ML20081A484
| ML20081A484 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 06/30/1983 |
| From: | WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19268E253 | List: |
| References | |
| SGPR-8302, NUDOCS 8310260286 | |
| Download: ML20081A484 (180) | |
Text
{{#Wiki_filter:- - - .
.- -. ., i-COUNTERFLOW PREHEAT STEAM GENERATOR TUBE V!BRATION
SUMMARY
REPORT SGPR-8302 June 1983 L WESTINGHOUSE ELECTRIC CORPORATION Nuclear Energy Systems P.O. Box 355 Pittsburgh, Pennsylvania 15230 i-SSNO P A 5 PDR 0855c/0119c/060383:5 1-
Table of Contents
-1. 0 INTRODUCTION
1.1 Background
1.2 Design Modific.ation Objectives 2.0
SUMMARY
f.ND CONCLUSIONS - 2.1 Field Data 2.2 Test Model Data 2.3 Analysis Models. 2.4 Wear Coefficient Tests 3.0 DATA COLLECTION 3.1 Operating Plant Data Collection 3.2 Test Model Descriptions 4.0 0ATA ANALYSIS' 4.1 Tube Vibration Measurements in the 16* Model 4.2 Velocity Distributions 4.3 Force Measurements 4.4 Work Rate Wear Coefficients 5.0 VIBRATION MECHANISM 5.1 Introduction 5.2 Summary and Conclusions 5.3 Mechanisms of-Flow-Induced Vibrations 5.4 Effect of Steady Fluid Forces on Tube Vibration 5.5 _ Indications of Fluidelastic Excitation in Counterflow Steam Generators ~ 5.6 References 6.0 DiERMAL/ HYDRAULIC CO. NSIDERATIONS 0855cIO119c/060383:52 [
T , a 1 Table of Contents (cont.) 7.0 WEAR ASSESSMENT METHODS / ANALYSIS RESULTS 7.1 Ga Method 7.2 'Non-Linear Flow Induced Vibration Analyses-Work Rate Method 7.3 Linear' Flow Induced-Vibration Analyses 7.4 Wear Depth to Volume Relationships-7.5 Statistical Evaluation of R49C53' Pre-Load Forces 8.0 SELECTION OF TUBES "0R EXPANSION / MODIFICATION WEAR ASSESSMENT 8.1. Data Base for Selection of Tubes for Expansion 8.2 Selection.of_ Tubes for_ Expansion 8.3 - Design Bas'is Modification Wear Assessment
- 9.0 SAFETY CONSIDERATIONS 1
0855c/0119c/060383:5 - 3 Il
1.0 TNTRODUCTION
1.1 Background
Preheater region tube wear was initially identified at Ringhals Unit 3, a plant with Model 03 steam generators. On October 21, 1981, tne plant was shut down due to an aoproximate 2.5 gpm primary to secondary leak. Subsequent examination revealed a through-wall hole in a single steam generator tube at a baffle plate elevation. That tube was in the outer row of the tube bundle
~
facing the feedwater inlet. Eddy current testing (ECT) of the steam generator tubes indicated tube wall wear in some tubes in the outer rows near the inlet nozzle. ECT was then performed at Almaraz 1,. a nondomestic plant with 03 steam generators: at McGuire 1, a domestic plant with Model 02 steam The generators; and at Krsko, a non-domestic plant with 04 steam generators. ECT of Almaraz 1 also revealed indications in tubes in the outer rows of the bundle. The plants with Model 02 and with Model 04 steam generators nad no wear . type ECT indic ations. These two plants had not operated above 50 percent power. To better understand the mechanism causing such wear, Westinghouse conducted an extensive test and evaluation program encompassing data collected from operating plants and laboratory test models. Additionally, analytical, models to estimate wear rates in tubes were developed. This report provides a summary of the data . collection and analysis effort undertaken by Westinghouse and the assessment methods used in estimating tube wear resulting from flow induced vibration in counterflow preheat steam generators. An. elevation view of a counterflow preheat steam generator (Models 04, 05 and E) is shown in Figure 1.1-1. As depicted, the preheater region is located on the cold leg side of the tube bundle and f aces the feedwater inlet. The lower shell internals including the preheater region are shown in Figure 1.1-2. Incoming feedwater enters the inlet water box and encounters the impingem,ent plate which directs the water outward to fill the water box volume and
-downward to the preheater inlet pass located between the B and 0 support plates (baffles). The water enters the tube bundle, then flows upward around 0855c/0119c/060383:5 4 l-l .
the tubes and baffles. It is in the outer rows of the tube bundle f acing the incoming feedwater that tube indications have been observed.
'l . 2 Design Modification Objectives The Model 04/05/E design modification objectives were:
- 1. To reduce tube vibration levels to levels comparable in modified Model 02/D3 steam generators and to vibration levels obtained in the Krsko steam generators while operating at a 70/30 feedwater flow split.
- 2. To reduce the potential for fluidelastic instability.
- 3. To achieve acceptable impact of the modification on the overall plant performance.
4 To assure that conservative wear estimates satisfy safety limits for tuce wall reduction. The first objective'is based on relating vibration levels to levels that have been shown to result in low wear rates in plant operation. The 02/03 wear rates with the modification are comparable to plant operation at 50 percent power prior to the modification. Figure 1.2-1 provides the design objective Ga values from Krsko at 70/30 and for 02/03 based on full scale tests with the manifold modification. The 02/03 measured Ga values have been[
.c.e.
to account for the ratio of Ga average wear coefficients For wear assessments based between 02/03 and 04 as discussed in Section 7.1. on work rate methods such as the nonlinear analysis model, the 04 design objective is to obtain impact sliding work rates, *less than the upper bound A c.,e.- l based on the 02/03 range of value estimated for 02/03 ofb e J l- Ior . Row 49 tubes. g The second oojective is to reduce the potential for fluidelastic instability. both the bypass flow and the expansion of tubes in high velocity regions In addition, plate contribute significantly to achieving this objective. position searches (in the 16' water model) were conducted to enhance the 0855c/0119c/060383:5 5 g-2
m. potential for vibration by either turbulent or fluidelastic excitation. Those
^
vibration levels observed were then utilized in the modification evaluation to conservatively account for fluidelastic excitation. if present. The third objective, to obtain acceptable impact of the modification on overall plant performance has been satisfied by selection of .the reference modi f ic ation. The mndification is a combination of tube expansion and feedwater bvoass to the steam generator auxiliary nozzle for Models 04 and 05 p l an t s. For plants with Model E steam generators, the modification c6nsists of tube expansion only. By utilizing tube expansion for tubes in the high vibration regions, the amount of bypass required to achieve the objective vibration levels in Models D4 and D5 is substantially reduced. The limited bypass- flow of the modification reduces the number of tubes requiring expansion and reduces the expanded tube wear rate by reducing the flow induced excitation' forces. This modification concept can be viewed as a three zone concept as schematically shown in Figure 1.2-2. With limited bypass flow, the tubes having potentially significant vibration are reouced from Zones 11 + 111 to only~ Zone 111. Detailed test and analytical evaluations, described in this report, were used to define Zone III. By expanding the tubes in Zone Ill, the tube vibration levels for these tubes are reduced to the design objective levels. The fourth objective of satisfying safety limits for tube wall reduction will be satisfied by conservative wear analyses between initial operation after tne modification and the first eddy current test. Af ter the initial eddy current test, satisf action of acceptable wear rates for safety considerations will be based on estimated wear rates based on rent results. 0855c/0119c/060383:5 6 _ l- $
1 a4al f /s% f n . D
'_/ .
Ml .
. 3
-( 6 I
=
- @ e,f-.r ..
p .
.e P "
g, ._. 3 r . mezzi. rune annone r-- a , n tti tti 7""*
- N , lil lll N
*"~
tii tiI III aaw [ em masser omms: Feesumeser Neszte 3 ,
+t _/
i I * -
% pr ses, s. amen g
N Tues snessN w pftIIIEr? CastafN N _
% Casemen Channes Fiqure 1.1-1 Counterflow Preheat Steam Generator 1
o . p sus. M-WIAPPER
$ 0 % $ th 4 /
/
1 1 % s -
- CAP PLATE PARTIT1t3s g
,,,,J.,-
Pt.n. t ,.
, , N
(- ___ as A ' t f f t 0 \
)
C f n < ( N N uotwaterr Purt - , r i S / / I l ! ! ! !
% % % % %A % b
/ 4 s /
s s s s s s s. l Figure 1.1-2 Counterflow Preheat Steam Generator Lower Shell Internals 9 O
-w. - - - - , . < , ,,--...-,um -
-,,m,y-- , - , - , ,-,..* . , , - , ewv, ,..,.--,y-we,wyww.w,wp-,-=m-ewws-.,,,w..e.
.eAm, _ 4 .* A 4 - p ..a - 4 -- -.-- p e .b
<r
-d . . .
6 i i N w
-h >'-.
.W J
<C
! C L
O
. so-A U
< t 9 8 C C. 8C I i U
=
b C
. tP e
e. w - N
@. Y C
N w O 0 (Ne ' ,
,y .
I i W l 8 6 l . I I ' l i I
, . . . . , , .- _ .~.. . ..._. . .-. . . _ , - .___.. . . _ ... . .. ._ . . .__._. . _ __ , _ . - _ - . _ . . _ _ . ,,,_._...- . . . _ . _ . . . . _ . , _ _ ,
2.0
SUMMARY
AND CONCLUSIONS The.following discussion provides a summary of the data collection and a.nalysis effort undertaken along with conclusions related to the mechanism and extent of wear on tubes located within the preheater region of counterflow preheat steam generators. 2.1 Field Data 2.1.1 Vibration Data . Operating plant tube vibration data were obtained at Krsko caring three phases of plant operation. These data provided information related to a range of' plant power levels, main and auxiliary steam generator nozzle flow rates, and feedwater temperatures and pressures. A total of 17 tubes were instrumentea as shown on Figure 2.1-1. Accelerometerswithinthese17tubesshowed( a ,b ,c., e. at 100 percent power and main feedwater nozzle flow rate. At 70 percent main nozzle /30 percent auxiliary nozzle feedwater flow, the highest Ga obse> ve was t 7.s,<.e-One tube (R48C55) was expanded at Krsko prior to tne third phase of testing. This tubgrior to expansion displayed at a peak Ga of(
. *Afterexpansion,thepeakGaobservedwas{ .
[ > "' The field ' data sh> that the primary response frequency varies fran tube to tube and can vary for a given tube with power. The frequency of the dominant response depends on the pattern of contact between the tube and the support plates. Linear analysis nodels have been used to relate measured responses to tube support conditions. Figure 2.1-2 shows the range of frequencies expected fo.r unsupported spans based on single span calculations. This approach is effective in correlating D2/03 data and agrees with multi-span linear model analysis results summarized on Figure 2.1-3. Based on this correlation, the support conditions associated with the main response frequencies in the Krsko z -\ -
plant data (and 16* model dats) are as shown in Table 2.1-1. The occurrence of these frequencies on the Krsko data is shown on Table 2.1-2. Depending on the magnitude of tne mean force between a tube and support plate and the magnitude of the fluctuajing force acting on a tube, impacting can o cc ur. The tube wear potential has been correlated to the Ga parameter wnich
. c. ,e_
r The highest Ga reflectsL c , c , e., e values have been f ound for tubes with ,a dominant response around( } Tne maximum Ga ' values from Krsko plant data for the accelerometer at the "E" baffle elevation on each tube are listed in Table 2.1-1. 2.1.2 Palled Tube Evaluations Three tubes were removed from steam generator No 2 at Krsko. One tube (R49:56) was removed in May 1982 after approximately 1600 hours of operation at 75 percent power (70 percent main nozzle feed flow, 5 percent auxiliary nozzle f eed flow. Wear on this tube was not detected by either fielo or laboratory eddy current examinations. Visual and profilometry examinations detected wear at five different baffle plate intersections. The maximtzn wear penetration depth was 2.5 mils (6 percent). Two tubes (R46C56 and R49C35) were removed in November 1982 af ter an additional 2850 hours of operation. Again, eddy current examinations did not detect tube wear. Visual and profilometry measurements did detect some wear, with the maximum wear penetration depth being approximately 2 mils on one tube and I mil on the other. 2.1.3 BaffQ Gap. Preload Measurements Tube-to-baffle gap measurements were made for selected tubes in both Krsko A steam generators and in two steam generators at Comanche Peak Unit No, 1. l I random sample of as-built tube hole sizes was obtained from manufacturing plant. The diametral gap sizes, as determined from these measurements, ranged from[ l ' w .e. i 1 "b
Table 2.1-1 Typical Tube Support Conditions Response Observr- Typical Support Conditions
- _+*
i
- Based on linar model analysis
Table 2.1-2 Krsko Frequency Response Groups - Highest Accelerometer Elevation i I l l. 1 I I l 1 l i l 0855c/0119c/060383:5 9 2-9
7 i ..
~
, i ~8j .
i.. .:!._ .;. ,-_., .
. w w -,
u-
- .< 3. 4. ...
- -. 4 , a
!.i . !. .
.i _ . ,i, . . . . . .Iw4 ::
;;1 i
. . . 4 ./ < ,
.! ... lists M
t((- i f . -t :
._;-. :4. 44 ,- -
i
-ts y7
- 96 ,3
* ; . j . ._ l -
t
*j-.. i i d; 'i i ' -
m u 111 (Sc r) a4 i
*3 N se II (5e2) 7" . . -
-.i:.
v 42
~.
- 1-oo 'i.. .. . I 349 844 C55 CS4 E S-D, t
--.t
---I -F0 i,
*1 349 CE2 8.D, D l'
~
R49 C59 B-D, E 849 C51 E -- 41. _-
, g '
4 Sea C51 E
.\.- - ..
M6 57 R21 C57 B-D B-D, D I
.-i
< .-. \\ - M R49 c35 844 cu e.
s-D, t D 3 u. s3 g., : . . 62 _ 4 - 3 M f ** * ** *"** Y *******T* . 1* m 1Qt . 1,
]
i. n e q& 3 . -: p: ,1 ,....o. ,' .y' -. :w,a. T _. ., 3 ~ .~.: . sA _as -
~~ . T . . _ f, Y,* : .;:.T:s.:
*i.. * ,
- p; I j . t . .n: e. ny-- ,
_ j. , ,-*-. . i, ; . 8 M
- . . = ..- .. e _J t
5 psto FEL.n tests ! l AccELEnoMETER LOCATIONS FIGURE 2.1-1
+- S-w,-r-w -=w,w--. ,,-ev---,,we-,- -r --
me- . ---m,ww, ___u--we,m-,--,- __w,-.-.m%w.=m-,=-nw w-owa em-e -en-w w m- --e-s--
.. - .. =... . . . . . .- -. - . . -. . - .- . -
e 4
^; -
- ., a,C, L 1
I 4 5 9
/
4 d 3 4 1 4
.i t
i t ,J f
,t h
Figure 2.1-2 Natural Frequency.of Steam Generator Tubes for Various Span Lengths at Hot Conditions f g t- -.ve-y , ,%.<. , ,,%y- -%-yw,-, y-*e- e - - , g ee w.,w -..w--++,,,t,-ve,--.,v---v-w-m--e-w-r-wy,--y,--,--,- we,yw,, v w wny-vw w-v v- y<,w e % ,w y- +w -e e W--V
T 6; e N v
.Y%
6 5 I M1.r. -
%usp .P e 5M
= ;;: r.
F, g 6 O<
~
s e E g K ~ an ur m= - E FS g 6 l E
.O
- 2.2 Test Model Data Test models used in the counterflow preheat steam generator program included a 16' f ull scale water model, a 2/3 scale water model, and a 0.95 scale air model. Measurements obtained from these nodels provided information related to vibration characteristics, velocity distributions, and the hydraalic forces acting on tubes.
2.2.1 Vibration Data Accelerometers were installed in selected tubes in the 16* model to characterize the tube vibration response, to simulate Krsko tu'ue vi-bration field data, and to evaluate design modifications. The validity of the 16* model was demonstrateo by initial tests wntcn reasonably reproduced data for the most active tubes observed during tne Phase I testing at Krsko. As demonstrated by Krsko plant data, and test runs in the 16* mooel leading to the du. plication of Krsko plant results, the occurrence of tne most ac
. .c.e_ To enable a conservative
) tube response is sensitive to plate alignment.
correlation between tube response and tube locations, plate searches were
,a,b.<.e-conducted to obtain a( jrerponse for each unexpanded window tube location. As shown in Figure 2.2.1, a spectrtsn of responses is achieved with a systematic change of plate alignment, encompassing the frequencies Plate associated with the typical support conditions described above.
searches were similarly conducted for non-window tubes and expanded tubes to seek a maximum response and investigate the sensitivity of tube response to plate positions. For the modified configuration flow rate (84 percent of In this configuration 18 Krsko flow) the results are 'shown on Figure 2.2-2. tubes near the T-slot were expanded to more closely represent the modified configuration. These Ga data are used in the selection of tubes to be expanded and in the development of correlations with 2/3 scale model data (velocity and fluctuating forces) to permit wear assessments for tubes outside of the 16* model full length boundary (i.e. Outside the region bounded by columns 42 to 73 and selection of tubes to be expanded). 2-6
2.2.2 Velocity Distributions The quasi-steady-state gap average and RMS velocity distributions are sno n in Figures 2.2-3 and 2.2-4. The highest gap average and RMS velocities are near the T-slot. 'the V RMS distribution is, however, reasonably flat across row
~
- 49. This parameter _ drops off significantly as the flow moves into..c.,=.-
the bundle. In general the VRMS at row 46 is approximately{ of tnat at Roa 49. 2.2.3 Force Measurements The turbulence inouced, dynamic forces measured in the 2/3 scale mocel are sh os, in Figure 2.2-5. The data show maximum values near the T-slot 'ans rapidly redating values at rows successively into the bundle. e z-9
1 5
- a,b,c.e
- Figures 2.2-1 thru 2.2-5 contain scale model' test
. data which are considered Westinghouse Proprietary.
