established limiting conditions for the rolling process which are dependent upon the dimensions of the tube to be sl eeved.

I Table 2-1. Load Conditions Seismic Loading case Static Transients and LOCA Design DW DP OBE Level A/B DW OP, ML, HL (consideration of OBE all specified level A/B ther-mal transient loading combina-

! tions)

Level D 1 DW OP, HL SSE 2 DW OP, HL LOCA DP -- Design pressure OP -- Operating pressure 3 DW -- Dead weight SSE -- Safe shutdown earthquake W l ML -- Mechanical loads OBE -- Operating basis earthquake i

HL -- Hydraulic loads g u

i I

E 2-8 gabcock &WIfcom I

a uronmos: company

I '

. l l

I I

I E l l

lI '

I i

!I I

I I

'I I

f l l I

i I >

I s

2-11 gakock &M8com 4

a McDermott company

I Table 2-3. Steam Generator Design Parameters Primary side Secondary side Design pressure, psig 2500 1050 Design temperature, F 650 600 Full-load flow,106 lb/h 65.6 5.6 Full-load operating pressure, psia 2200 925 E (OTSG inlet) (Steam outlet) E Full-load inlet temperature, F 604 460 Full-load outlet temperature, F 554 570 l

Full-load pressure drop, max. psi 33 50 I

I I

I I

I 1

l I I

i l

I I

I I

l 2-12 M &Mhz j ,

a u<oermois company g

I I

I l I 3. RESULTS AND CONCLUSIONS I

3.1. Leakage Tests

~

The expected leakage per sleeved tube was determined by measuring the leak-age of each sample over a range of axial loads, fitting regres-sion lines to the observed data, and factoring the control or as-installed specimen leakage by the ratio of service cycle leakage to control leakage.

This results in an end-of-life leakage curve that can be compared to the as-installed leakage curve. The maximum expected leakages for normal opera-tion and accident conditions are read from these curves at the appropriate axial loads. They are conservative because the samples were made using low-yield tube material, which is the worst case, and conservative applied I service cycles.

The mean value of measured leakage fJr each applied axial load is plotted I in scatter diagrams in Figure 3-1. The control , axial / thermal , pres-sure , and vibration-cycled groups of specimen data are each represented by l

a linear regression line fitted by minimizing the sum of the squares of deviations from the line.

I Leakages for zero axial load were not available because internal hydrostatic pressure on the tube end- plugs created an axial load of almost 400 lbs.

f l The regression lines of Figure 3-1 indicate that leakage is I

Table 3-1, these relation-ships have been expressed as factors that are ratios of conditioned leakage to control leakage at each test load. The cumulative effect of all the conditioning could be expressed as the product of these three factors, but l

l l

3-1 Babcock &Wilcox a McDermott company j

T' I

the healing effect of cycling has not been confinned in conjunc-tion with vibration cycling. Rather than underestimate the predicted leak-age, all ratios of less than 1.0 were arbitrarily increased to 1.0 in order to assure conservatism. The resulting overall service factors and cumula-tive leakage af ter 40 years' simulated service are presented in Table 3-1 and plotted in Figure 3-2. Within the range of normal operation the maximum leakage is well within khe 2.5 ml/h average leakage objective.

The tightness of a ralled joint is highly dependent upon the yield strengths of the materials used. In this qualification, both tubes and sleeves were type 600 Inconel per ASME Specification SB-163, which pennits a minimum yield strength of 35,000 psi. All sleeves were made from a single heat of material which had a yield strength of I

Figure 3-3 includes a scatter dia-gram of the mean measured Icakage of these high-yield samples at various axial loads and a linear regression line fit to these points. A comparison wi th the regression line for the low-yield tubes, also plotted in Figure 3-3, confinns that the stronger tubes are also much tighter. Table 3-2 tab- g ulates specific values fran these two curves and the ratios of these leak- 5 ages are used as factors to predict the leakage of high-yield service-condi-tioned sleeve joints at various axial loads.