1 m I s 1 0020G/FTE/7-28-83 2-IU F w.4. , . - . , . , - * ..m,..
% - - , . - , - - - ,.-w,.,,--,..,,-,w...,.c._,-., , , . , . -,,.,_-.m - ,-mw, .-n,- w..wr., e-..,.r.,-.v~.,, , , . ,
a . 2.3 Analysis Models 2.3.1 Nonlinear Analysis Model The run-linear (FIV) analysis techniques developed during the D2/03 design modification prog-am were also applied in the 04 program. These tecnniques are used to evaluate designs, and proposed design nodifications on the Dasis of work rates and type of motion (rubbing or impact and sliding) . The wear coefficient appropriate f tpe type of rrotion are applied to the work rate to determine the wear volume rate. Initially the non-linear model was used to reproduce wear rates and scars compatible with those observed on the pulled tubes from Krsko. Then, witn the help of a non-linear, preheater model developed to establish nominal D4 plate positions at operation temperature, work rate results were obtaines f or a matrix of cases chosen as representative of all tne D4 tubes. This mat-1x includes both beginning of time (BOT) and series runs, where a seqsence of increasingly larger gaps are introduced for expanded and unexpandet tubes. Limiting wear assessments were performed for expanded and unexpance: tutes for safety related evaluations. ( mIne.addition upper bound wear distribution relations and fluctuating forces were used in these calculations. Results from this matrix of analyses were than used as a basis to assess the modified D4 design and in the development of single plate volume to total tube wear scar volume relationships. 2.3.2 Linear Analysis Model Linear model analyses were performed in which elevations at which thea .c. tube
. e. is The in contact with the support plates were modeled as results are used as an aid in the interpretation of tube vibration responses, as discussed above and for calculations of fluideiastic stability ratios.
2.-l)
l i Fluidelastic calculations were performed with a fluidelastic parameter values for support conditions consistent with conditiens indicated by tube vibration data. The inlet pass velocity is calculated as the SRSS of the gap and skimming velocities measured in the 2/3 scale model. 2.4 Wear Coefficient Tests The wear coefficient, Kg, that relates work rate to wear volume was obtainec frcrn wear tests conducted in Westinghouse and AECL laboratories. In these tests the tube support plate configuration is simulated and tube trotions characteristic of tube motions measured in the field are imposed at plant temperature conditions. The tube motion and contact force between tne simulated tube and. Support plate are measured for eacn test 50 that, witn tne measured wear volume, a wear coefficient can be determineo. Work done on the model D2/03 steam generator program resulted in tw; wear coefficients.{
~
- , c. , e_
The counterflow preheat steam generator tube-support plate hole configuration is nominally the same as the model 02/D3 configuration. Test results inoicate similar wear coefficients for the stainless steel-Inconel pair as for the carbon steel-Inconel pair. For these reasons, the wear coefficients developeo on the D2/03 program are applicable to model 04, D5 and E wear assessments. Tests have also been conducted at the AECL f acility for clearan.ces and motios that simulate expanded tube ' parameters. These data for( minal cap
}no,c.,e_
indicates wear coefficients similar to those obtained with larger caps {, a
2.5 Ga Wear Prediction Method Values of the Ga parameter have been correlated with wear measurements on KRSK0 pulled tubes. The Ga parameter
- o. , c. , e
~ ~
a , c. , e_ Wear coefficients, Kg, have been determined to related Ga values and total tube wear using the equation V = KgGaT A good correspondence has been found between wear predictions based on conservative model nonlinear analysis model and Ga wear predictions based on plate search Ga's. 2.6 Application of Ga Method Wear assessment with the Ga method utilizes direct application of maximum Ga values measured in the 16' model. The 16' model_ Ga values used in the assessment of the design modification are maximum values at each instrumented location obtained from plate position searches. Three tube configurations have been evaluated. The tube response resulting in the highest Ga values for these configuration are: 4 Unexpandedwindowtube-[ '**
- e ' .,c.e.
8 Expandedwindowtube-( 0 Non-windowtube-{ Tne Ga vs. turbulence force correlation shown on Figure 2.6-1 was developed using 2/3 scale modal rms turbulence forces, maximum Ga values from 80, 90 and 100% flow 16' model data and 100/0, 80/20 and 70/30 maximum Ga values
,e., e_
from KRSK0 plant data.- For a Ga value of[ the correlation
,e,s,c.e_
indicates a turbulence force - a at full scale. I 2.7 Selection of Tubes to be Expanded I ! The data correlations available for selection of tubes to be expanded are
- listed on Figure 2.7-1. 2.-11 l
The general boundaries of the expanded zone have been established based on G: distributions. G4 values from 16' model data are used for columns modeled as full length tubes in the 16* model (columns 43 - 57). The Go force correlation together with the turbulence force distribution from the 2/3 scale model were used for tube locations beyond the 16' model boundaries. Nonlinear model results were used to support selection of key tubes at the expansion zone boundary. 2.8 Generic Estimate of Tubes to be Expanded The assessment method described in Section 2.7 was the basis for a generic estimate of the number and location of tubes to be expanded in model D4, DS and E. The present results of the assessment
,. %c.,e .
indicates expansion of
- % c., e.
1. tubes on models D4 and D5'and tubes on Model E steam generators.
- ., L The boundary of the expanded zones for all models can be described as
- o. 3c. , R
~'
2.9 Wear Assessment of the Design Modification An extensive data base is available for assessment of tube wear with the design, modifi cation. These data generally indicate: _..,c,,,_ 1-14
1 Table 2.6-1 - WEAR-ASSESSMENT OF KEY UNEXPANDED TUSES FOR EXPANSION BOUNDARY ASSESSMENT , 4 h TUBE < TIME TO 40% WEAR DEPTH Upper Bound Nominal Model Noniinear~Model 4 R49C35 s
- R48C41 .
R44C56
\ /,
GA Method
- R49C35 l-R48C41 lt R44C56 f
R49C84 , , R21C58 kg g mode tube response ( Gamethodcorrelatedfor( y i i \( , g... 1 -16 e z. !
., , , _ - , , - - , -- . , -. .-=,---, -,... ,.- , _ ~ _ - --- __m. , _ _ . -
8 v ,
?
- o
(
)
\
L s, ' . e ') i
./
~
, en.,b,c,L 4
s i I L t i i e t, , , . f
'. j_
\
w J s ame Figure 2.6-1
=
-) s
^
0) c' d~ l 1
).
s r tJ i
, r k
~ N C2 y
4 I
-( l.L i .,
e i 1
') '}
l 8
e 0 v [ l- g n i J.
\h M
3 3 1.T. . I 1
FIGURE 2.7-1
-DATA BASE FOR SELECTION OF TUBES FOR EXPANSION The vibration data base used for selection of the expanded tubes included:
- a. KRSK0 data - G1 values for 16 tube locations including tubes in both steam generators.
- b. 16 Model data - GL values for about 50 tube locations in the 16" sector.
(Figure 2.6-2) .
- c. 2/3 Model data - Turbulent tube excitation forces for about 40 tube locations. (Figure 2.6-3)
- d. Nonlinear model tube vibration analyses - These analyses utilized the 2/3 model force data to calculate tube wear at key tube locations and aided the identification of tubes requi. ing expansion. (Table 2.6-1)
- e. GA vs. turbulent force correiations - Measured Gt values have been correlated with turbulent force data to obtain Go estimates at tube-locations not included in the 16 model.
- f. Ga wear correlation - Ga values have been related to wear based on measured GL values'for tubes pulled from KRSK0 to obtain direct wear measurements.
g-i9
2.9 Conclusions-
- 1. Cross flow velocities in the preheaters of unmodified model D4, DS, E counterflow steam generators have the potential for causing excessive tube wear. rates at plant full power with all feedwater entering the main nozzle.
- 2. Tube vibration is sensitive to the extent-and pattern of contact between a tube and the support plate. Adverse support conditions can result in tube vibration activity that is{
} o.,c than,ethe activity at the same location with a beneficial tube support condition.
- 3. Adverse support conditions are evidenced in tube vibration data by response frequencies that are consistent with lack of support atasupportplate.(,
, o. , c., e a , e.,e 4.
- 5. Expansion of a limited number of tubes and incorporation of bypass flowrates of 10%, (4 loop models 04 and DS) and 18!
-(2 and 3 loop model 04) generators reduces tube wear potential to levels consistent with other modified preheater steam generators operating at 100% power.
- 6. Tube expansion for Model E generators without increased bypass flow results in reduced wear potential comparable to the D4 modification of tube expansion and bypass flow.
2.-2 O
7, , _
.s.e'
~
- 7. Tube wear: predictions generated with the -non-linear ' analysis
, ' model, .using very conservativeltube-. support and wear rate
~
^
distribution! assumptions, demonstrates that the. minimum full
~
- .~powe'r. operating period. to incur a wear. scar depth of 65%.of the
^ "'
-tube wal1-is".-- .L .
f 4 9 e p. {< . .-[
^
Y
~
i '. , h h I' l -r a b L- z 2i e
3.0 ~ DATA COLLECTION 3.1 Operating Diant Data Collection 3.1.1 Description of Plant Tube Vibration Measurements Plant data were acquired at the Krsko Nuclear Power Plant with three sets of accelerometers, identified as Test Phases 'I, II, and III. During Phase I testing, data from an initial set of accelerometers were recorded for tests performed from January 1982 to May 1982. Steam generator number 2 was instrumented with accelerometers mounted on the steam generator shell and installed within four. tubes (R49C56, R48C55, R46C56, and R45C56) . Phase II' testing was performed with a second set of tube accelerometers installed within four additional tubes of steam generator number 2 (R49C35, R49C59, R49C62, and R21C57). The purposes of this instrumentation were to obtain data for plant operation 'at split main / auxiliary nozzle flow rates, variations in f eedwater temperature and pressure, vibration of tubes away from the corner of the T-slot, and to assess the effect of three months operation on tube vibratory behavior. Phase Il testing was performed during June and July '1982 with additional testing performed in October 1982. During Phase III testing, steam generators 1 and 2 were instrumented with accelerometers on the shell and within steam generator tubes. Phase III testing was performed during November and December 1982. The purpose of Phase III instrumentation was to expand the data base to include both steam generators- and additional tube locations. Data from all three test phases were used to characterize the tube flow induced vibration issue. Along with other analytical and experimental data, the data were also used in the evaluation of stesn generator modifications. The Krsko operating conditions for which data were acquired are tabulated in Table 3.1-1.
' 0855c/0119c /060383:5 14 .
E-\
~
, o 3.1.2 Accelerometer Configurations and Locations The bi txial accelerometers used to measure tube vibration were fabricated into an assembly of the type shown in F,igure 3.1-1. The assembly includes a spring assembly to provide contact with the tube inner _ wall, hard-line tubing for instrument leads, and a forward attachment used to pull the assembly into a tube and determine the installed accelerometer orientation. The spring assembly that supports an accelerometer applies a force to the tube wall-
' sufficient to increase' the natural frequency of the mounted accelerometer e 7 a,u. G assemblyabovetheprimaryrangeofinterest,( 3 The hard-line tubing is also supported by springs to eliminate impacting that would induce spurious signals into the recorded data. Plan and elevation views of tube accelerometer locations are shown in Figure 3.1-2 through 3.1-6 for the three test phases. Locations of external accelerometers mounted on the steam generator shell are shown in Figure 3.1-7.
3.1.3 Data Acquisition and Reduction Upon completion of accelerometer installation and data acquisition system set-up, a pre-test check-out was performed to verify the operability of the system. Upon determining that required plant operating conditions were achieved acceleration data were recorded on magnetic tape for post-test reduction. In addition, some data were reduced on-line to verify its acceptability, and to provide preliminary information. Post-test data reduction provided the following typical reduced data forms for use as noted: s f f-l '. ,
)A b bRMS acceleration spectra - used to identify frequency con of acceleration.
., c, o 2.
(displacement
}RMS displacement spectra - used to provide tube RMS values.
- 3. Time history visicorder traces - used to provide peak-to-peak acceleration values and tube-to-baffle impact information (initial r
impacting, number of impacts per unit time, level of impacting). 0855c/0119c/060383:5 15 Z-3
v l' . - o-4
.4 Phase and coherence plots of upper and lower accelerometer pairs, tube-to-tube accelerometer pairs, or steam generator shell-to-tube accelerometer pairs - used to characteri::e tube and shell motion, and to identify common signal' sources.
- 5. X-Y displacement time history plots (Lissajous figures) - used to study and identify tube motion patterns at different flow conditions.
In addition to' the spectra generated using the Nicolet Fast Fourier Analyzer, the analyzer has the capability to provide RMS acceleration and displacement values by integrating over a specified frequency range.
.The data forms . described here were used along with analytical and experimental
- model data to identify tube vibratory character _istics such as frequency, amplitude, tube-to-baffle support conoitions, and to provide Ga values wnicn have been correlated to tube wear.
3.1.4 _ Data Parameters Used to Evaluate Tube vibration
- Four major parameters were used to evaluate tube vibration. The first parameter is the tube response frequency or frequencies. {
p
'The second parameter used to evaluate tube vibration is a peak-to-peak A
acceleration value measured with the tube mounted accelerometers. peak-to-peak acceleration value was obtained by visual observation of a time history record. In the determination of a peak-to-peak acceleration value, a typical ' impact transient was selected to weight the number of impacts
^ occurring at various magnitudes.
The third parameter used to evaluate tube vibration is the root mean square displacement (RMS 6) . The RMS displacement was computed from measured tube
- acceleration signals. Using a Nicolet 660 Real Tim,e Analyzer, an RMS 0855c/0119c/060383:5 16 33
,c.,4.-
- displacement- spectrum was ' generated and integrated over the(,
frequency range. -- a , c, i. 3.1.5 Discussion of Tube Vibration Data i l As discussed in Section 3.1.4, various f actors may affect tube vibration re activitge. a.s. The highest tube activity occurs for those
-a, tubes %, ,g,pond responses. The
! The majority of tubes 1in Phases I and 111 had lowest tube response frequencies from Krsko plant data are tabulated in Table i 3.1-2, along with other frequencies of response, h t , accompany t the lowest tube response indicates the _ ,
,,esponse r frequency. The occurrence of a , 3 0855c/0119c/060383:5 17
(- 3-4
_*,ic.L s The Ga values at the inlet ' pass and E levels are compared in Tables 3.1-3 and 13.1-4 for 70/0, 80/0, 90/0,100/0, and 70/30 main / auxiliary e- nozzle feedwater r . ,=,b.c., determined for the E flow rates. The highest' GA .value of t }was
.0855c/0119c/060383:5 18 B-6
elevation'of tube R49C56 in steam generator No. 2. The acceleration and
' displacement values associated with this Ga value are( -
- 4, b,c, t- - .r TheGavaluesatthe{_
a c, t. s
.. a,c,e 3.1.6 Summary / Conc'lusion After Phase Ill Based on evaluations of Krsko Phase I, II, and Ill data, the following conclusions are made:
Tube Vibration Activity _$ 1.
-~ _]
- 2. Vibration of Tubes R49C56 and R49C59 (5.G. No.1 and S.G. No. 2)
The vibratory responses of tube R49C56 measured in steam
- generators number 1 and number 2 were essentially the same in frequency content at approximately equal main feedwater flow rates. The Ga values of R49C56 at the E level were of comparable magnitudes in the two stesn generators at 70/0, 80/0, 90/0, 100/0 and 70/30 main / auxiliary nozzle feedwater flow (Table 3.1 4).
0855c/0119c /060383:5 19 3> -Gs
^
3 ..- .
?
-' _%C D I
p
- 3. Effect of Steam Generator Overall Motion on Tube Vibration (Steam ~
Generator Number 1)
-4 Vibration of Window and Non-Window Tubes R45C56, R44C56, R21C57 _
o ,q t 1 0855c/0119c/060383:5 20. 3 -7
. .c S,c,
- 5. : Tube vibration on Opposite Sides of T-Slot
. a , c ,( l l
- 6. Results of Tube Expansion (R48C55) i In November 1982, tube R48C55 was expanded at the B and D plates to evaluate the effect of this modification upon tube. =vibrations.
, e.,e.
Prior to expansion a diametral gap of approxi.mately , _ existed at both the B and D plate. After tube expansion, gaps offL
~
,- e. ,c .e.
were measured.
.L 0855c/0119c/060383:5 21 3-8
Ouring Phase I testing-(before expansion) at main feedwater fl,ow, rgtes of 60 percent and . higher, tube R48C55 accelerometers had a{ response. c >
!a,c,e_
.The inlet pass accelerometers had responses at L , 5,c,^ ,l while the E level accelerometers had } responses. At 100 percent main feedwater flow, all accelerometers displayed (.
s,c e.