Two samples were tested to assess the effects of test temperature on tube leakage. When the specimens were maintained at 388F, the mean leakage with

! no axial load was At the ambient temperature of 67F, the corre-sponding mean leakage was -

Although it was not practical to leak test at the full design temperature, the hot test results show Since both sleeve and In service, the annulus between sleeve and tube may tend to insulate the tube so that the sleeve is hotter I ,

1 l

l 3-2 shock &WWilcon a ucoermore company g

1

Samples were tested to evaluate the effect of tube surface condition on the quali ty of the joint.

I I

The measured leakage of these specimens is tabulated in Table 3-3. The tube samples all show very little leakage with no significant differences The mean leakage I

I Space limitations require that the sleeves be bent into a gentle arc out-side the OTSG and straightened as they are fed into the tubes.

E Two sample sleeves which had been bent and straightened were expanded into tubes and leak tested to verify that the insertion process does not degrade the quality of the expansion joint. The measured leakage is tabulated in Table 3-4. As these expansions were made in high-yield tubes, the leakage in similar high-yield samples which had not been bent and straightened is also listed for comparison.

The measured leakages are quite low at all axial I loads, I The axial load on a tube during operation is a function of the pressures, the position of the tube in the OTSG, and the tube and shell temperature di f ference, which in turn is a function of the service transient. For normal operation, the transient that results 'in the greatest total tube load is a cooldown from 15% power. Duri'ng this transient, the load on a l

.,_u ,_

I .3 a McDermott company

I center tube reaches 649 lb while the load on a peripheral tube reaches 1107 lb.12 The worst accident condition is a main steam line break (MSLB),

which results in a 3140-lb total load on a peripheral tube of high-yield strength,12 or 2620 lb if the tube has a low-yield strength, according to Figure 2-2. For a central tube, the maximum load of 1585 lb is a result of a loss-of-coolant accident (LOCA). These loads are plotted in Figure 3-4 against expected leakage curves for high- and low-yield . tubes. This indi-cates that under maximum operating load, the greatest expected leakage for a peripheral tube would range from It also indicates that under maximum accident load, the greatest leakage for a peripheral tube In order to p.* edict the leakage in a sleeved 0TSG, it is necessary to make some assumptions regarding the location and yield strength of the sleeved tubes. Figure 3-5 shows the radial distribution of upper tubesheet tube ed-dy current indications at a typical plant.

l l In Figure 3-4, dash lines are shown connecting these mean loads between the high- and low-yield leakage lines. It is reasonable to assume that the yield strengths of all of the

! tubes in an OTSG are for a typical OTSG in Table 3-5. Thus, the predicted maximum leakage for under normal l l operating loads If there were 10,000 41eeved tubes in a j plant (and all of the tubes leaked), the predicted leakage in nomal opera-l tion would be of the usual

( Technical Specification plant shutdown limit. For accident conditions, the i rate would be These predicted leak rates are conservative for nomal expansions. It is difficult to quantify the amount of conservatism, but factors that contrib-ute to the overall conservatism are as follows:

3-4 Babcock &WHcox a McDermott company

l l i I i

1. It is assumed that all the sleeved tubes have through-wall defects. If a degraded tube remains leak-tight, sleeve leakage would not contribute to overall primary-to-secondary leakage.

2. The maximum operating load is not continuous. It occurs only during The axial loads during steady state and the lesser transients are much lower.

3. The leakage at is consid'rably e less than However, all of the leakage predictions are based upon 4 tube surfaces tend to result in tighter rolled joints, but some of the qualification tests used In the event of a complete failure of a rolled joint, such as a full circum-ferential tube crack at the lower roll transition, the tube and sleeve have been designed to remain engaged under worst-case accident conditions The maximum leakage which could occur in such a failed tube has been calculated as For comparison, the maximum leakage for an unsleeved ruptured tube has been calculated 3.2. Joint Strength Tests The expected joint strength was detennined by measuring under a range of axial loads, fitting regression lines to the observed da ta ,

i This results in an end-of-life curve which can be compared to the as-installed curve.

The mean value of measured each applied axial load is plotted in scatter diagrams in Figure 3-6. The groups of sperimen data are each represented by a linear regression line fitted by minimizing the sum of the squares of deviations from the line. The sample data are al so plotted in Figure 3-7 I

I 3-s - - = = = -

. A McDermott company

Il Il Il 3;

The regression lines of Figures 3-6 and 3-7 indicate that joint strength is reduced the individual conditioning samples at each axiol load is tabulated in Table 3-6 and plotted in Figure 3-7.

This represents the cummulative effect of all the conditioning.

i I

The strength of a rolled joint is dependent upon the yield strength of the I

materials used. In this qualification, most of the tubes were made from I

All sleeves were made from l material . Figure 3-8 includes a scatter diagram of the l samples at various axial loads and a linear re-gression line to fit these points.

Table 3-7 tabulates specific values from these curves, l

I Samples were tested to evaluate the effect of tube surface condition on the l

strength of the joint. Table 3-8 reports the various surface samples at several axial loads.

I 3-6 hock &M8com a ucoermore company l

t

I I

I I tube joints ma'de with sleeves that had been bent and straightened were measured at various axial tube loads. The results are presented in Table 3-9 and campared to joints where the sleeve had always been straight.

1 I l I The expected displacement of the rolled joint after 40 years of service is compared to the design requirement in Figure 3-9. The acceptance limits for high- and low-yield tube material from Figure 2-2 have been replotted, and the maximum operating loads for center and peripheral tubes have been added in the manner of Finure 3-5.

It is evident that the expected joint is well within the acceptance limits for both maximum operating and accident loads.

I The are based upon low-yield tubes on the outer I periphery of the OTSG undergoing the maximum cooldown rate or maximum hypo-thetical accident conditions. They also presume that the sleeved tube is totally severed so that the rolled sleeve joint must carry the entire axial load on the tube. In actual practice, the tube would normally share the load with the ' sleeve because a severed tube is quite rare. As previously discussed in section 3.1, E,

Thus, it would be a relatively rare occurrence for a sleeve in a low-yield peripheral tube to be subjected to the full max-imum operating loads. However, when this happens, the rolled joint is cap-able of carrying the load without failure.

g 3., - ,_

a McDermott company

I Although no acceptance criteria were established for the ultimate failure of a rolled joint, most of the test specimens were pulled to failure.

These failures included l failed specimens had a mean failure load maximum load that a joint could experience under accident conditions.

3.3. Light Expansion Tests The tightness of a rolled joint depends upon the amount of roller expan-sion. In the qualification tests, all sleeves which were rolled at the tubesheet end had a regardless of tube dimensions and strength. The

. free-span expansions were ,

Figure E

3-10A plots test data which confinn a diameter increase depends on In the event that I

may have less strength and leak more.

assess the effect of light expansion. Figures '3-11 and 3-12 show the scatter diagrams of the light expansion mean measured leakage and joint strength and the regression lines for both light and normal expansion.

Tables 3-10 and 3-11 tabulate specific values from these curves, and the determine the leakage and strength of service-condi tioned lightly expanded joints at various axial loads.

The effect of the amount of roller expansion 09 the quality of the joint is also shown in Figures 3-13 and 3-14, Again, I

3-8 mayylicou a uconmore company l

I I they show that I they clearly indicate that the 9ffect of light expansions on both leakage and displacement is The sensitivity of the joint to light expansions is I -

The roller expansion tool I

However, these dimensions vary as I illustrated by the sample distributions of Figure 3-15. The normal expan-sfon that has been qualified I

I I

l lI

(

lI I

I 3-9 Bah &Mm a McDermott company

'I

The effect of light expansions on an OTSG can be estimated by applying fac-tors to the expected portion of light expansions. Figure 3-16 indicates I