'During Phase 111. testing (after expansion) at all feedwater flow rates, includ'ng split flow, both inlet p3ss acce,lerometers displayed only low level responses in the frequency _ range of _
a , b ,c.,( M ' Comparisons of expanded and unexpanded tube responses (Tables 3.1-8 to 3.1-10) shm reductions in both accelerations and displacements for the expanded tube. Peak-to-peak accelerations were reducea by a factor between[ ~
}a,c.,e. Comparisons of Ga values for expanded and unexpanded conditions show expanded tube Ga values lower by a f actor of -
m a ,c., e. ( At the inlet pass elevation for the 100/0 flow split, the Ga value af ter
- e. , % , c., e.
expansion was a significant reduction from the pre-expansion
,, 4.
e Ga value of( 4 similar reduction of Ga was noted for the 70/30 flw split. Based on the average variation of the Ga values for all accelerometers at the inlet pass level, the reduction in Ga values due to. tube expansion was equivalent to reducing the main f eedwater flow rate from 100 percent to 80 percent ( s,c, e._
'0855c/0119c/060383:5 22 g-1
3.1.2 Palled Tube Evaluations ~ Three tubes" were. removed from the cold leg (preheater side) of steam generator No. 2 at 'the Krsko Nuclear Power plant. Tube R49C56 was removed in May,1982 af ter approximately~ 1600. hours of operation at 75 percent power (70 percent main nozz!e . feed flow, 5 percent auxiliary nozzle feed flow). Wear on tnis tube was not ' detected by either -field or laboratory eddy. current ex ami nation s. Visual and profilometry ' examinations detected wear at five dif f erent baffle plate intersections. .The maximum wear. penetration depth was - 2.5 mils. In November,1982, af ter an additional 2850 hours of operation, tubes R46C56 and R49C35 were removed from the cold leg of steam generator No. 2. The field eddy current examinations and similar laboratory examinations' revealed no wear indications, but sensitive laboratory eddy current procedures did detect some
-wear. Visual examinations and profilometry measurements verifiea the presence of wear. Tube R46C56 had measurable wear depth at two baffle plate locations with a 2 mil maximum depth. Tube R49C35 had measurable wear depth at one.
' baffle plate location.with a 1 mil maximum depth. ,
Figure 3.1-9 shows the locations of the removed tubes in the tube bundle. Table 3.1-11 presents the wear depths and volumes for the wear scars at each baffle plate. Figure 3.1-10 illustrates the defined tube orientation. Figures 3.1-11 through 3.1-13 are photographs of wear scars for tube R49C56. The wear depth profiles for this tube are presented in Figures 3.1-14 through
. 3.1-16. ,
' The wear scars all have similar patterns. The tool marks from the baffle hole
- drilling were evident on the tubes. The wear surf aces have many round or oval dimples over the entire worn surf aces. These patterns indicate that , e the
, e_.
The wear wearing -mechanism was ~ surf ace on tube R49C35 had deposits over the wear scar which implie$ that the wear was not recent. Possibly the wear had occurred early in the plant life. 0855c/0119c/060383:5 23 g_fo
3.1.3 Baffle Gao/Preload Measurements Tube-to-baffle gao measurements were made by two different nethods. I'n one method a rotating mass was inserted into the tubes at baffle plate elevations
' B and D and the resulting tube motions were measured to determine the tube to baffle . gap and preload. This method was used in both Krsko steam generators and in the Comanche Peak Unit 1 steam generators 1 and 4 The second method for determining tube-to-baffle gap sizes was by actual tube hole diameter measurements made during steam generator manuf acturing. These measurements .were not matched to a specific tube hole or steam generator, but provide statistical data for determining probable gap sizes. The tube diameter. was assumed to be 0.750 inches for these gap determinations.
3.1.3.1 Rotating Mass Gap Measuring Probe The rotating mass gap measuring de'vice was developed by ANCO engineering to detect tube-to-baffle clearance. The vibration probe utilizes the simple concept of moving the tube and measuring the motion of the tube in the support plate hole. This concept was implemented with a probe inserted in the tube through the steam generator primary head. After the probe was positioned at a support plate junction, an eccentric mass within the probe was rotated, producing a rotational force on the tube. Two miniature accelerometers mounted in the probe at right angles to the tube axis measured the oribital motion induced by this force. A correlation between the rotational frequency of the mass, the measured acceleration, and the displacement was then used to calculate the diametral clearance. The motor to drive the eccentric mass was outside the steam generator and a flexible cable linked the mass to the motor. The probe has an eddy current
- device for locating the baffle plates. The probe was then clamped into place at the baffle plate elevations. The tube motion was detected by the two accelerometers. Double integration of the acceleration signals provided the tube displacement information. The preload forces were estimated based on the eccentricity of -the tube motion and the applied probe r
force.
- The measured 7 e- less for the and, gaps are considered accurate for preload forces of [
0855c/0119c/060383:5 24 3H y -y- ,- y -,r- -
,--,m ,- ,
. . o ,c..e.
and less for the Commanche Peak data, wnich is the
, Krsko data and preload range-for most of the tubes.
~
3.1.3.2 ANCO Probe Data for Krsko Steam Generators The Krsko tube-to-baffle gaos'were measured in November,1982, using the ANC0 probe. Thirty-one (31) tubes were tested in stesn generator No.1 and 61 tubes were tested in steam generator No. 2. The measurements were made at baffle plates B and D. Thetubetobafflediametralgapsrangedfrom( e., b,c ,e. 3.1.3.3 ANC0 Probe Data For Commanche Peak Steam Generators The tube-to-baffle gaps in stesn generators 1 and 4 at Comanche Peak Unit 1 were measured with the ANC0 probe at baffle plates B and D. Ninety nine (99) tubes were tested in steam generator No.1~ and 100 tubes in steam generator No. 4 . The mean gap diametral clearance was( , a.A c., e The maximum,
,l minimum, and average values only include the data which was determined to have
%C.,S.
a preload force of which includes 68 percent of the data. The accuracy of the data for higher preloads is questionable. 3.1.3.4 Tube-to-Baffle Gaps From As-Built Dimensions A random sample of tube hole sizes for as-built Model 0 steam generator
- baffles were available from the manuf acturing plant in Tampa. This sample included 240 holes which ranged in diameter from ,
6, b, c , t. 0855c/0119c/060383:5 25 7 S - t-
~
- i. .. .;.t.-
, s 73 - -
13:1.3.5 .y
.Sumary. of Data -
.m Theadiametral gap sizes,;as determined by! the ANCO: probe' and the
' manuf acturer's' measurements ' ranaed from[ -
,% c., %
.L -
i t P <s i x 4, 4 s .d855c/0119c/060383:5'26 ::' 13
--- ', , , , . . , , c4.--,-,,,s... . .m,, .. - , , , _ , . , , _ , , -- . .,,..,,-m,.,4., , . . , .......,..,,,,-.__,,,---.,,_...,,,,,.s . . - . , , ,
t 3. c Table 3.1-1 Operating Conditions for Krsko Data Collection Parameter Range from 40 to 100 percent in 5
~
- Thermal power percent increments
~
Main nozz1e feedwater flow from 40 to 100_ percent in 5 percent increments
- Auxiliary nozzle feedwater' flow from 0 to 30 percent in 5 percent i ncrements*
Feedwater temperature from 172 to 230*C (342 to 446*F) ' 0
*100 percent of _ feedwater flow corresponds to 4.09 x 10 lbm/hr.
I [' i: i l I i-i I Y 0855c/0119c/060383:5 27 3-/f-i. t - -
.- - .- - . - . . ~ . . _ . - _ _ . _ . _ _ _ _ _ . _ . ._ _ . _ . . _ . _ _ . _.
1 Tab 1e 3.1-2'
-Inferred Tube-Support Conditions Response'0bserved Inferred Support Conditions
- _%c,c .
s
-
- Based on linear model analysis i
6 l0855c/0119c/060383:5 28 :j - IG
a,b,c.e Tables 3.1-3 thru 3.1-10 contain Krsko Ga data which are considered Westinghouse Proprietary. 0020G/FTE/7-28-83 ' 3 t C,
Table-3.1 . Wear. Depths and Volumes for the Cold Legs of Tubes R49C56, R46C56;' and R49C34' from Krsko Steam Generator No. 2 Maximum Penetration Calculated Volume Angle to Maximum Mils (Inches) Penetration-Degrees
- Tube' Baffle- Pe rcent .
~
. 4 0855c/0119c/060383:5 37 3_;)
- ' a,b,c.e Figures 3.1-1 thru 3.1-8 contain details and -
results from the instrumentation program at Krsko which are considered. Westinghouse Proprietary. I r 1-I l l 0020G/FTE/7-28 3 -/5
1
,eteei- %
,,,,t13
"--""""?I x _
3f111 I g,,too., , t es j8 b - J?Inge l , ie_q es I i, . . . . .........rtrrr M 2_...N - T -.. . Lef t ns 3 Phase I Accelerometer Location F !Q t s1 y de 88 2 -\ e7 lE ,R49 C56 8-D, D-E b
's m eL es y <
R4B C55 B-D. E :3?.. '., "Co3
< -w. ep se_-e,g 1 Mase II R46 C56 B-D E 4
- Accelerometer R45 C56 8-D. E
-d 8s
- 52:8%'
=,
e4 e2 83 83 Location PULLED TUBES 4= ?l . ec 7e R49 C62 n..... ..., , ...... U .) ', N 777g B-D 1: '-w ' 74 l ; ** ' - l '. ,
-~
72 M ye 73 D
,aa
---- ges '
y - a E N ~j ieelf
. ,m,
<I..
,,' es es - es e R49 C59
. . .% + p i..z:
gg B-D E_ _- x- - ---- 7 g-
-se 57 E
' -55 **
- ce ! % _aeoesessenesessee . 6' l ?--
~ %" -~C
' . . k 52 53 t
rw - ee ,j:' - -== =:-5 ~ a u7
- - .k R21 57
< ised
- 42 4, q -- 4c ~gg
~ ~
-33 37
. --: -- 6 O S 36
.- - - A. & ' 34_.35 33 y . --*- -- [ .,,,,', '
m g3 %e
, , . L l' i ^
_I, L,,,
; . C w.ea - ---- :::- 2.=:',". ,'>
2s 22 R49 C35 _ =.=- t ee ps,. ; ', ' B-D 3es3' : *- 2 p ,=7,
': '-.1 EE . Lri = 9
," a 22 2
0 , n-- f:::=-+ w_a_. L n--g$: w e{1 tas.-g "
-2eTh::
1 e -- PULLED TUBE N 1 g=-g -te 17 N , M ,=-m_p.g ';- . .g Mr.=". 4., ._ . 14 is 33 I 1 '
- m;; ;-.;.- d gey 12 Fl 1 , ,
c g +;: ma-- p- _"- , _y ie ,1
- ; :::: t= 73M 3
.. ; . ^yw ; - a _7
- : -.+ r W.--: - 4e _5 I ,eo.:e# .~
1 , i i o 1 e**s i. 1 . W. g. g
,,gese n
; i :
l in ;
*'f,
~
[
+
,li o
. i .l r
j N,m u ww w.ua!AAAA g E n a a m -* - - N mn N'8g'I"I Yf. ' 3 k-
.M .. .-.mwo-emse-wave-.AAs, -
--- nnnun.l..W..Affffi mwo 2
m FIGURE 3.1-9 L Pulled Tube Locations for KRSKO Unit 1 Steam Ge
Note: Convention shown in this figure is for the tube to be stationary and the viewer to move around the tube. During actual tube examination, the camera or profilometry equipment are held stationary and the tube is rotated I d clockwise as viewed from the tubesheet end looking upward. [P
./,
To adjacent g . Row 48 tube 2 Toward Feedwater No-dle 4 2JO
\
0 180 I % M l l l I l l FIGURE 3.1-10 Orientation for Nondestructive and Destn2ctive i < Examination of KRSKO Steam Generator Tubes. 1
1
- 1 t a
~
g 1
~
, J .
~
. j f i
; I -
t 1 l p f f - I b l l
\
l l 2 ; ; 1 - l , . - l O' 90* t a 0 4 a . I ! L . 180* 270* l FIGURE 3.1 1]; KRSKO tub'e R49C56, cold ic;;, from steam generator 2, at intersection with baffle B. i m_ :
I '
, s
. , i <>
\
's s
I
.,
- 7, ' ' '. ... ..
S ., Jk
~
r ,..7.'* . ,. : ..g .; -
, s . .- :
. , ; .m
. . , . , 4, 7.. ,.
.e p
j q
~
. .,p; [
.-, , . ,:- 7
' *,, . 'l..
A..
\
l c .W n4
- ~ _ c -.' :~.' .> y k:c-.
.. .v ., . ..
+ . ' .s.: :.: -.: .
m p. .~ g. , '
..?
Sk
'W.'c~ b ;J.,"'.:,.
A. -h 'D[' '.*. .,> '.k. .' ,-' ' :-: t ? i
- f. :.. ^ M;%.
l
.y ;;%; ', :o ; : 4 '.-
~.k + :;.- > _
. 6 ;. p .' : .' *
(. q'
.?
r.? - u = .. . . .
.. .:. [
. %. 8, .
s {? e I
. c . .. ..
. = . . . !
l O' 90* i i ~ b ;Nf k h . f 5 9g4 -:f ex( e.: - . 5.,,*.[*- f-
. $.f
.k '"
.Q &
- l, YI ,
l 4
- l. ?. If f. ,
pu . I '
' h-[ -{. '})
Y h.M
,e.
$gm?N .. - A. . f i.@9 g;; s j.h\ .i 4.'.. gg.kqM . . .. .s.,w 1
; . /w f.
? 'l.?. .(y' ? Y:f.. ~A
- h. . $h .u . .9.( .N.
> s. nl..:. . ..
> :2 - %
f .h . '.n. . - .-{I-
- k. 3.
u s .2 a
- {.,' ~g;..
e? Y" ' . S % a1, Y y: _. ' ; ,. . ..; x; -
.,e , y. ,
>:;{
. s. . V , _
q, >
.m 3 1;.
.gg e. ,
1,. -I. > 1l nnoi . E q: ' :k:5 : . . y i 'i .? r .. g '*.
.. l'..: : .' ,' . :q l .V, t ': -4,r . y.. g' y .:, y .
4-
.".? * *'t.' .h 4-
'.E
, ~
v.4. %,i f ,1 Js-y9 < !., .,,,i. I : .' p4%?j .
.h 7i c.7 . C.g: , --f 4 .-
,N' r..
.+-.4 .,
#:) r "
;,.. >A 'q+.
w< .
.r
.x u ... . .? w . :-
~.~.,,. % J . 9,,. :..g . . u .a .
y&%%hn ;.l,. ,3. ,y h
- p. ..
E
-3 E :
$[if?.* Qf p K g',;' ;
O & & & '?.?l.,j~-[ p.((f:(p. N e.ij,,k. E .'= !Y $ , (% 3k6.M .. h,,j>c :Q{vm L .~9
. - , n 1e
.- .e . ys-
.c.,... ,,
,y
# .,.% ,s.-
3 3 4 . +, . g
' :$ $$'i L & .\%. : % . k
- J.cjQ"? 'y
'{ JO I l 180* 270* .
l
\
FIGURE 3.1-12 KRSKO tube R49C56, cold leg, from steam seaetator 2, ; at intersection with baffle D i RM-97352 i 6 t l
--__.--_-_-.}
;; '- .[, ; - ; p ;;.g_.:
.. -[ m.giW:7 ' g.}.;.- ( )_ .,,
f 3. y ;
. ..: 7. : : ; ,,; .. .;.c, < f;4 s.
v . . .;. - e 1. ?
.p,
- ~
- e. - .
f' i . . . 9.4 :%w. ..: - _..,:,?.}., . . p' - .g'}, ', *. '.. .v'p_. I
,} ,l
,v 5::: .1
.ff I
. '& (y6
- c
..t, . ; e'. ~ n ;..a
~
- y. :- ?
l, ':[ ~_ ; _ T U T ';D,.}w g:.,et' ' ';t ,f;l R ) :Q >l f , f .:. g. . },
. . . p- .. .. ,.- .a: . - . . - . . , . b .. : .;- , ~ . -
..u,... .. -: - ;.
. k. , . . .. j,,l&..$*:, *:'. A -. _. _-f:h::.G T
, f. ':Q I .f. . . NA. '
.- . . . g.
3 .
- . : < G d._
Tf;.+ . _ *f ~ *. *: . n ' j Ry.; * {r .N
,,'.. .:' .';. h ... ~.. .~..
l p;'; .,,; . i ':.. :,
,n ? 4. ,. 4-. :.. O-. g %.*..x.
.~. :: .
m.
. ..;'N A h * . 's l., ',:, . ': - - ,. l -h;;
....s.,. e 4;: .; /.; . -J: ._-,.. ,;'.', .*'.;'.;:-gi!
a o..: . ;.- . u.- . o
....y' ,
' l 4 9 f.v'h. e- - ...- - *.. -.-_- f ?.,
1 . .. ~. j-r , e
,. c -
g.u.
.c
' .- r-
,r . . s : .
L ' Y , ; ;: . M "- -- , l-f. v. , :. - q's.f., . c,3;,hy +; . . *.: . . ;,;; 3. ; ; ;, . ."_, .,Q _. I ' g. g. . ' e l Y b i.. f* : _ ff _. _[_* l:
~
} -
Y.'.
~
09" _h. .. . :,. ,w: .. .- 4
; 2-
.,R . ApJ.1+.
- : c u .p.;p
,L~."-
~
r..
,-..~.;.,.,._... 'N ; * . a: .
. . . . . . ".; 7; j . .- .
t,
.. . '. : . /. , ,.. : , ,s.3t.
. ,. _e. e,.. .,,, -
.w ws y
.a
. . f ., .,
g . : . O' 90*
*e
'V z, ' ^ ' :,: ) -:V.. s . .?;, , . u NY. ; -
- j. .[,' ;.
'~~.
[ s .
- i. -
&.. .;. N.} .e4p^.>.
.%. ,R,. .- i. yv-.. _
,~. v .
- c. . . . '
. %.}
1 s a . . Y gl.'p .- ; .,. Q:* (.1 -
'< . > . . . . &.:. :.. . ,v.$
u . ..-,;
- 7. . f;m..
k, . , , ,; - . . . . , .. :. . g -
- 84. ', J #v'i' >Q ,:$f;. a. q.*\
' 9.4
}}- 4T 6 %?g::g .. , ;M v . * ' x :'.
N
'c.-- _ jf 1 u ' f.*h k . ;p.nk %'
.;s f.
y .4 . r, ; 2: s ? : i et ; .; . ;Q.Q g
,3" i V :e.:.
3. h - y n,h..g g ,;f.5
, 4 g :.' ..r ,g...
L
}yk#Rr ,Qg ?, yM;.5l* . .. Q: ,. u ',:o .,,,
- y. 4 )% . ...m
. ,.. .,L .m+ ,..r ..m.