I I

Figure 3-17, which is a modified version of Figure 3-10, shows within the strength acceptance limits. If the expected leakage curves of Figure 3-4 are as shown in Figure 3-18. As described in section 3.1, these I

I I

the predicted leakage for the total plant would be of the usual plant Technical Specification limit. Under the worst normal operation loads, the expected mean leakage

! I i

3.4.1. Corrosion Tests Accelerated stress corrosion cracking tests were performed on sleeved tube mockup specimens to determine whether residual stresses from the sleeving process are sufficient to cause stress corrosion cracking of the sleeve or the OTSG tube. The mockups were rolled at the tube-l sheet joint, In the free-span expansion, the tube and sleeve walls were respectively, which is greater than expected in a normal expansion. In the tubesheet expan-E sion, the tube was in the initial roll and the sleeve was E by the sleeve rolling, which is the normally expected I

3-10 l ,ucm , ,,,, ,,,ny

r 1 i

I I amount at Specimens were removed from the tubesheet sleeve joints, free-span sleeve joints, and bent and straightened sleeve I samples and mounted in autoclaves along with control specimens from both U-bend and virgin sleeves and tubes. In the autoclaves, they were exposed to a 10". sodium hydroxide solution at 550F with a +190 mV applied potential The results of the corrosive tests are listed in Table 3-12.

~I l I

All U-bend con-trol specimens cracked which demonstrates that the test was rigid enough to produce severe cracking in hignly strained specimens.

Use of the accelerated caustic corrosion test results to predict failure in service with all-volatile-treated (AVT) water is based In autoclave tests, the time to crack initia-tion was measured at various strains for specimens subjected to 650F AVT water and 550F 10% caustic solution with an applied electric potential .

I a linear curve was fit to define the relationship between failure times Although the data are limited, this corralation indicates that a I

I 'These correlations are for well controlled AVT water and may not apply to other water treatments I or water with impurities which may be present in steam generator opera-tions.

3-11 Babcock &WIfcom a ucoermott company

I l I l

i I

Actual steam j generator water chemistries and loading conditions may reduce this predic-tion I

I l

3.4.2. Flow-Induced Vibration Analysis I

The effect which a sleeve has on the vibration characteristics of a tube is not obvious because the sleeve stiffens the tube and tends to increase the system damping. However, the additional mass tends to reduce the natural freque ncy. The characteristics of the sleeved tube depend upon I'

I I

l i

i I I I

These averaged curves are plotted in Figure 3-20. '

I 3-12 M & Micom a ucoermott company l

I I Finite element models were set up with different assumed rotational spring ,

I I constants at the lower face of the tubesheet in order to match the computed and the observed frequency in the mockup. The modal frequencies compare as follows Computed Hz Mock-up Hz Unslee/ed tube Sleeved tube j i

Sleeved tube severed 1 at tubesheet  !

I The response plots from the in-air tests (Figure 3-20) were used to deter-mine approximate critical damping ratios These i values were rounded off to conservative test damping values, and then more conservative operational damping values for the analysis were obtained I These damping values are as follows:

' Hal f-power Test Assumed for method (rounded) analysis Unsleeved tube Sleeved tube I

Sleeved tube severed at TS Sleeved tube severed at TSP A NASTRAN finite element model was used for the analysis wi th a at the tubesheet face and damping as listed above. Secondary side crossflow velocities of OTSGs with both internal and external auxiliary feedwater headers were evaluated. The resulting worst-case generic fluid-elastic stabili ty margins and randam I vibration and vortex shedding responses are as follows:17 3-13 Babcock &Wilcox I a McDermott company

~

I I

Sleeved tube Sleeved tube Unsleeved Sleeved severed severed E tube tube at TS at TSP g Minimum fluid-elastic E stability margin 5 Random vibration response max. displacanent ms, in, max. stress ms, psi Vortex shedding response max. displacement, in.

max. stress, psi I

l Therefore, it is concluded that flow-induced vibration will not be detrimental in any OTSG tube sleeved in the upper span, even if the tube is completely severed anywhere between the upper tubesheet and the number 15 tube support plate.