,$$Qi$$$hg.. .
k..
./
c.
.Er f(&..
sq N r, p 3- ' !Q p i
> ~:
..g 7 . % i:g.*
.n . .
., .-M; ' "~
,s^..,.
D, .1 : .y u vs... gy .f. . f,!.< g .-_39,,'.e.. . r . ,, n.
. A.
.. .?
u .. . q
. - * . , gj *{
._ .,: ; .M
,. , l Y.. ,, , $ . , . ) .Q
.x ,,
fA b y i'. R_ [ f[K [Q L%x.hh. b) k h.M;-Q'.; ;MDM) 180* 13p FIGURE 3.1-15 KRSK0 tube R49C56, cold leg, from steam generator 2, at intersection with baffle C. w- m s -
*=v W -a m Primary -
_ _ - ~- 0.25-incn Flow O.004-inch i I i - I i i ! . j i
... : .g . .l . _ . . . .
.. :I. . ; 1
.l.. .
l.
. l i .
i _ _ _
~ '
22.5' l
..j...... v, .j. . . ,. . . .
j.. . . . . .
, e l l l 7 i .
i i *
. l 1 l .
1 l
.s.. . .. . ..
l t, . .,. 3 .. y I i ! ' l u ! i e ! l t - 1- 45' 3 , y_ , .a.. .p.....,.. ...u_. _. s . . . q. .. .j _. ... { _ ....~..,..i.......
'
- l
. I I
- i i i . .
~
I . 1 1 l' 1. l I .I i i i
- l l l 1 I
.t I i.
'. .- i /
i 8
- .. L ..L. ! 67.5' i '
.!...t,.,...
/ ., j r ;/ ! !. - I.
l.. _
- j. V - 6 ;
;~f l
']i.. . ,. . . . .j ., . . . , . . ." ]
l-. - ' ~ f" ' :'- ~~
'--1.-
'~ ~l' q .r-
. t q .
. l :, .
I i e-
.. . I !
i
- .I ,
- i. l i
J I- i i i j . l j
'- i i ! l ~ . 9 0'
_' l- l -l i l
.... . . . 0 v44.- i i ; ,
- ! .- .l.- ,
. ,. _a l
j l
-i l
W . .. [ 1
.. ..i . .
. .* . . . . . . .a. . s. .
. i ; . : .
I i . j
. J .
s i1 , I . I e .
' t
' ? ,
l g.' .!
. i '* k '
I - I h *" ' !*
' .g. l
= '
112.5' Ah
. s
=
- i g e
l
- 8 l
l 1 . . s; -
-? -
7 I I I I l l l l 1 i i
. l .
. i i
. i
..+-. . .I ... . . . _
. 't * "t-
' ~~ ~~1'~'
-_ i.
.l, 135' 8
' j .
- ' ' I . -
w! '
! .. $ .. .'.... ..... 8 ..g ..l 4 g*g-< =
- d, " ='.
.i I
1 l . ! l ? i
; ,l Cold leg of tube R49C56 from steam generator 2,
~
FIGURE 3.1-14 l'RSKO unit 1. Axial profiles of the tube at intersection with baffle B. .
l _e.,, .h
- .li = u.. e
- .d =.:.::.
Primary jn'iiiNi!Ij{
._!l 0.25-inch ,
F1ow
.1 .
.w.__._._.. ,{
- ---. ~.- - .
".*m. ......
0.004-inch
. q. ,...,y_ ._. . . .
. , _I~ . -
1
, , . i :
i i ) l
. .. . p,-- J .
p 225'
.,i: ; .-.. .
,r i :
i l.. , 1 ll , i , i I .I ! ! ri ; , , ; . 2. _.q.._. j. . . . ;. . . . . . _ _
. , . ...:j_. _.
l I i 8 i e _p -
._ p . , . ; * . ,;,.. 4 4 .j
; . j.... ....
.. ; .l. .
247.,,.
._ . .H -
l l l l l y l l i , . .
- - [.m ,--,.i l l- . .F i l i i l-.I r , l i ' .
' .i i
;i 270, 4 _. _1 -q ..\l.W' . t. .. .
I. i l
! ,I i
- l i i ! ! l
.i 7._.
.jp.it:. ' 1..1.i 8l .. !..l.
.. . .. I. --. .ll .' _ - l L l i-i .. ,._. i .
i . . ..i ... i i I ... a ... ___. l _. . _ ..l.
.....;._. .i H . . ! _. ... [ ...
. i.
.. _ ... - 8 ._ .J, . ._l - i- t
' _ -l.
. t- -
; i .- j . ii, 292.5'
- r. '-
_.I yl 7 l i i I i ' i I . I ! ! ,
- I ,
'tijl
,l . l . ! i. l . i I i . . . .t . ;. ' _ ..! . . .
i,I 1 l t i l ~
- I L: -'~ 8 31 5' 7, .
1
....g .
i l . .. _, . . l I
- 1 l l
- l I g 8 I -
! l
- , i l .
- i
;i ~.dl..
1 l : l ' i l i i l e FIGJRE 3.1-15 Cold leg of tube Rt.9C56 from steam generator 2, KRSKO unit 1. Axial profiles of the tube at intersection with baffle D. Continued in Figure 3.1-17b i
._.____._.___..___-.____._..__m-__ -_
. , , ,. n . . . . . . , ,
1 E
.4..F. h=: .i. f. ={. . *i:E...
Primary i:F..~d!.5Ni~.f 0.2,.-inch
]!llgggg_
F1ow _: 6MEili~-p.E[ a
- 0.004-inch
... . J, __j..
l.
.n_ ... l ._
, [.
.. .. l .. .. . . . _ _ : . __. i
! - 337.5,
?
- m. l :
i ( - .I.y .._.... _.y...l' i l.
;i.
4.. . firi- . l 1 i ! 1 l l !. i:
+
g l . 1 l .
. . _. ,... ( .
. . . _. . . . ... i.... . . . .
m, .- . 0
.. F. .
i
.l .
[ i
..!._ ._. .. .. .. .. .. .. l . . :_ . . .l ._
. Yl - . ' , . _l l . . I i l
!.l .
l- l. ! .l.
-l l .l::.- l:. _.l-.. l..l::
. . 1 p ,1_.. l
'_l
- - w i i
!}
'. -l-l
- l. u 22.5'
. ;. - 4 .. - 6 1 ,.
l ! . i ; : . v l-
. l
. i l
...l:.
. 1. l: ! l . :l .
.1- .
l .l.
..l.
. . .l.
. . . l- -
.l...-l-:-
I: -.j . ...l- .. j. g . .a j
' :. ! l- !
. , i l 45'
. _. ! . . ..l f . .!. . . .j _ _ J._. L .: {
M.... ..;._._.....
,g .g
._a..
1
.1. '. 1 . f . f
. l
! l 1:
l ' .l l i i ! e l i . - - - - . . - . . , - -
...J -. .~ .
.. . .l.
1 i !
; .- . t . l l- .
a
. 1. .
. i. l .
. ; g .
1 I
..l. .l 1 .
67.5'
-l t
______:..l....l ; : l l I I l 4 l , l l l l l ; a
- . . ; i i. i .
i.
.l. I l l-l i i [ I- l- I. t ..
FIGURE 3.1-15b cold leg of tube R49C56 from steam generator 2, K SKO unit 1. Axial profiles of the tube at Continued f rom Figure 3.1-17a intersection with b ffle D. i l l l l l l
~
_A f M..iN[l'~ill3}i[.*.lMih
-I.Il..ll,dilil...siii u -
Ji .. r - 0.25-inch Primary c:.n c,:_..ccr:lh. .6_ _ {- rIow
- E p i.m:-cr.:n=1
- == ~us;=: . . . .
0.004. inch
, ! I !
i i ~ . . . . _.,_ _ . . ._.
, o . - . ; i
....t_.. j _ . .I 1 L 4 j o
._ . :. . . . . _ g ._
I
- l 'l. ..I i i I
[ . 22.5 I I 1
~. . !. i.I. . . r r .
r.
.) .. ... .,.. .j. . .
.;.r .
t ,...., . i .
- . I . _ . , _
t , . 1
._..,_n...
,.. .l._ _ .._ . _ ,.. . ! _. ,.l, s..
!. l_. .
0
- ! , l 45
! l l ! !. !. ..
l- !' I l ! b ,. I J p , . i i .I i ; i .
- l. f 4
, l l i l
,I
; a i. .i i j !.
nl1if.IJil.l'eI l . #- i
.! _ _q . .
i t
....p
.__;_. . . 1 . . , _ . o J t 6,e ' 5 i 1
_. 4. . 1. . . .J,.. _.!, J ~
.. . 1: l i I i
- 1. l -
.; ._. . . . . i_T_. w.
i I i i 1
~
1 i ld, _ .2 .-
. I
.. _. _..f :
I'. j . g. . ).
, , i. i ..l. ..
t.j.
; . j .. .1 :l 1 . . .
=]>. l.
,.. . i. r .. . .
l; ,. g 4_ ~+l.. _ ; .. ,. _.p.._ 4. . .. ' __ i..u..jJJ..:.... . . i. . . . . .. 90 U w.,
. . '. I: I .I
-- :- -g ,
j 4? i ! l I.l (. t... e l.I . .. ii- ., l
, i ,. i. .;.
- 7. : .
..- (r :
e . .j :- i ,
.1.: ..l.
i I.
. .l . . . _
, .. ! l j i I
.l .. , .
'l .. .
3 l s ... ., -~ 8 ! ! l .. . l . . . . ,, .. . , _ . . }. g . . ..' 112.5 j j f E- I - G.!. .+__ ,
.1 I.- .
s 8 A l L. ,'s _.._ :> .;.. f 9,
..t. 1
. .I . .l il I l. l. l l [ -l I g: ;
t
,. .I-i.
i. t g l l j.
. I j ; ,
O F t _1. _.l :. # ,
. l.5 a
i _ a f y-I
#~ f
! l ! -l 'l i j I. l .I ,
- i. l l .e.. l l l-
. l.
1 l f
'l- l - . _ ; .. .
l _ .,.= .
.:.l*a !r t.J.vi
' i ; 0
. 1._.. p ;.o. ' ni -c* r._. .. .r . __ _1- 1 , '
, ^
157.5
.T
_' ;. L
%. ..'.; . i. ..,. l . .7. . . .
I l ,. I.
.l..
.l.; _ .!- /
. i . :~ . . l. . .
I. i i l i l l.-
. I i : l - i1 ,
l :
.- l
- }
n t cold leg of tube AxialR49C56 profiles offrom the cube steam at generator 2, incer-FIGURE 3.1-16 KRSKO unic 1. seccion with baffle G.
3.2 Test Model Descriptions Test models used in the counterflow preheat steam generator program included a 16' full scale water model, a 2/3 scale water model, and a 0.95 scale air model. The 16* full scale water model was used to provide tube vibratory response data in relation to inlet flow conditions and varying tube support plate conditions. The 2/3 scale water model provided measurements of feedwater flow velocities at selected locations within the tube bundle and measurements of the hydraulic forces acting on tubes. Determination of the steady ' state velocity distributions of the fluid in important regions of the preheater was also made in the 0.95 scale air model. 3.2.1 16' Full Scale Water Mode' 3.2.1.1 Description of the Model , The 16' water model included half of the preheater inlet and half of the tube bundle, bounded by a plane along the center of the T slot, the center partition plate at the center .of the steam generator, and the wrapper. A diagram of the-model is shown -in Figure 3.2-1. A rectangular section of the tube bundle,16 tubes wide from the T slot and extending the full depth of the bundle, was composed of long tubes extending from the tubesheet up to the L support p1 ate. The remainder of the tube bundle was composed of shorter tubes extending from the tubesheet up to the D support plate. The model inlet included half of the reverse flow limiter and half of the water box, from the centerline to the side corner. . All dimensions in this nodal corresponded to the dimensions of the prototypic steam generator. The positions of the B, D, E, G, H, and K support plates were adjustable in the plane of the plate. Feedwater entered the model at location I and exited at locations II, III and IV (see Figure 3.2-1). Outlet Il removed flow which passes through the B plate. Outlet III removed flow which passes through the D plate above the shorter tubes'in the bundle, and Outlet IV removed flow which passed through the length of the preheater along the long tubes up past the L plate. 0855c/0119c/060383:5 38 3
- The first 5 rows of long tubes nearest the feedwater inlet and tubes along the T-slot were prototypic Inconel 600 material, while -the remainder of the tubes -
were stainless steel of prototypic diameter, to simulate preheater flow resi stance. Tube support' plates were 405 stainless steel. The model enclosure was carbon steel with a protective coating inside to prevent oxidation. The water supply was'provided by a recirculating water loop with pumping capacity up to 5324 gpm at 100 psig. Flow through outlets 11 and 111 was controlled by valves in each return line. All flow tests were carried out at about.100'F. All flow connections to the model incorporate flexible sections to isolate the model from loop vibrations. Loop water was partially filterec during_ operation.
- 3.2.1.2 Instrumentation in addition :to normal test loop instrumentation, the following instrumentation was utilized: o A. Accelerometers Bi-axial accelerometers were installed in selected tubes to characterize the tube vibration response, to determine the repeatability of tube response at various tube support plate alignment conditions, to simulate Krsko tube vibration field data, and to evaluate design modifications. Six external
. accelerometers were installed on the outside of the structure to allow "
The dissociation of flow induced vibration from structure induced vibration. external accelerometers measured accelerations in the x and y planes at the top, bottom and midsp an of - the model. Tube support plates B and D were monitored for vibration by two biaxial
-accelerometers located in the T-slot region of each plate.
0855c/0119c/060383:5 39
- E-t:f L
1:. .. . . t[ g. i S. Dynuic . Pressure Dynamic pressure measurements were made in the test . loop. to. determine the _ loop-induced vibration, and in the model . water box to check pressure- fluctuations-in the-model inlet, i
=
t f I i
?
f _ 0855c/0119e/060383:5 40 - 3ao t m . - _.___
. _ _ _ _ __ __ _._-___,_._m._ _ _ _ _ _ . - _ _ . _ _ _ _ . _ _ _ _ _ _ - _ _ . _ .
c-3.2.2 2/3 Scale Water Model 3.2.2.1 Description of. the Model The 2/3 scale water model simulated the portion of the feedwater channel comprised of the reverse flow limiter, the feedwater entrance nozzle, the water box, and the first shell-side preheater pass (See Figure 3.2-2). The test article consists of the full complement of 4578 tubes having an 1/2 inch 0.0. (3/4 x 2/3 = 1/2 inch). A side view of the model is shown on Figure 3.2-3. The model provided a 2/3 scale representation of the following features of the feedwater inlet region of the 04 Steam Generator: A. 180' segment of the wrapper on the cold leg side B. lnlet nozzle with reverse flow limiter C. Water box with vertical flow ribs and impingement plate D.- First tube pass between the "B" and "D" baffle plates containing tubes arranged as in a D4 steam generator and incorporating the centerline "T" slot. The spacer pipes located between rows 35 and 36 and columns 45/46 and 69/70 were included. E. Provisions for incorporating a divider plate along the "T" slot and waterbox centerline to simulate two 90* sector models. F. Provisions for installing instrumentation for measuring fluid velocity and tube hydraulic forces. The water supply was provided by a recirculating water loop. Bypass flow across the B plate was controlled by a valve in the return line. All flow tests were performed at slightly above ambient temperature. 0855c/0119c/060383:5 41 3 - 27(
3.2.2.2 Instrumentation A. Velocity Measurements Dual pitot-static tar.gential velocity probes were used to measure the tube gap velocity (approximately mid-way between tubes) . Commorly called "Cownorn probes", these probes .ere raised or lowered, allowing velocity measurements at various elevations along the tube between the "B" plate and the "D" plate. The probe was rotated 360' in the horizontal plane so that the angle of approach for this velocity resultant could be determined. Stagnation pressure taps ("cowhorns") protruded from both the right and lef t side of the tube body. The velocity was derived from the pressure differential between these stagnation pressure taps and static pressure taps located on the tube surf ace above the "cowhorns". A variety of three-dimensional prism probes were used in making velocity measurements in areas of the vessel away from the tube bundle. The probe bodies, though they varied in length and diameter, were all of the same concept. Ports on each probe allowed for the determination of the direction of. flow. B. Hydraulic Load Measurement Hydraulic loads were measured with a force tube assembly which utilized the same oversized shell penetrations as the dual pitot probes. Tubes used for measuring load had special end connectors. These passed through a pair of 3-dimensional load transducers mounted above and belcw the structural shell. Each transducer was held between a set of stress distributor blocks which contained water-tight seals. Preloading the transducers by putting the tube in tension was accomplished by tightening the hex nuts at each tube end connector. This preload allowed the piezo-electric crystals in the Kistler transducer to respond properly to the lateral and longitudinal forces applied to the force tube - assembly. 0855c/0119c/060383:5 42 3-33
3.2.3 0.95 Scale Air Model : I 3.2.3.1 Description of the P.odel Three passes of.the steam generator preheater section were included in the model (See Figure 3.2-4) .- The flow was guided to one end of the assembly before entering the tube bank. A return flow through the first three passes of the preheater section created the necessary conditions for replicating the flow patterns throughout the distribution duct, pass 1, pass 2, and pass 3. The base was constructed of angle iron with a plywood overlay. The baffle ! sheets were supported by this base and spaced by shaf t collars on selected tubes. 'Through bolts and end angles stabilized the assembly. The flow distributor was f abricated from polycarbonate and acrylic sheets laminated to .the proper thickness and. supported by the baffle plates. This section also included the mounting flange for the venturi nozzle assembly. The edges of the baffle' sheets were reinforced by a steel angle formed to a quadrant of a circle. The wrapper, a clear polycarbonate sheet, resteo on the formed angle and was held in place by a series of tensioned steel belly bands. The tubes were removable to allow removal for insertion of instrumentation. A wooden mandrel of a single venturi nozzle was turned. Four individual nozzles were formed by successive layups of polyester and glass on the mandrel. The.four venturies were mounted on a plug. A sheetmetal sleeve was
~
fitted around the assembly and filled with polyurethane foam. The air supply was furnished by a 40' hp, 3600 rpm blower which delivers approximately 4200 cfm at a pressure of 42 in-H20. A fiberglass duct connected the blower to the model . A long radius (1.5 r/0) prototypic elbow connected the duct to the nozzle. f 0855c/0119c/060383:5 43 33S
3.2.3.2 Ins trumentation A. Velocity Measurements Flow velocity was measured using a cylindrical probe, consisting of a 10-foot lengtn of tubing identical in diameter and material to the other tubes in the model. Any of the tubes in the model could be removed and replaced by this probe to measure the air velocity distribution over its surf ace without
- disturbing the flow. Static pressure taps were circumferentially imbedded in the periphery of the tube normal to the surf ace. There are five such stations, each at 3.0-inch intervals along the tube. Each tap consisted of stainless steel hypodermic tubing with the tubes led out to one end of the probe where they were connected to a scanning valve through Tygon tubing. The valve was connected to a calibrated pressure transducer.
A Prandtl-type pitot tube was used to measure air velocity in the cap holes of the downcomer of the air model preheater. The pitot tube head was piaced paral'.el to the air stream to preclude errors in the total and static pressure readings. . Two strain gauge bridge-type transducers were used to measure air velocity in the model preheater. The first transducer was used in conjunction with the cylindrical probe and the scanning valve. An appropriate signal conditioner was used with each transducer as a balance unit. Calibration of the transducers and their balance unit was achieved using a regulated source of compressed air and a micromanometer. B. Temperature Measurements Three copper-constantan thermocouples were used to measure air temperature in the thermal sleeve of the preheater model, and ambient dry and wet bulb temperature. Calibration of the thermocouples was accomplished by check.ing the ice point and boiling point of distilled water. Outputs of the thermocouples were fed directly to a data logger system and were converted to temperature readings by means of an internally-compensated isothermal block. 0855c/0119c/060383:5 44 3-31
4
~ ,
'O .. . c-
\ ?r h,gyi ,' 't%
s r. 1 y% n: ,
\'t 4
,-- v)..
X)
= ; ~ gx93
\y
._) w .
h -- Ic-
,,-.49 --
7g t
/ )..
,. ,, ,., , ,.. a -
-p
["s [v:c:"n...
'~'
s AF q d- fy[-
' y,'.'Is- t -
~
,. . > ' y Y
n... y ,
%W (
; il , nai. i i o.
f.. $: Y N-
- 16' WDE COORDg g ts e
, a .mL # .
i FIGURE 3.2-1 nov28"# 4 10 S** *
\
S S e
-a+- + - - , - y - , - , -, -,v--------r-es--+-----ee,--er e e- ow -ww--,.--ye,-,- m,y, --e,,-e---,---a-++w-w+ wee- ,w- , w ce , +,vw ww w ww e e-- ow-e e-r w+ e n+--- vm *- w e we g-we= mwev--
l a CAP PLATE l WATER BOX ,/
/
CENTER OUTLET PIPE \ IMPINGEMENT RIS f TE OUTLET P1ANIFOLD PIPE , ' i
~ u SIDE RIB
'fj rf WEAR LINER
\~
STIFFENER BEAM { "'
~
e, ' COVER PLATE tr t MG, i v i-j L} 'i;l..;r;t
! {
,I.
l
' i ll 'i "B" PLATE
! M' .h
~
I [g'll.lllis LOWER VE5SEL T. ui LLi>
. Z 6
l Y- X
\ - - - - - - - , _ _ _ _ FIGURE 3.2-2----Isometric View of the 2/3 Scale Model
SKETCH SHEET e... res-- WESTINGHOUSE ELECTRIC CORPOR ATION Mock Impingenent Plate
& -a Center Rib 9/ 2 ?? '@
555
/ ~;O ,'
Inlet 1/2" Tangential Probe Di Nozzle
/ }_'? \
/
l y [ Stiffener / Beam :
/
1'
- - - ~
I l
,, %:u \
se i Tie Rod -
/, 55 -
s
\ . i' ! (
N Access
/
- 2@- Ports
\ - n .
1, l [ sN i 8 A x x xx x , . , _ p J-. a,c,e .
- i
- \
l 1
\
; < N ll 'l b C ,
] ,/ / / / / / //////// s. W r7-5 't "L i \ v II
; m,
\ '.
I utlet 1 Manifold - Tubes % I 15" Stiffener Beam Outlet Manifold FIGURE 3.2-3 SIDE VIEW 0FTHE 2/3 SCALE MODEL
a
- St'WCtu'.1 V9e tu r t he#21g [*LIG"'*T nssema r ,
[# -' .s .
#Ia= Oistettutoe ,
/ , L- /
' f.i...t.,
,<g h~ $ ~s ' , y/ o......
!! g
( \,
\ s O
i l 'x s l l l l FIGURE 3.2-4 0.95 SCALE MODEL 1- .
,,-...y.-, . - - - . , . . _ , , . - , - . _ - .
w- :-
, - 4.0 D ATA ANALYSIS d.l' Tube Vibration Measurements in the 16* Model 4.1.1 Test Parameters The 16' model was- designed to simulate the preheater section of a mocel 04 Site test data available from Krsko were used to establish
~
steam generator. the validity .of the 16* model responses. The test cases evaluated with the
. 16' model consisted of the variation of three primary parameters:
- 1. Inlet Flow Rate This' parameter was varied as a percentage of full feedwater Euler Number for a 2-loop Model 04 steam generator for baseline tests and as a percentage of modified 4-loop flow rates for "TBX" tests.
- 2. Plate Positions
' Tests were conducted at a total of 73 plate positions' to obtain relevant representation of Krsko responses. The plate positions were varied in later tests to maximize tube response and to relate tube response to plate position.
- 3. Design Modifications These modifications altered flow distributions or changed tube-to-support plate contact conditions. Tests were conducted to determine the effectiveness i of several conceptual design modifications and to evaluate tube vibration for
' the final design modification.
4.l.2 Test Data Reduction The accelerometer signal tape records were reduced using a Nicolet 6608 FFT Analyzer. . This instrument performs dual channel f ast fourier transf orm
-analysis of . data yielding a variety of frequency, amplitude, phase angle and
-0866c/0120c 060S83:5 1 bk
WESTINGHOUSE PROPRIETARY CLASS 2 real time-displays. Four types of processed data displays were used for data eval;ation: RMS acceleration (g's) vs. frequency; RMS displacement (double integration of acceleration) in mils vs frequency; coherence between pairs of data channels .vs frequency; and Lissajous plots of dual channel X vs Y real time displacement signals. The frequency displays represent the average of a number of samoles of data (typically 64). Averaging of the result from each data sample led to the identification of periodic or deterministic narrow band random responses from which statistical evaluations were applied. Data was also reduced in the form of time history chart recordings. Figure 4.1-1 describes typical output of a Nicolet 660 FFT spectrum analyzer. A description.of each form of data reduction follows:
- 1. RMS Acceleration Spectra The RMS spectra were obtained from a f ast fourier transform which is then represented as the true RMS amplitude for 400 discrete narrcu band f requencies comprising the entire evaluated bandwidth.
ar,o displacements were reduced in.a manner , o similar
, c. . o to the Both RMS displacements acceleration spectra. J were reduced.
Tube frequency responses between( }owere
,% R-used for displacement data. Due to the low acceleration levels at low frequencies, the actual signal was indistin uishable low frequency noise. Typically,
, u ,d c
displacements above( {werenotsignificant. For cases where line noise (59-61Hz) represented a significant part of displacement, it was
~
removed from the displacement by square-root-difference- of-the-squares method.
,g o Forthemajorityofdatareduction,( }RMS acceleration spectra '
were produced. When the analyzer was used while in this mode, the acceleration input was internally low pass filtered to prevent aliasing errors. 0866c/0120c 060883:5 2 4 -2.-
m .. . For this reoort the units displayed in the various figures are g's acceleration (336 in/sec 2)-and mils (0.001 inches) displacement.
- 2. Peak-to Pest Accelerations The peak-to-peak acceleration values' were obtained.by visual observation of a 5 - 20 seconc time history oscillograph trace. :In the determination of a peak-to-peak acceleration value, a judgement of relative impact levels of'each time history was made to eliminate low level impacts and utilize impacts having levels. typical of the majority 'of higher level impacts. Impacting was characterized by a sudden amplitude spike in tne time domain. The sudden amplitude change rather than the actual ampiitoce values characterized the impact condition. The high frequency signal of 750 Hz to 2 KHz, which normally fullowed an impact, was the natural frequency of the accelerometer mounting and did not represent tube motion. This reduction was perf ormed independently two times for each time history and the independent values averaged. This reduction technique provided an average acceleration value, and an indication of the repeatability of the process.
- 3. Phase and Coherence Plots Phase- and coherence are dual. channel functions perf ormed by the Nicolet 660B FFT analyzer. These plots were generally produced for a{
frequency range.
'4 Lissajous Plots Lissajous plots were also produced by the Nicolet 6608. However, this analyzer did not have the capability to double integrate an acceleration time history to produce a displacement X-Y time history plot (lissajous plot) . To obtain the time history displacements, the analog acceleration signals were double integrated external to the 6608, then processed, resulting in the. Standard format seen on these plots.
0866c/0120c 0608S3:5. 3 4-S
L'ESTINGiOUSE PROPRIETARY CLASS 2 4.1.3 Data Analysis and Interpretation . The three main products of *A da g reduction were r-average peak-to-peak 6.c . O acceler a tion,[ RMS displacement i RMS displacement L 4 spectra), andt }uacceleration
,< , o spectra. Coherence and phase between two data channels, Lissajaus (X-Y time history), and 0 - 10 KHz RMS acceleration spe:tra were also reduced but not produced in the standard data reduction procedures for each test. Interpretation and explanation of each ,
f orm' of' data reduction is, presented here for reference throughout the report.
'1. RMS Acceleration Spectra
- o.,e , t The frequency response spectra were used to evaluate tube response characteristics. The vibrational frequencies and average amplitudes were determined from these plots. Predominant vibrational frequencies were characterized by amplitude peaks. (
1 *>% t-J Some of the acceleration spectra showed sharp peaks at 60 Hz and 180 Hz. These frequency responses were electrical noise, not tube motion. The{
]a,c., ospectra were normally.used because this band includes the frequencies of tube' response that result in the majority of tube displacement. No significant tube responses occurred {
- 2. Average Peak-to-Peak Acceleration Review of 10 KHz spectrum plots indicated that the impact responses observed on the time history plots did not represent actual tube mation.
0866c/0120c % 0883:5 4 -}-4
a: .
- The main content of: these accelerations was contributed by the resonance of the accelerometer'on its mount. Review of many of these spectra indicated a similarity between accelerometers produced by the same
. manufacturer. (Note that= this discussion applies only to tube _ mounted accelerometers.)
' '3. _ RMS Di splacements
, o.p e The third' main parameter used in evaluating test cata was a
.RMSdisplacement.[
g'ure 4.1-2 is a typical . example of a displacement plot. The RMS displacement is an averaged quantity. Actualpeak-to-peakdisplacementswouldlikelybe( a sc ,e times the RMS value, dependent on the waveform.
' 4. Phase and Coherence Plots Phase and coherence spectra were used together. For the ideal case of a
. constant parameter linear system with a single clearly defined input and output, the coherence function will be unity, if x(t) and y(t) are completely unrelated, the coherence function will be zero. If the coherence function is greater -than zero but less than unity, one or more of three possible situat' ions. exist.
- 1. Extraneous noise is present in the measurements.
- 2. The. system relating x(t) to y(t) is not linear.
. 3.- y(t) is an output due to an input x(t) as well as to other inputs.
Phase and coherence were selectively processed for pairs of transducers to verify suspected mode shapes or to characterize tube-to-tube or tube-to-shell behavior.
- 5. Lissajous Plots Lissajous plots (X-Y displacement time histories) weree used to provide an indication of tube motion in the x-y plane. Interpretation of the motion l 0866c/0120c 060883:5 5 d'3
l w'as.made on a tube by tube basis and in conjunction with other forms of
. reduced data.
4.1.4 Summary of Confiaurations Tested Three 16* Model test t'ypes are discussed in this report. ( A summary of tests is _ given in . Table 4.1-1. ) l l
- 1. Baseline Tests These tests were performed to establish a reference point for subsequent modification tests and to compare 16' model tube vibration responses with the 1 corresoonding site test results for model D4 steam generators.
- 2. Tube Expansion Tests These tests were performed with diff erent numbers of expanded tubes to
-investigate the eff ect of tube expansion. on tube vibrational response.
- 3. TBX Tests i These tests were performed at flow rates based on design modification flow
- conditions and 18 expanded tubes. Tube vibration at a large number of tube locations were investigated during these tests.
l 4.1.5 Baseline Tests e 4.1.5.1 General To evaluate the eff ect of various modifications to reduce tube vibration it was decided to simulate the most severe conditions seen in Model 04 site test data. The most severe conditions for a D4 steam generator were reported in the Krsko Phase. I ' testing. l By varying support plate positions the response of the tubes was varied to duplicate as closely as possible the results obtained during Phase I at
~
0856c/0120c 060883:5 6 44, i h.-
Krsko. Tube R49C56 was used for comparison between the 16* model and Krsko test results a , b.<- y_.
. A total of 73 plate positions were tested to reach an acceptable vibration .resoonse. Figures 4.1-3 to 4.1-8 present a spectral comparison of commonly instrumented tubes (note there is -no correspondence of "X" and "Y"
- directions of accelerometers in the 16* model and Krsko) . Examination of
.these figures shows good ~ correlation of response frequencies. Tables 4.1-2 to
- a. ,c., e-4.1 4 are tabulations of . rms displacements, . average peak-to-peak a-Ga, for Krsko' baseline tests ano
}o_A,e accelerations ~ and{.
subsequent tests. These results showed similar orders of magnituce and were.
. considered an adequate representation of site responses. Responses of otner tubes instrumented in the 16* model are also presented. The remaining
. instrumented tubes showed response levels too low for evaluation of modif ic ations. The motion of tubes with low response was characterizec by test fixture' motion, which is described in the next section.
= 4.1.5.2' Test Fixture Motion - External Accelerometer Responses RMS acceleration spectra for accelerometers located externally on the 16*
model are presented in Figure 4.1-9. External steam generator vibrations measured at Krsko are presented in Figure 4.1-10. Coherence between external accelerometers on the 16* model are presented in Figures 4.1-11 and 4.1-12. Coherence between internal . and external accelerometers of the 16* model -is presented in Figures 4.1-13 to 4.1.-16. Figures 4'.1-13 to 4.1-15 indicate [
, .m c.
J Examination of RMS acceleration spectra-for these three tubes
* . c . *-
indicated a response ~ level that is( r Figure 4.1-16 l r indicates g 3* The RMS acceleration spectra for this tube indicated a relatively( [ '- (with respect to the tubes presented in Figures 4.1-13 to 4.1-15) . { se , e-0866c/0120c 060883:5 7 d-7
i c
-, o 8ee-cause tne test
}
fixture motion responded with frequencies not characteristic of ordinary tube ' response, it'was readily apparent if fixture motion was a significant part of the tube response (this can be seen by reviewing the tube acceleration spectra and the external accelerometer spectra) . This eliminated the necessity of producing a coherence plot _ to determine if a tube response was due to movement within the support plates or external fixture motion. 4.1.5.3 REPEATABILITY A number of plate positions and modifications were evaluateo during the testing phases. As a measure of the repeatability of the model responses, the original base case was repeated after the first phase of modification t'es ti ng . The new base case was designated as llRS. Figure 4.1-17 is a
. typical comparison between the two base cases. Tables 4.1-2 to 4.1 4 present typical acceleration and displacement values for the original (5R3) and tne repeat (llRS) base cases. The valves indicate good repeatability.
4.1.6 Unexpanded Window Tubes 4.-l . 6.1 Tube Response Modes
. Figure 4.1-18 indicates typical tube responses observed during testing of unexpanded window tubes. This progression of responses was obtained by varying the plate positions. AD refers to the distance in mils that tne O plate is moved parallel to and away, a from
,c, e., b the feedwater inlet. The 60 = 0 e
position resulted in a( jresponse. As the plate position changed, a { [m'oYIppears. Furthermotionoftheplateproduceda( } respense. a.h .e 0866c/0120c 060883:5 8 4-%
8 0 [the following three categories.]e ,c.,e_ Generally, the responses were classified b 3 ,c.,c in A. 0866c/0120c 060883:5 9 k3
P g a,b,c, i l l-4 mamme 4.1.6.2 Tube Response vs. Plate Position The position of support plates has a' direct effect on tube-to-support plate interaction. A number of factors influence the position of the plate and tne contact condition between the. plate 'and tube. These include:
- 1. Dimensional tolerances - As-built conditions.
- 2. Tube bowing (straightness of tubes).
3.. Static tube displacements from steady-state flow induced forces. 4 Amplitude of vibrations due to oscillating flow inouced forces.
- 5. Local flow variations.
.The existence of clearance' between the tubes and tube support piates present the possibility of numerous contact conditions and consequently many possible frequencies of response.
0866c/0120c 060883:5 12- J - .' O L
As discussed previously, the 16* model support plates were moveable to provide various< plate alignment configurations. Figures'4.1-34to4.1-36are{, - a , t, t. at:eleration spectra comparisons for tube responses versus plate positions
~
f or three unexpanded window tubes. The position of the plates (in mils) is
-given on these plots to the right of the V/EU. The position. is designated as a 6x and 6y. These spectra exhibit the response frequencies previously mentioned.(,
o.,b, c. , e , Vibration amplitudes vary significantly with response type and, consequently, vibration amplitudes vary significantly with plate position. Typ ic al [, } ' g6 values f or diff erent types of responses at location R45C56 f or 100 percent flow can be seen in Figure 4.-1-18. As shown on this figure g5 variesfL a,c,e over the range of plate positions shown.