3.4.3. Strain Tests Specimens were processed to evaluate the sleeve and tube elongation due to sleeve rolling. During the tubesheet roll, the sleeve is I

I l

I I

3-14 Babcock &Wilcox g a McDermott company g i

I -

I I

I 3.4.4. Adjacent Tube Tests There was some concern that rolling a sleeve in a tube adjacent to a tube which previously had been sleeved may loosen the first expansion, causing increased leakage.

I I it is conservative when applied to rolled sleeves in the OTSG tubesheet.

3.5. Analysis 3.5.1. Performance The installation of a significant number of sleeves in the OTSG could re-I duce the OTSG's thermal performance due to the insulating effect of the sleeve (especially the annulus between sleeve and tube) and the change in primary flow distribution caused by higher flow resistance. The net result of these effects I The analysis of thermal and hydraulic effects assumed that  :

5000 80-inch long sleeves were installed in the peripheral tubes of each OTSG. These worst-case assumptions reduce primary flow by I The effect of this reduction in superheat temperature on plant opera-tion is considered to be minimal . The first OTSG put into operation was I warranted to produce a minimum of 35F superheat steam at full power, The new oper-ating point for the OTSGs, turbines, and feedwater control system would be I

3-15 Babcock &Wilcom a McDermott company

I I l I

3.5.2. Structural Tlie minimum acceptable wall thickness for degraded sleeves was detennined in accordance with the allowable stress and pressure limits of ASME Section III and NRC draft Regulatory Guide 1.121. Primary membrane stress, burst g pressure, and fatiguc analysis were considered for normal operation, and a primary membrane stress, burst pressure, collapse pressure, and primary mem-brane plus bending stresses were considered for postulated accident condi-tions. In addition, primary plus thennal stresses were evaluated. The minimum sleeve wall thickness was calculated for these eight different ac-ceptance criteria. For the expected type of defects, the greatest required minimum wall was found to be a 70% through-wall defect would require that a sleeve be removed from service. This compared to a 69% defect limit for the OTSG tubes.

I The sleeve must be bent and straightened for installation in the outermost OTSG tubes. This results in a slightly elliptical cross section, which was evalua ted for buckling pressure. The maximum expected ovality (i.e., dif-ference in extreme ODs at any one cross section) was found to be inch based on sample dimensions.2 The critical external pressure depends on the material yield strength.

I Under the maximum secondary pressure of 1050 psi with no primary pressure, neither tube nor sleeve would collapse.

In the Obrigheim steam generator, tube blisters were found inside the tube-l sheet between tubesheet rolls. This plastic deformation of the tube was at-l tributed to water that had leaked through roll transition cracks and become trapped in the annulus between the tube OD and the hole ID by corrosion products. Upon heatup, the water expanded more rapidly than it could leak out, causing the tube deformation. Should water be trapped in a similar 3-16 Bat > cock &WHcom a McDermott company

manner between the OTSG sleeve OD and tube ID, I

I the annular pressure increase is more likely to binw out the corro-sion products which plugged the leak than to collapse the sleeve. Thus, the likelihood of sleeve collapse is very small.

3.5.3. Process Control The parameters of the free-span expansion were evaluated to assure that the amount of expansion can be controlled within a qualified range. A general equation was developed to define the relationships among tube, sleeve, and expansion tool parameters before and after expansion.13 This equation is:

'I I

'I This process was qualified for a range

!I I

I I

l 3-17 Bakock &Wilcox a McDermott company I

l t

Mest of the qualification samples were rolled to represent the normal range of expansion.

E to evaluate the effect of light expan- E sion on leakage and strength, and another set was rolled at an average to evaluate the effect of cold working on corrosion resistance.

The range of nomal expansion is while the range of assure that any tube diameter will be expanded within the qualified range.

I I

I I

I I

I I

I I

s 3-18 Babcock &Wilcox a McDermott company

I I

4.

SUMMARY

AND RECOMMENDATIONS I Mechanical tube sleeves have been qualified for use in degraded OTSG tubes by a series of tests and analyses. The design is strong enough and suf-ficiently leak-free to be used as a permanent remedy to keep degraded tubes in service. The criteria for this qualification have been summarized in Table 4-1 along with the results, which show that all the criteria are sat-isfied.

It is recommended that up to 10,000 of these mechanical sleeves be in-stalled in the OTSGs of any plant as needed to correct or prevent tube degradation which would otherwise require that the tube be removed from ser-vice.

I

.I I

'I I

lI I

I I

I 4-1 E E" a McDermott company

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

Table 4-1. OTSG Sleeve Qualification Criteria Reference Sleeving criterion Justification Results section Installation Capable of installation in any tube in Degradation has been found in tubes Tooling designed to pemit the Installa- 2.1 an OTSG at depths up to 80 in. below across the face of the OTSG, concen- tfin of an 80-in. sleeve in any tube in the top of the tubesheet. trated at elevations in or near the upper an OTSG.

tubesheet and at the 15th tube support pl a te.

Leakage Average leakage of no more than 10,000 sleeves with this average leak The average leakage for a sleeved tube 3.1 per sleeve under nomal operating con. rate could be installed in a plant, and undar maximum operattnq loade was found to ditions. the overall steam generator primary to be under maximum secondary leakage would be only one-tenth loads.

of the leakage which requires plant shut-down.

3 Strength b Pullout strength of at least Accident conditions could load a sleeve No joints failed at loads below the tse 3.2 up to the yleid lief t of the t@e in yleid strength, which it is insta11eo Process Control The sleeve installation process must be The ilmf ts of the amount of expansion light expansions are expected, 3.3 controllable so that predictable joint must be established in order to adequate- and these would have adequate strength quality is maintained. ly control the expanston process, e

$ Corrosion 2 The corrosion resistance of the installed The tube / sleeve assembly is required to Accelerated corrosion tests representing 3.4.1 g

o sleeve is to be at least as good as that remain functional in the OTSG. primary and secondary coolant tests indt-D of the original OTSG tubes when stbjected cated '

to condltions expected in the OTSG. sleeved tees w11 not cract 4 M

.M .M M .m W W W <

M M M M M M W W W W W

m M M M M M M M M M M M W _ m W Table 4-1. (Cont' d)

Reference Sleeving criterion Justification Results section Vibration A sleeved tube, either severed or un- Tube / sleeve integrity is required under The sleeved tube, whether severed or not, 3.4.2 severed, is to be adequate for the 40- normal nperating conditions. has a smaller maximum displacement and a year design life as confirmed by an greater fluid-elastic stability margin ASME 111 f atigue analysis. than the unsleeved tube, and the maximum stresses are well below allowable.

Tube Strain Sleeving installation should not leave Dif ferential expansion in postulated ac- Sleeving will increase the tension in a 3.4.3 -

a tube in compression or pennit tube to cident conditions must not permit adja- tube. Maximum s11ppage tube contact. cent tubes to contact each other. would not bow the tube in neatup enough to contact an adjacent tube.

Adjacent Sleeves p Sleeve installations must not damage a Quantity sleeve installations must not be Adjacent sleeve installations are unaf- 3.4.4 w prior sleeve installation in an adjacent restralned by sequence or location Ilm- fected by a new sleeve Installation.

tube, i ta tions.

Performance The ef fect of sleeving up to 10,000 The thermal / hydraulic ef fects of sleeving 10,000 sleeved tubes would reduce primary 3.5.1 tubes must have a tolerable ef fect up- must be acceptable, flow and reduce full power on plant performance. steam umerneat The effect of this would have a minimal effect on plant operation.