.i 4.1.6.3 Response Variation with Flow __t , '. t-0866c/0120c 060883:5 13-4-ll
" *D
. _,i y, .
. sI g,
.l'. j .
, r 4
f y, s"'
, g. \ tg
+\ s k
j{g .
. T, t, i
~ , s .
g l t l , < ,i _. s t' t h.
-, t ,
f b
^l '
. . s '
,s.g t *
,j +g tt 7 4 .1. 6.' 4 Sumary s
, ~\ t r '~
1 1 i , 3 y x
. Generally all unexpanded window tubes tended to show similar responge u N' v ~
-frequencies and sensitivity to plate position and flow rate. The main f v
variation was seen in vibration amplitudes. Tubes located further f rom Row 49 , tended to show 1ower g6 *s. - f c , ,g ., , i ey _ n; ; a ,, N,
.d -
~%.' %
[
)
\
14 0866c/0120c 060883:5 g3
, ._y,, c . c;
.m..(~,
j y
. f 7.( s (4.l.,7 dxp anded5 Tube Tests After' completion of-a baseline configuration test,.a number of tests were Tubes were expanded at the B h carfo7med fi th expanded tubes (see Table 4.1-1) .
and 0 ' plates. Figures 4.1-38 and 4.1-39 are typical comparisons between the exoan,. fed tubes and baseline data. Figure 4.1-40 and 4.1-41 are typical comparisons between non-expanded tubes for the base and expandea tube data y s et s .0 a, b , c.( 9 4.1.7.1 Tube Response Modes
, o >% L Figures 4.1-38 to 4.1-41 are typical ( }RMSaccelerationspectra suntnaries for expanded tubes at different plate positions. At all locations where the accelerometer was placed at a support plate, the response observed i- was very los level and characterized by fixture motion (see Figures 4.1-38 and 4.1-39). The monitored locations at the midspan between support plates show two different responses. The first can be seen in Figures 4.140 and 4.1-41.
3_ Figure '4.1-42 presents an acceleration displacement spectra corresponding to a { i. N c., t: o, Figures' 4.1-43 and 5.1-44 show the second type of response. This e i > 15 'the , ] o , c , 9--response of ten seen on unexpanded tubes, described previously. A11.of'theresponsesonexpandedtubesWerelow(
- a ,c e. , ,
relati,ve to unexpanded tubes (see Tables 4.1-3 to 4.14) . O and G levels again showed low responses typically characterized by fixture motion. Figure' 4.1-45 presents E and G level g6 values for an unexpanded tube. Figures 4.1-46 to 4.1-49 present the same tube responses at the E and G level for an unexpanded configuration. Displacements are given on Figures 4.1 48 and 4.1-49. Comparison of displacements for the E level for expanded and 0866c/0120c 060883:5 15 4-/3
F-unexoanded configurations indicated a significant drop in displacement. The G level displacement response also showed a significant drop in the tube expans. ion test. This was an indication that the process of tube expansion dio not cause higher motions at the G elevation. 4.1.7.2- Response vs. Plate Position and Flew
. Figures 4.1-3E to 4.141 indicate that the response of expandeo tubes did not charige significantly with plate positions for the positions' monitored.
Tests TE-3.and 18 TBX were tests of instrumented expanded tubes for a number of. plate positions. The results of these tests indicated an insignificant change in tube response frequencies with flow and plate position variations. An' example of this can be seen in Figures ,4.140 and 4.141 wnich show expanded tube; response asf a function of plate position for 100x percent flow. This- testing " also showed lit 31e variation in g6 with plate position for expanded tubes. X Tube oisplacements for expanded tubes (based a ,s,con,9-TE-2 values) drop S. 4 of their base case value. Average signific antly a p , e., t
/ . .
peak-to-pea,k accelerations also drop
)s,b,c.gs f their original values.values TE-1 also showsfor drops inTE-2
~
(based onNE-2 Values) . The
,e displacements"and acceleration values that result in 95 values approximately
,'* s,4, < . C Lqof the base values. Unexpanded tubes tend to show displacements t-[ ,
ag
~
and' accelerations that are similar to base values.
- 'a .
( ; e / 4 6
+
i ,
-0866c/0120c 060883:5. 16 1 -/4 L
~* ,
[Q e , m 4.1.7.3 ' Tube Response Variation with Position in Tube Bundle Tube -expansion' testing was . performed for various numbers of tubes, the maximum
- being'24 The results showed relatively~ small ga values fer all expanded g ' tubes.h-
.a
}g#e,e .igure 4.1-50 shows a typical ga distribution.
. 4.1. 7- Simultaneous. Impacting of Expanded Tubes -
, ,,,e
.I f
4
.~'
s a 10866c/0120c 1060883:5 17 q . f'5
.,mg - ,,, _
s . 1- . 4 1gure l =[9.c 4.1-53 isl a typical: ex' ample of two types of-. waveforms. Note that these two responses were at the same . location . but in orthogonal. directions. The apparent : impacts which occurred in the "X" direction did not. appear in the "Y"
~
' direction.
4.1.8. Non-Window Tubes-4.1.8.1 Tube ~ Response Modes , 9 t I I . I F
<.. I
_.1 i;
'4 .1. 8. 2 Response vs. Plate Position i
I
=0866c/0120c d60883:5 19 4,.fg
^ ^
Q.%t 4.1.9 : Tube Expansion:in Conjunction with Flow Bypass Four' series of Ltests (labeled TBX) were conducted at various 16' Model water flowL rates (see T able 4.1-1). Plate searches were conducted :during each test series to attempt to identify the hi.ghest Ga value for eacn. instrumented tube loc ation. Twelve tubes were instrumented, with the instrumenteo tubes being relocated for each test series (giving rise to the series numDering) as follows: Test. Series ISTBX - Examination of expanded, window tubes LTest Series 19TBX - Examination of unexpanded, window tubes Test Series 21TBX - Continuation of-as 19 TBX Test Series 22TBX - Examination of unexpanded, non-window tubes located in the T-slbt region Generally, the The resulting highest Ga values are shown in Figure 4.1-55. results for the expanded tubes indicated the following: e q, 4 1. 2. 0866c/012d'c 060883i5 20 4 -17
- 3. The Ga values were generally very'small~ compared to the values for unexpandeo tubes. Additionally, these values did not show a high
- sensitivityL to. plat'e positions.
For.-the unexpanded tubes, the' emphasis was put on plate positions where the e a,c,t.. response. Observation of the results of Test incividual tubes -showed[J
-Series 19TBX and 21TBX indicated the following:
__. o. ,c , 9 r-- 1. 2. L.~ Observations of results for non-window tubes located in the T-slot region (Test Series .22TBX) indicated the following: __a>c,e._ 1. J 2. r 0866c/0120c 060883:5 121 4'/I
TABLE 4.1 16~ Model Test Series Test Series Searches. Flow Rates Expanded Tubes- Instrumented Tubes 70,100,112 49/56,53,48,43
.5R3 1 48/55,53,49 46/56,55 45/56
~
44/56 21/57 70,100,112 Same 11R5 ~ 1~ 12TE-1 11 70,85,100, 49/56-51 Same 112 48/56-51
.47/56-51 46/56-51
'13TE-2 1 70,85,100, -49/55-51 Same 112. 48/56,54,52,51 47/56-51 46/56-51 100 49/55-51 49/56,53,51,43
' 14'TE-3 88 48/56,54,52,51 48/55,53 47/56-51 47/54,52 46/56-51 46/56,55,51
' 21/57 25 80,90,100 49/56-51 49/56-51 18TBX 48/56-51 48/56,55,53 47/56,53 47/56-53 46/56-55 46/56,55 0866c/0120c 060883:5 22 4 - ///
TABLE- 4.1-1 (continued); Flow Rates- Expanded Tubes Instrumented Tabes Test Series- Searches 47 -80,90,100 49/56-51 '49/47,45,43
'19TBX 48/56 ~48/49,47,45,43 47/56-53 47/51,45,43 46/56-55 46/51 45/55,51 .
89,90,100 49/56-51 49/46,44 21TBX- 72 48/50,48
'48/56-51 47/56-53 47/52,50,48 46/56 65 46/54,53,47 45/47,45,43 80,90,100. 49/56-51 44/56,55,54 22TBX 75 48/56-51 43/56 47/56-53 42/56,55 46/56-55 40/56 37/56 33/56 29/56 25/56 24/56-
.21/55 i
f
+
0866c/0120c 060883:5 -23 , ;, py
.t V , _ , a,b,c.e
. Tables 4.1-2 thru 4.1-5 contain data from in-plant and scale model testing which is
+
considered Westinghouse Proprietary. o w == 1 1 i. i. i n 0020G/FTE/7-28-83 d, ' 2 /
- , , . - , , , , , . . . . - , , . - . . . . . . - . . . . . - . , . - , . . . . . . , ~ , - . . - . . . . - , - - - - , - - . . . - - , , , . - - . . - - . - , . , - . -
1 3 P
~
-- a,b,c.e 4 Figures 4.1-1 thru 4.1-55 contain data from
., in-plant and scale model testing which is considered Westinghouse Proprietary. - me l-l 0020G/FTE/7-28-83 4-db
., =
4.2 Velocity Distributions 4.2.1 2/3 Scale Model 04 Test Velocity Data 4.2.1.1 Tube Gap Velocities Tube- gap velocities were measured at the 2/3 scale model 04 test facility using the dual pitot-static tangential velocity probe ("cowhorn" probe) as described in Section 3.2.2.2. For each of the specified tube gaps, velocity
~ data were measured and recorded at eleven elevations between the "B" and "D" baffle plates. Normalized to the vertical distance between the "B" and "0" baffle plates, these elevations were nominally designated as 5, 10, 20, 30, 40, 50, 60, 70, 80, 90 and 95 percent height'. The nominal normalizeo heignt actually refers to the height of the total pressure taps ("cowhorns"). As the
- static pressure taps were located above the total pressure taps, the velocity data were ther, assigned to the elevation midway between the total pressure taps and the static ~ pressure taps. Area weighting factors were applieo to each of the velocity measurements when calculating the average gap velocity and the RMS (Root Mean Square) gap velocity. The average gap velocity is the area-weighted average velocity through the tube gap while the RMS veloc~;ty is the ' area-weighted square root of the sum of the squares of the eleven gap velocity measurements. The average gap velocity was used to calculate flow through each gap and to calculate a continuity-based scaling f actor which was
-applied to each row. The RMS gap velocity provided an indication of rela +ive q %C e loads on tubes. Tube gap velocity data were recorded at the i
t- [ minimtsn tube gap cross-sectional area, thus giving the maximum gap velocities. Typical curves of gap velocity as a function of elevation are presented on Figures 4.2-1 through 4.2-4 for "cwhorn" probes positioned in row 49 columns 37, 45, 51, and 59, respectivsly. The figure for each probe shows the velocity profile in the tube gap to either side of the probe. The eleven individual velocity data points for each gap are plotted at their nominal elevations. The figures show the general shape of the curves and the similarity of flows between gaps to either side of a given tube. The typical row 49 velocity profile can be described asf - , a A, c., e. a 0855c/0119c/060383:5 45 q43
o . a , b,c., An iso-velocity plot of the~ continuity-scaled and averaged Row 49' tube gap velocity _ data is shown on Figure 4.2-5. The plot shows contours of constant velocity. It can be seen that there was[, a , b , c. , e _ Figures 4.2-6 and 4.2-7 present both the scaled polynomial curve fit and the scatter of the scaled data points for the _ Row 49 average tube gap velocities and RMS tube gap velocities, respectively. Figures 4.2-8 and 4.2-9 show plots. of the scaled Row 47 average and RMS gap velocities, respectively. The Row 47 data was curve fit without the gap -(column)_30.5 data sets. Gap 30.5 had considerably lower velocities than the gaps near it. The gap 30.5 data is shown on the plots. The scaled Row 45 average velocity and RMS velocity curve fits are presented on Figures 4.2-10 and 4.2-11, respectively. Figure 4.2-12 ' combines the RMS gap velocity curves for rows 49, 47, and 45. Figure 4.2-13 combines the average gap velocity curves for the same rows. The f a i L -, ,w,.,e_ j Figure 4.2-14 presents a good. example of this change in velocity profile with row nunber. The figure shows the velocity profiles for the gap between I . 0855c/0119c/060383:5 46 4,74 i
b-.- .; columns 41 and 42, designated gap 41.5, at rows 49, 48, 47, and 45. Note that
'the row'48 data were actually measured in the symmetric gap, gap 74.5. The row 49 prof ile had(
. a ,b,c. 4 Figures. 4.2-15 and 4.2-16 present in another form the general trends in the change of velocity profile' with row number. Figure 4.2-15 is a plot of the RMS gao velocity as a function of , ow r number at several tube gap (column) locations. Figure 4.2-16 shows the same plotLusing the average gap velocity. __a,q r
l i l l i __J. l i 4.2.1.2 'T-Slot Velocities Figures 4.2-17 and 4.2-18 show velocity data along the T-slot and at the rear of the.T-slot. The number recorded in each tube on Figure 4.2-17 is the mean value of the RMS velocities measured with both the lef t and the right "c owho rns" . The mean value of the left "cowhorn" average velocity and the right "cowhorn" average velocity is recorded on Figure 4.2-18. P 0855c/0119c/060383:5 47 4 26
Average normal velocity-for the iront of the T-slot was determ_ined from measarements made' with- a three-dimensional prism velocity probe. The incividual' velocities which were used to deduce the average velocity _wer.e
. measured at 11: elevations, three row positions and two column positions in the T- s l ot'.
Measurements were 'made1at tube-location-equivalents of Row 24 columns 57 and
~ 58, Row 36 columns 57.and 58, and Row 42 columns 57 and 58. At any one row the pair 'of measurements at each elevation was averaged and thel average
' velocity was plotted vs.' normalized elevation in the B/D inlet pass. Figure
~
4.2-19 shows curves; of the average velocity going back along the T-slot vs. . normalized elevation for all three rows. Integration of each of the curves for the three rows then gave the following average velocities: Row Number Average T-Slot Velocity
- a 3 b , c. ,e.
-24 36 42- -- -
These velocities were plotted vs. row number and a curve drawn through the plotted poi [its (Figure 4.2-20). The curve was then extrapolated to obtain b,c. the e-
-velocity corresponding to Row 49. The velocity obtained,( .}%was thus Note-regarded as the average ' normal component of flow velocity at Row 49.
that this velocity represents full scale prototypic velocity at 'the entrance to the T-slot at 100 ' percent flow. 100 percent flow is based on Krsko 100 percent power conditions. 4.2.1.3 Tube' Bundle / Wrapper Annulus Velocity The average t.ube bundle / wrapper annulus _ velocity was determined near the front of the B/D pass from measurements made with a static-pressure indicating two-dimensional velocity probe. Both this velocity and the average velocity
'at the front of the' T-slot were used together with the average velocities of all tube gaps in Rows 49, 47, or 45 to obtain a flow continuity . check and, thereby, a correction f actor f or the average tube gap velocities in each of 0855c/0119c/060383i5 48 V2 es
f the-flow vector was .
. Note- that. only the normal flow component o these rows.
useo to determine the . average velocity.
~
insertion positions. The _ average annulus ' velocity was measured 'at three i radial sitions between. the wrapper' wall and -the tube bundle and at three elevat on po The average annulus (tangential) baffle plates. between the "B" and "D" lized radial velocity _is shown 'in Figure 4.2-21 as a ' function of normaA normalized ele
- insertion f or .various ' normalized elevations.
"D". baffle
~
plate and a norma lized insertion of represents an elevation at the 1.0 represents an insertion to the edge of the tube bundle. direct The average velocity at each normalized elevation was obtained from a
-inteoration- of each curve. The f ollowing resul.ts were obtained:
AVERAGE VELOCITY NORMALIZED ACROSS ANNULUS (FT/SEC)_ ELEVATION o. ,b , c. , e 0.14 , 0.35 0.83 L- _J the annulus Normalized elevation was then plotted , a ,% c , <- vs. average ve as shown in Figure 4.2-22. c
} f ort the annulus.
100 percent flow plotted points yielded the average velocity,[ Note that this velocity represents the prototypic velocity a (Krsko conditions) . 4.2.2 0.95 Scale Model Test Velocity Data . heater Separate test nodels of the 04, 05, and E2 baseline regions have been constructed. The data from these scoping with air. as the medium simulating feedwater flow. in the gaps between tests were reported as prototypic feedwater velocitiesComparisons of th tubes at selected locations in the tube bundle. dels are gap velocity distributions from the-respective baseline test mo , discussed below.
. 0855cl0119c/060383:5 49 9-2'7
4.2.2.1 - 04.and D5 Baseline Tests The first feedwater passes in the 04 Land 05 preheaters are very similar basea on' nominal- dimensions. The nominal specifications for the tube bundles, i.e, tube pitch, tube . diameter, bundle layout, etc,- are identical for the 04 and D5 models. .The differences mentioned below are shown in Figure 4.2-23. The following differences are relevant to flow distribution. a ,c , t (11 i i i l' (2)'
-t i
l, L_
~
The measurements of gap velocities across row 49 in the 04 and 05 models reflected' the basic similarities of the two models, as shown by the VRMS data-in Figure 4.2-24. This data is representative of 2 and 3 loop plants and was scaled to Krsko 100 percent flow conditions for both 04 and 05 models. The discrepancies between the 04 and 05 data points appear random in nature, which would result from random influences normally present in a test apparatus of this type. (, ._. p.<.e
- m ,c.,e_
0855c/0119c/060383:5 50 ,gg
. a,b,c,
~
A plot ofi iso-velocity contours across the: plane of row 49 is shown in Figure 4.2-26 for the 04 model. Since the 95 scale air model contained only a quadrant 1of the tube bundle, ' symmetry about the T-slot was assumed for
. illustrative' purposes. (, .