Code Tube / sleeve funtional Integrity must be Tube / sleeve integrity is required under The minimum required sleeve wall for nor- 3.5.2 maintained under stress and pressure postulated accident conditions. mal and accident conditions

., I Ilmits of AS4E Section III and NRC pennits sleeve defects less than /tts through-wall.

g Regulatory Guide 1.121.

k$ Collapse BQ A sleeve must not collapse under exter- Sleeve collapse in a leaking tube would The sleeve is about as strong as 3.5.2 2, g nal pressure. create a primary to secondary leak path, the tube under external pressure, and g collapse is unifirely, a

i-

=8M

I

5. REFERENCES

1. 177-FA OTSG Tube Sleeve Design Criteria, B&W Document No. 51-1146626, Babcock & Wilcox, Lynchburg, Virginia, February 1984.

2. 177 OTSG Tube / Sleeve Loads, B&W Document No. 32-1147602, Babcock &

Wilcox, Lynchburg, Virginia, May 1984.

3. Qualification Specification for OTSG Mechanical Sleeves, B&W Document No. 11-1147592, Babcock & Wilcox, Lynchburg, Virginia, November 1983.

4. Test Plan for OTSG Mechanical Sleeve Qualification, B&W Document No.

51-1148156, Babcock & Wilcox, Lynchburg, Virginia, October 1983.

5. Sleeve Installation Rolling Procedure, B&W Document No. 51-1148145-01, Babcock & Wilcox, Lynchburg, Virginia, February 1,1984.

6. Initial Leak Test Procedure, B&W Document No. 51-1148146-01, Babcock &

Wilcox, Lynchburg, Virginia, February 1,1984.

7. Thermal Cycling Conditioning Procedure, B&W Document 51-1148147-01, Babcock & Wilcox, Lynchburg, Virginia, February 1,1984.

8. Axial Load Condi tioning of Sleeved Mockups, B&W Document No.

54-1023433-02, B&W R&D Division Technical Procedure ARC-TP-611, Rev.

2_,

Babcock & Wilcox, Lynchburg, Virginia, November 15, 1983.

9. Vibration Condi tioning of Sleeved Mockups, B&W Document No.

54-1023434-02 and B&W R&D Division Technical Procedure ARC-TP-612, Rev. 2, Babcock & Wilcox, Lynchburg, Virginia, November 10, 1983.

Conditioning Procedure, B&W Document No.

10. Cyclic Pressure 51-1147649-00, Babcock & Wil cox, Lynchburg, Virginia, October 19, 1983.

11. Leak and Load Pressure Testing Procedure, B&W Document No.

51-1147650-01, Babcock & Wil cox, Lynchburg, Virginia, February 1, 1984. s h M Icom I 5-1 a McDermott company

I

12. Determination of Minimum Required Tube Wall Thickness for 177-FA OTSGs, BAW-10146, Babcock & Wilcox, Lynchburg, Virginia, October 1980.

13. Mechanical Sleeve Free Span Expansion Study, B&W Document No.

32-1148881, Babcock & Wilcox, Lynchburg, Virginia, February 1984. ,

14. G. V. Theus, Stress Corrosion Cracking Tests of Alloy 600, EPRI WS-80-136, Electric Power Research Institute Proceedings, Palo Alto, California, June 1981.

15. Thermal / Hydraulic Performance --

80" Sleeves, B&W Document No.

32-1149041, Babcock & Wilcox, Lynchburg, Virginia, January 1984.

16. Flow-Induced Vibration Analysis of Three Mile Island Unit 2 Once-Through Steam Generator Tubes, EPRI NP-1876, Electric Power Re-search Institute, Palo Alto, California, June 1981.

17. Flow-Induced Vibration Analysis of Sleeved Tubes in the 177 OTSG, B&W Document 32-1151170, Babcock & Wilcox, Lynchburg, Virginia, June 1984.

I I

I

~

I I

I I

- . m., a 5-2 ,u,o,,,,,,,,,,,oy g

. - - . . _ _ _ _ .