A , c . e_, The D5 model iso-velocity contours are very similar to the
- 04 contours at equal flowrate.
4.2.2.2 04'and E2 Baseline Tests
' The primary, differences between the 04 and E2 in the region of the feedwater.
entrance to the tube bundle are as follows: (1) ;The tube pitch for the E2 model is slightly larger than in the 04/D5 models, while ~ the tube diameter is the same in 04/05 and E2 models.
.(2) The first tube row in the E2 model is row 48, which is eight columns wider than row 49 in the 04/05 models.
u
-(3) The distance between the impingement plate and the first tube row is
{ tIeE2model. The above characteristics of the E2 model allow more flow area entering the tube bundle, suc'h that the velocities entering the tube bundle are lower after 6 ' accounting for the increased feedwater flow rate, 4.24 x 10 lbm/hr for E2 T
, 0855c/0119c/060383:5 51 g_g g
(4 loops).vs.' 4.07 - 4.09 x 106 lbm/hr for 04/05 (2 and 3 loops). Using the limiting case _ of completely uniform gap 1 velocity as a basis of c arison, tne
~
- average: gao velocity in row 48 of the E2 model is )h'dthe a
average. gap velocity in row 49 for the 04/05 model is{ } .s.c, The e increased distance between th'e-impingement plate and first tube row in the E2-model also allows more volume for the downcomer feedwater flow to become more evenlyf distributed across the plane of row 48 prior to entering the tube bundle. Test data for V RMS at the first tube row of the E2 model representative of 41000 plants is shown in Figure 4.2-27 and compared to the row'49 data for the 04/05 models. The flow characteristics of the E2 steam
- generator resulted in significantly reduced velocities.
-V data from the third pass in' _the preheater (between E and G baffles) is RMS
. shown in Figure 4.2-28 for 05 and E2 models. The VRMS distribution was
' observed to be nearly uniform across the bundle.
A compa'rison of final pass gap velocity profiles for 04 and E2 models is shown in Figure 4.2-29. The'-lower velocity magnitudes in the E2 model in pass 1 are evident. l r 0855c/0119c/060383:5 52 L
a t. 1 a,b,c.e Table 4.2-1-and Figures 4.2-1 thru 4.2-29 contain scale model test data and steam generator design details which are considered Westinghouse
~ '
Proprietary. J j' < 0020G/FTE/7-28-83 ,3f
~
c.- : 4.3 Force Measurements . 4.3'.l' Objective Force _ measurements were made to experimentall'y determine the turbulent forces acting on the tuDes in the 04 stesn generator inlet pass. In addition, the flow induced forces acting on the' force measuring tubes. in the 2/3 scale model
"' were used as input forces in t!'e WECAN non-linear tube vibration model.
4.3.2f Summary of Results:: 2/3 Model Force Measurements 4.3.2.1 Dynamic Force Dynamic force measurements were made at 70 percent, 85 percent,100 percent
' and 110 percent of f ull flow (3966 gpm) in the model. 100 percent flow is based on Euler number similitude and accounts for the difference in density between the model water temperature (typically 100'F) and the prototype (430*F). These measurements were recorded on FM tape which were subsequently analyzed' using_ a Nicolet 660 FFT analyzer to provide plots of RMS c force, cross-*>'.t-spectra, coherence and pha.se vs. fr.equency. RMS forces between)- a
.,< e.
and between _ were computed for the X- and Y directions for both
.the top and bottom of the force tube.
The square ' root of the sum of the squares (SRSS) of these values were then obtained to provich single value representing the sun,c, of 4-the dynamic loaos on the tube. Figure 4.3-1 depicts the SRSS 1 values for row 49 tubes at 85 and ,100 percent flow. conditions. Force neasurements from the 2/3 scale model were also used to provide verification of the 16' full scale model. Since the 16* model was a 90' sector of the preheater including 1/2 of the T-slot and 1/2 or a 90~ sector of the water box, measurements were made in the 2/3 model _ configured in a 90'
, sector. A partition plate was installed on the inlet pipe, water box and down through the T-slot. A' comparison of forces, at selected locations, between L
the 180' _ and 90* configuration of the 2/3 model is graphically shown in Figure l 4.3-2. Figure 4.3-3 shows a comparison of force in the T-slot region (column l 0855c/0119c/060383:5 54 ._a, 2
O .. 56). These ' comparisons: show that the 90 configuration did not understate the
-J dy,namic : forces even for tubes along the .T-slot. 'Thus the vibration excitation of: tubes along the.T-slot.was conservatively simulated in the 16' model. -
~
'4.3.3 'Ouasi-Static Tube: Drag Forces Prototypic steady drag _ loads were obtained at eight tube locations from the 0.95 scale D4. air mode?. -Incremental forces at ten (10) elevations along each ,
tube were calculated by integrating the measured pressure distributions at
;each elevation over 'its respective area. The resultant force on the tube was obtained by. summing the. ten incremental foces. The results are shown in Figure 4.3 4 t
0855c/0119c/060383:5 55 ,$- t
. - - . - . = -. - , .- . . . .
t 1 t 3 9 b d
- a,b,c.e Figures 4.3-1 thru 4.3-4 contain scale model test ,
- data which is considered Westinghouse Proprietary.
ei d T 1 I e i .
+
l l-l l I' - l. l f l ~ r
. 0020G/FTE/7-28-83 4-39
-r-- ~ --
,~w. yp, v- ,,g, , ,,,-~nwa.w,,,r,~ rw,w,e+,s--n,w~-,ve,,.,,,-- ,--,,-g,---,w .,w-,r-,-,a-,-e ,s--.-r ,,-,w-
8 s 4.4 Work Rate W' ear Coefficients 4.4.1 Introcaction Laboratory wear test programs were' conducted for the Model ~D4/05/E design modification evaluation program at both the AECL-Chalk River Nuclear Laboratories and at- Westinghouse. These test programs examined in a parametric manner the wear interactions of the various prototypic tube versus tube support plate -(TSP) material couples- with prototypic. geometries Specific areas addressed in this section are: (1) the absolute magnituoe of the work rate wear coefficients for specific steam generator models, (2) the , relationship between tube motion / force and wear coef ficients, and (3) the
-relative tube to TSP weight loss and wear coefficients.
The purpose of the parametric wear test program was to determine from experimental methods the wear rates for various tube-to-baffle. interactions under prototypic conditions. These interactions included those associated with the Model D4 as-built units throughout the operating range and the interactions anticipated for the modified Model 04 steam generator with expanded tube configuration at full power conditions. The wear rates associated with the Model 04 as-built units were confirmed by comparison -to wear conditions observed in the field in Model 03 steam generators having the same tube / TSP material couple. In addition, wear tests were performec for tne Model 05 tube / TSP material couple using previously tested Model 02/03/04 test parameters (i.e., motion, force and geometry) to evaluate the relative wear properties of- the different material couples. Several Jinvestigators have reported a significant change in the wear mechanism between two materials as a function of the type of motion that is being exhibited. Three types of motion have been investigated: " rubbing" or
" fretting" defined as one where the motion does not result in separation of the two surf aces; " sliding-impacting" defined as one where the motion of the specimen results in separation (lif t-off) of the surf aces; and " impacting only". Orders of magnitude (10-100) difference in the material " wear rate" are not uncommon between the three types of notion, with the " rubbing" motion having a lower wear rate than " sliding-impacting" but higher than " impacting 0866c/0120c 060883:5 24 4-35
.w ., --
only." The . reduced wear rate -associated with " rubbing" motion has been attributed in -the literature to a build-up of a debris layer of wear products tnat acts to recuce metallic contact I) On the other hand, when separation of surf aces occurs, higher-wear rates result because of the removal of these wear products. I Additionally the nature of the applied impact, direct vs. 031ique, has been cited as a major difference in tests reported in the-
-literature.( }
4.4.2 Wear Test Motions The wear test tube' motion matrix consisted of:
- 1. Tubes rubbing against supports without tube- lif t-of f.
- 2. Tubes rubbing in small arcs on the baffle hole with occasional lift-off.
- 3. Tubes orbiting in the baffle hole with considerable lif t-off and impacting.
] ,r_, emotions for both the as-built and modified steam generator configurations were represented by the wear test matrix. .
Typical representations of type 2 and 3 motions above are shown in Figure 4.4-1. These tube motion patterns were obtained from wear test machines with lift-off capability. . a , c., e_ In these figures the motion is displayed with variable scale f actors to allow visualization of small motions, such as case "i" which has 'a 5-7 mil travel as opposed to large motions, such as case "f" which has a travel equivalent to the as-built tube to baffle hole clearance. 4.4.3 Wear Test Evaluation Wear test results presented in this section are from Model D2/03/04 test programs at both AECL-CRNL and at Westinghouse. Also included are the results of the Model 04/05/E test program at AECL-CRNL. 0866c/0120c 060883:5 25 4-34
.Weignt loss. .or wear volume determined from the measured weight loss, was utilized -to judge the wear rates observed in the tests. The independent-parameter' chosen to measure the severity of a test or level of effort applied
.to the .tes't -specimen was the product of force and distance. This parameter was chosen because it oefines work from first principles. Sliding and rubbing
' reciprocating sliding) wear coefficients have also historically been defined using these parameters (11' . Considering the more complex tube-baffle interactions. involving impact and sliding it was found that previous
. investigators studying compound impact-(impact and sliding) in fundamental experiments observed that the degree of sliding motion (defined by a slip term k) in the studies) associated with impact was a first order consideration .
To quantify these parameters in the tests reported here, the force associated with each increment of' contact distance was integrated to characterize the work. A work rate was derived by dividing by the total time in contact. Inis method of evaluating the test data had a similar objective as that used by Ko(3) in which a histogram of force levels was used to weigh impacts observed. .For well defined motions this work integral simplified to the product of constant normal contact force per cycle, distance per cycle and number of cycles per time interval. This approach was used in tests with exact repeated motions (i.e., Westinghouse Rubbing Tests).
's ,
For the AECL tests, inductive-type high temperature displacement transducers were used to monitcr tube movement continuously throughout the testing. ~The reaction. forces at the TSP were monitored only during room temperature pre-test and post-test calibration runs by using four miniature force transducers (piezo-electric type) . An assumption was made that the forces measured at room temperature were equal to those experienced during the high temperature and pressure testing. This assumption was reinforced by observations that the tube motion did not change significantly due to f acility heatup. The forces reported were the average of the resultants of the pre-test and post-test measurements. The AECL work rate wear coefficients were calculated by dividing the measured wear volume by the product of the work - rate and the testing time. These wear coefficients were based on corrected wear volume using control test specimens. 0866c/0120c 060883:5 26 y.- 27
LTable 4.4-1 presents the AECL-Phase I wear test results for the Model 02/03/D4 steam generator' material couple (i.e., SA285 Gr. C Carbon Steel with Inconel 600' mill-annes!edi. Table 4.4-2 presents the AECL-Phase II wear test results f or the Moael 05lE steam generator material couple (i.e., 405 Stainless steel with Inconel 603 Thermally Treated), and for the Model D4 expanded tube configuration.
. Presented in Table 4.4-1 are five unique tube motions ( A through E), developea by using unique tube " excitation force ratios. The applied contact force magnitude was determined by choosing suitable eccentric nesses for each of the two drive motors. Photographic records were made of these tube croits
-throughout the tests. -These records were reviewed to group the actual tube 1 motions.
Presented in Table 4.4-2 as tube notions. (M) and (N), are the Mocel 04 expanded- tube test results. The primary purpose of the subject tests was to determine if small gap sizes associated with expanded tubes led to recuction inthewear' coefficient.{ e% b, c .e.
-s Fretting tube motion wear tests were performed at Westinghouse for the Model D2/03/04 program, using the SA285 Gr. C carbon steel (TSP) and 1600 MA (tube) material couple. Table 4.4-3 presents these fretting test results for both prototypic temperature (motions J and K) and ambient temperature conditions (motion L), all at prototypic pressure in AVT water. All Westinghouse fretting _ tests were conducted with a reciprocating tube motion having no lif t-off and with relatively large forces as compared to AECL testing. For all fretting tube motion wear tests at prototypic temperature, 9 o A , c. e.
the work rate wear- coefficients were . J In performing the Westinghouse fretting tests, a multi-specimen positioning and loading fixture was used'to test eight pairs of specimens simultaneously. This provided the same relative displacement under identical temperature and 10866c/0120c 060883:5 27 4-39 m
orgssure conditions'for eight separate wear tests. As noted in Table 4.4-3, forcomparisonofwearcharacteristics,testswererunwithTSPspecimens. _ .,c,e_ 6 -' 4 The applied normal contact force for the Westinghouse fretting tests was obtained by compressing an accurately calibrated coil spring, calibrated at both prototypic and ambient temperatures. The forces reported in Table 4.4 - are the pre-test' nominal. values whicn were used to calculate the wear Coefficients. For' all wear testing, accurate weight loss measurements were made on b3th the tube and TSP wear specimens. Both wear specimens had a control test specimen
~
that simultaneously experienced all test conditions (except wear) and cleaning procedures. The control specimens.were also weighed to allow deterT:ination of a corrected weight loss for each wear specimen. These measurements provi;ec a comparison of relative tube to TSP wear volume and wear coefficient. AECL sliding-impact test results for the Model D2/03/04 material couple
.indicat.ed
- , c e.
i a weighted tube to TSP wear volume or wear coefficient ratio of( Similarly for the Model
- s , c, t.
D5/E material couple a weighted tube to TSP qa.,c,Q. wear ratio of was indicated for prototypic conditions and for ambient test conditions. Westinghouse test results for the Model 02/03/04 material couple for f retting wear'with prototypic test conditions and TSP roughness - a , c indicated
- e. a tube to TSP as compared to the AECL wear volume or wear coefficient ratio of a , c., e_ - J r
sliding-impact value of } 4.4.4 Wear Test Conclusions As identified by previous investigators a significant difference was observed for,the Model-02/03/04 impact-sliding and rubbing motion tests. This was observed in both ambient and prototypic tests in several test configurations. 0866c/0120c 060883:5 28 4 39
. :3 The prototypic temperature rubbing test wear coefficients werefL
- o., b, c., o and the prototypic temperature imp, act/ sliding coe#ficients for the asgi{t configuration were characterized by a value.of
( Figure 4.4-2 presents the wear volume rates from
~ prototypi.c rubbing - testsiand prototypic sliding-impact tests plotted as a a , b, c, e.-
f unction of work rate. The value of was used as a
- 9 -
a , % , c., t wear coefficient for rubbing motions, and - for impact / sliding motions. .'In cases where impact leveis were very low a 5, b, c, o
.was used.
conservative value of ., g' S 4 The W-AECL measured Model 02/03/04 as-buil't swear cbefficients were compared to that deterT.ine- ' ram ur?.~ i#ied Model 02/03 'ield /dat'a.A, . he field data e , G-ind'cated typical vaIaes of abbut and the test value a,b,c, ,- was The prototypic W-AECL tests resultea in ? higher coefficients than similar impact and sliding tests perf ormed a,t ambient ,
%c., e_ '
8 g ^l. m , As was presentedlearlier, tNe 500*F carbon steel /I600 MA wear pair exhioited comparable wed tcefficients over a wide range of force and motion types ,
. tested at AECL. 1m Thu's, no correlation of wear coefficient with' any of the test us parameters was apparent in the 500'F AECL tests'. The 500'F Westinghouse fretting -' test data exhibitedk i
\ - .L
' '- his was due to a combined change in
~
l '
. motion and force.
{N y 4 i Testing at- 100'F and 100 PSI A for the 405' SS/I600 MA wear couple showed an r , o., c i e-approximate f actor of { 'gn wear coefficient over the 500'F/900 PSI A tests. Besides the diff erence in test temperature and pressure, these specimens differed in TSP surf ace finish. The TSP. surf ace of the 100'F (low wear) tests hada.,c.e-a smoother finish than the'500*F tests
~
Because of the magnitude of the AECL material properties ano surf ace _ roughness diff erences they were considered as secondary f ac3grs and y thus the _ test temperature and pressure were considered tof3,ejthe primary factors for the wear reducticn. Between temperature and pressure, temperature had the more significant effect on wear rate as determined frun the Westinghouse fre,tting data base, t, s:- t q A y ; 0866c/0120c 060883:5 29 .s4-40 .
- __ ._. h __ ._
!3.'
uh r
,; 4.f.S^Re'erences g 1. Sarkar, A.D. " Wear .of Metals", Pergamon Press,1976.
flh <,W, . _ thf ( , -/,.
- 2. Engel, P. A. et.al. '" Impact Wear Model. f or Steel . Specimens", Wear, 23 (1973)f135-201.
. )), [y(',k,. ,., e,; -, v/ \ t
'of ~
" Heat Exchariger Tube Fretting-Wear: Correlation of Tube
~
,, 3.y K6; P.L'. ~et.al.
a'#Motton and Wea'r",- Presented 'at the ASTM Symposiun on Materials Evaluation .
\ U der Fretting-Conditions, Gaithersburg,'MD, June 3,-1981.
r. 475 4kurr'icks, P.L'. .
- ".The Mechanism of Fretting-A Review", ASME-WEAR. Vol.15.
-. t i o pp. 3892409', Aoril '1970.
' 5. Ko, P L. "We'arlof Zirconium Alloys Due -tc Fretting and Periodic .
Impacti ng", ASME-WEAR, . pp. 388-390, 1980.
- 6. : Ko, P.L. " Experimental Studies of Tube Frettings -in Steam Generators and
* . Heat Exchangers", Journal of Pressure Vessel Technology, Vol.101, Ma/,
1979.. '/ 9 s t
x E
e 4 gh - 30 0866c/0120[ /060883:5 4-4 l
} ,q , l' q -~ --- - . . _ . _ _
e a,b,c.e Tables 4.4-1 thru 4.4-3 and Figures 4.4-1 and 4.4-2 contain wear test details and results which are considered Westinghouse Proprietary. - t 9
/
e !7 0020G/FTE/7-28-83 4 Ab , u
5.0 VISR ATION MECHANISM 5.1. Introduction The purpose of the following discussions of the mechanisms for flow induced tube vibration in counterflow preheat steam generators is to provice a basis for evaluating the modified steam generator configuration. The modification is shown to be eff ective in reducing tube vibration, independent of
- assumptions regarding the partir.ular mechanism (s) that operated in the as-bui't configuration.
The three main flow-induced vibration mechanisms thdt can cause tube vicration in heat exchangers are: o Vortex Shedding o Turbulence o Fluideiastic Excitation Other possible mechanisms, which are not believed to be relevent to the counterflow preheat steam generators, are pump pulsations and primary flow excitation. The relative significance of the three main flow-induced vibration mechanisms are discussM in Section 5.3 with respect to the types of flow fields that exist in the steam generator. With respect to the natural frequencies and mode shapes for counterflow
. preheat steam _ generator tube flow induced vibration, the field vibration measurements and various model tests indicated that{
- 7. +
5.2 Summary and Conclusions
- 1. The unexpected mean velocities and fluctuating velocities generated by the preheater inlet configuration are the main cause of the tube vibration in the front rows of the preheater.
0855c/0119c/060383:5 56 ~ 6-l
4
- 2. Bypass of a percentage of the main feedwater flow and the expansion of selected tubes at B and D baffles significantly reduces 'the peak velocities and turbulent forces acting on the tubes and greatly reduces '
vibration and wear potential.- ,
- 3. For the as-built configuration, field measurements showed that(
s, c., e_ ,
'4 The nonlinear tube motion involving impacting and rubbing of the tube within the support plate clearance complicated the interpretation of response spectra and time histories of the tube vibration.
5. n view
]a.I<. e. of these observations, the potential for fluid elastic excitation was considered in the design selection for the modification, the planning of vibration
' tests and the methods used for wear assessment.
- 6. Wear assessment methods reflected the wear measured on tubes pulled from Kr sko. Results of nonlinear analyses based on t'urbulent excitation were conservative when compared to pulled tube wear scars. The Ga method measured tube. response. Ga values- and correlations with RMS force included Krsko and model data from tubes estimated to have stability
~
ratios greater than unity. The plate search method used to establish Ga
- values in the 16* model resulted in data which .were in modes that maximized tube respon5e and included the modes that had s'tability ratios exceecing' unity.
- 7. . Vortex shedding is rot believed to be a significant cause of tube vibration in the counterflow preheat steam generators.
0855c/0119c/060383:5 57-
O e l 5.3. Mechanisms of Flow-Induced Vibration 5.3.1 Vorter Sheddino Methods for calculatin tube vibration caused by vortex shedding are available-in the -literature U'2' ' . However, the practical significance of vortex shedding in closely packed-tube arrays is cuestionable. In a recent paper
'PaidoussisI ) notes that, "to this day, no ;.ne has shown that the observed periodicity in cylinder arrays corresponds to anything resembling .
conventicnal, von Karman vortex shedding". He suggests that' within an array there may not be discrete periodicity, but simply a quasi-periodic tak in the turbulent energy spectrun as postulated by OwensIOI. When periodic vortex shedding does occur, the alternating forces on a tube occur at a shedding frequency f3 given by SU (1) f
- ir s
where S,_ = Strouhal number, a constant for a particular tube pattern U = fl ow 'v elocity D = tube diameter Vortex shedding is essentially a forced vibration and significant tube vibration would occur only when the vortex shedding frequency is near a tube natural frequency. Therefore, spanwise variations in flow velocity '
. greatly reduce the effectiveness of vortex shedding excitation since the shedding frequency will vary in the spanwise direction. In investigations conducted by AECL II f or nonuniform flows similar to those in the wrapper 1 5ynchronization. effects can occur (4) such that resonance can occur is the over a natural flow velocity range in the vicinity of f 3 = f n, where f n frequency. ,
0855c/0119c/060383:5 58 { -3
inlet region of recirculating steam generators, no significant vortex shedding excitation was observed. No evidence of vortex shedding has been found in the various fim ano vibration tests that simulate the counterflow preheaters for both the as-built and modified configurations. Hence vortex shedding is not believed to be a significant source of excitation in the counterflow units. 5.3.2 Turbulent Excitation The character and intensity of the turbulence in the' approaching flow depends strongly on upstream conditions such as impingement plates or fim smoothing devices. Turbulence causes narrow band random vibration of tubes in a linear system at about the natural frequency of the tubes in the fluid. Figure 5.3-1 shows a typical plot of the tube vibration. The vibration amplitudes vary randomly in time and in direction. IOI for linear systems based A simple, semi-empirical formula can be derived on concepts presented by Fung I9I and Keef e for a single cylinder. It predicts the vibration of a single-span tube with uniform cross flow over 'its entire length: y oD 2 0.5 0.5 [=C1 (,o ) h)n S (6 n ) ( ) (2) where y = root-mean-square (rms) vibration amplitude in the nth mode n of vibration Cy -- empirical constant related to the magnitude and spatial correlation of the excitation forces S = empirical constant related to the shape of the power spectral density curve for the excitation forces. f = tube natural frequency in the nth mode of vibration D = tube diameter Axial flow turbulence also causes tube vibration; however, it is usually small compared to vibration caused by cross flow. , 0855c/0119c/060383:5 59 g.g
m y ., A eg= fluid density m - tube mass per unit length including added mass of fluid outside the tube L = tube length e n
= . damping. in the nth mode of vibration expressed as the logarithmic decrement U- = cross fi m velocity ,
Typically 1 < S. < 3, and, S and C idepend on the tube pattern and sp' acing. The purpose for providing Eq. (2) is to identify the main parameters involved in turbulent excitation of tubes. For tubes on multiple supports subjected to spanwise flow variations, a somewhat more complicated formulation is required. In the case of the counterflow units the effects on tube vibration of the _ turbulence generated upstream of row 49 are significant as inoicated by nonlinear analysis results. Data on turbulent excitation were obtained from the 2/3 scale model. It provided quantitative information for the forcing function that was usa 1 in the nonlinear dynamic analysis model to predict tube response as discussed in Section 7.2. 5.3.3 Fluidelastic Excitation A major cause of tube vibration that has resulted in rapid deterioration of tubes in heat exchangers has been determined to be fluidelastic excitation, a self-excited vibration mechanism. The' mechanism is characterized by a critical flow velocity below which the vibration amplitudes are small and above which the amplitudes increase rapidly, see Figure 5.3-2. The vibration occurs at a natural frequency of the tubes in the fluid, and the tubes vibrate with orbital patterns as shown in Fig. 5.3-3. The orbits may be steady or they may precess. The fluidelastic mechanism for tubes in cross flow was first identified and characterized in 1970(11) . The simplest design formula for estimating the above which large-amplitude fluidelastic vibration critical flow velocity (8) initiates is given by 0855c/0119c/060383:5 60 65
U,, m5gn (3) f3 n
.s eD j
in its simplest form by ( ) The effective velocity U e is d? fined 2 2 g 2, 'n U (z)d (z) dz (4) e ( 617
- 2) dz g
When the ratio eU /U c is greater than one, fluidelastic vibration may initiate. The above definitions and equations are similar to those used by others. (2,13,14,15) Recent experimental ano theoretical studies ' ' ' suggest that the critical velocity may be better determined by the relationship U m a y (6) (5) dn = B.( f 2) where e, y, a may vary depending on whether liquid or gaseous flows are cons idered . There is considerable work being done in this ar,ea, however, conclusive data and agreement is not yet available. Accordingly, the simpler formula given in Eq. (3) is still usually used for design, particularly as a relative normalization tool for comparing designs that operate in the same fluid property range.- A fluidelastic type of vibration can also be caused by axial flow (I I , however the velocities required fcr their initiation are very high and generally out of the range of normal operation for steam generators.
- In' addition to cross flow that penetrates the tube bundle, skimming flow Relatively little across the outer row can also cause fluidelastic vibration.
information has been reported for this type of flow. However, that which has been reported (20) indicates that the general form of the instability
- equations for skimming fim are similar to those for flow that penetrates the tube bundle. Hence both types of flow can contribute to the potential for fluidelastic vibration.
4 6(z) is.the mode shape and U(z) is the spanwise flow velocity distribution. 0855c/0119c/060383:5 61 g L.m
/' i ' i k i _ i i i l 5 l m 5.4 Effect of Steady Fluid Forces on Tube Vibration -
. a ,c,4 h
h r i 1 1 r 0855c/0119f.f060383:5 62 g7 .
a , e , e. 1 (
~t 4
o
~
5.5 Indications of Fluidelastic Excitation in Counterflow Steam Generators Krsko and 16* model data were evaluated for indications of fluidelastic excitation. 0855c/0119c/060383:5 63 6-E
y
- p. .
)- o., c., G F i i l i t l - i i i Fluidelastic stability calculations are discussed in Section 7.3. 0855c/0119c/060383:5.65 6-7
t J
- a,b,c.e Figures 5.5-1 thru 5.5-3' contain plant and scale model test data which are considered Westinghouse
,, Proprietary. _,
F d 0020G/FTE/7-28-83 b'
.j.
5.6 References
- 1. M.1J. Pettigrew,' Y. Sylvestri, and A.0. Campogna, " Flow-Induced Vibration Analysis of. Heat Exchanger and Steam Generator Designs," AECL 5826, August 1977.
'2. F. L. Eisenger, " Prevention and Cure of Flow-Induced Vibration Problems in Tubular Heat Exchanger," Flow-Induced Vibrations, ASME, p47, 1979.
- 3. R. T.. Hartlen, Experience with Heat Exchanger Tube Vibration in a large Utility; A Viewpoint Regarding Field Incidence, Design Analysis Methods, and Preoperational Testing," Flow-Induced Heat Exchanger Tube Vibration, HTD vol 9, ASME, pp 35 42,-1980.
- 4. 'H. J. Connors, " Vortex Shedding Excitation and the Vibration of Circular Cylinders," Flow-Induced Vibration Desion Guidelines, PVP-Vol 52, ASME, pp 47 73, 1981.
- 5. M. P. Paidoussis, "Fluidelastic Vibration of Cylinder Arrays in Axial and Cross Flow-State of the Art," Flow-Induced Vibration Guideline, PVP-Vol.
52, ASME, pp 11-46, 1981.
- 6. P. R. Owen, " Buffeting Excitation of Boiler Tube Vibration," J. Mech.
Eng. Sci ., 7, pp 431-439,1965.
- 7. M. J. Pettigrew, J. L. Platten, and Y. Sylvestre, " Experimental Studies on Flow Induced Vibration. to Support Steam Generator Design, Part 11:
Tube Vibration Induced by Liquid Cross-Flow in the Entrance Region of a Steam Generator," presented at the Int. Symp. Vibration Problems in Industry, Keswick, United King' don,1973, Paper No. 424 (1973). H. J. Connors, " Flow-Induced Vibration and Wear of Steam Generator
~
8. Tubes," Nuclear Technology, Vol. 55, Nov. 1981.
- 9. Y. C. Fung, "Fluctating Lift and Drag Acting on a Cylinder in Flow at Supercritical Reynolds Numbers," J. Aerospace Sci., 27_, pp 801-814,1960.
0855c/0119c/060383:5 66 5-II -
.~ _
-l
- 10. R. T. Keefe, "An Investigation-of the Fluctuating Forces Acting on a l Stationary Cylinder in a Subsonic Stream and of the Associated Sound i Fielo," UT/A Report 76028-1, Univ. of Toronto, 1966.
l
- 11. H. J. Connors, "Fluidelastic Vibration of Tube Arrays Excited by Cross Flow," Flow-Induced Vibration of Heat Exchanaers, ASME, pp 42-56, 1970.
- 12. H. J.- Connors, "Fluicelastic Vibration of Heat Exchanger Tube Arrays,"
Trans. ASME, J. Mechanical Design,100. pp 347-353, April 1978. l l
- 13. M. J.- Pettigrew and P. L. Ko, "A Comprehensive Approach to Avoid Vibration and Fretting in Shell and Tube Heat Exchangers," Flow-induced Vibration of Power Plant Corr.,,onents, PVP 41, ASME, pp 1-18, 1980.
14 W. J. Heilker and R. Q. Vincent, " Vibration in Nuclear Heat Exchangers Due to Liquid and Two-Phase Flow," ASME Paper 80 C2/NE-4,1980. l 1 l j
- 15. R. E. Franklin and B. M. H. Soper, "An Investigation of Fluidelastic Instabilities in Tube Banks Subjected to Fluid Cross Flow," Trans. 4th Structural Methods in Reactor Technology Conf., Paper F 6/7, 1977.
- 16. S. S. Chen, " Design Guidelines for Calculating the Instability F low Velocity of Tube Arrays in Cross Flow," ANL-CT-8140, December,1951.
- 17. H. Tanaka and S. Takahara, " Unsteady Fluid Dynamic Force on Tube Buncle and its Dynamic Effect on Vibration," Flow-Induced Vibration of Power Plant Components, PVP 41, ASME, pp 93-107, 1980.
- 18. H. Tanaka and S. Takahara, "Fluidelastic Vibration of Tube Array in Cross Flow;" J. Sound and Vibration, H (1) pp 19-37,1981.
I
- 19. M. P. Paidoussis, "The Dynamics of Clusters of Flexible Cylinder in Axial Flow: - Theory and Experiments," J. of Sound and Vibration, 65 (3), pp 391-417, 1979.
l 0855c/0119c/060383:5 67 g-fZ
L-.., i , o < - H. J. Connors, "Fluidelastic Vibration of Tube' Arrays Excited by e; . 20 . Nonuniform Cross Flow," Flow-Induced Vibration of Power Plant Components,
'PVP;41, ASME, pp 93-107, 1980.
9 ( f
\
~
4 I J J
'0855c/0119c/060383:5 68 5-13
a a i l l l l p l M - ~~. ~. ' . s~-
.y .
?
g, '
* 'r. -
'[ i-$ . ' iS' ~_ ;', '. M t' I
W II l l l ll I
- n ll lll Tee '
' ' (b) l j ll{ijklil !! i lMTOTNj,I' I n?M i li"Ik d l','/['N '!. /I,'lI l i fee (c) T*'*3 *"" " * ' ' '"6' *d 6Y "
- Fi9ure 5.3-1 turbulence: (a) typical s-y picts of tube muispan am-phtude. (b) tune hastory for transverw (y) directaon, and (c) tune lustory for streamwsse tz) dmcuon.
o 9
- - - - . - - - , . . , - - - , ,_._...-....m.. . _ , . . ,,m-,,,,_--m _._..,-_,m_,__.___-.m .-.._,,_...,_,..,m_.---.y,--.w., , _ -
(
)
i 03 , ,
)
a.is - _ E= I a i - _ a k< .ia
.I s i t
$ 4 *
\ Y 0D5 ,n
,f t
,?
, 3,'
,o a.5 is u Cretuut veiecq Reno,U/Uca l
! Figure 5.3-2 Typu r.p.a r . name er uon arnputude I that occurs when the threshold instabdary velocity Uc for the array is escreded. ( ., L l l
- < J l
l 't l l e _, ._..-,.. - , , _ ., _ . _ , _ , . , , , _ , _ . - - . . , - . . . . - _ , . . _ . , , _ , - . ,, , _ , ~
l l ( EEEEEEEEEEl Emmum El
.MEMEREE I f EFJ F E5EMI EIEEE56@J,1l2llkiiF22 EEEEEG EMII EEEMEE 15E555555EE1
, _ s n.. t .
w .. . . . Tuess*. e nut nrtting, an.the snyssa
- ///////////////////
_u.LQOOOO ; ( 00000 , 13 ,, , t 33
,% 00000 o 6 n ..
O0000'o OOOOSf ' 00000
///////////< // ////
, s Figure 5*3-3 W =ms han d rods *
- w %
dag) array. Lift direcuan a vertacal and drag darecuan is horuontal. Flow a fraan left to right. l l i i I
i
- 80
..1 i !
- 40 40 --=
l i 1
- - - - Z 33.35 H
,.e ~
O (e) W t* Z 524 H2 L - 4. 0 (b) V hf A=,
- - h7 A
Z 8.34 Hz 0 A (c) l. Y t* - - - - W Z 33J5 Ha l ks n A G m (el
,# = m % %% a
,' , *~s %
/
7 - - - - D <s2.o He 0 (el Mode shapse and natural frequencies with and with-
' out dearance at tne center moport and ample Figure 5.4-1 appans at each end. '.
(9) zero clearance, (b) zero clearance. (c) v4 ration with antinode at center apport,(d) v - bration made has node at center mpport where the clearance enasts, and (e) v* ration with unpactang at center apport. l . O e l l -_ . . -
.. m 6.0 THERMAL /HYORAULIC CONSIDERATIONS The relative thermal / hydraulic ' effects of operating counterflow preheat steam generators with. and without external split' flow were evaluateo. Comparisons were made in the following areas:
- 1. Steam pressure performance
- 2. Moisture carryover
- 3. Steam generator stability 4 Preheater conditions (void fraction and velocity distributions)
- 5. Preheater sludge deposition .
.6. Preheater vapor blanketing 16.1 Thermal Hydraulic Perfonaance 6.1.1 Overall-Steam Generator Performance Operation with bypass requires an increase in primary temperature to maintain steam pressare performance. The increase is approximately 0.5"F for eacn 10 percent bypass. With this change the overall performance including moisture carryover will be the same for the modified plants and the original design.
Without a primary temperature increase, steam pressure decreases by 5 psi for
, each 10 percent bypass.
6.1.2 Stability Split feedwater flow reduces stability but stable operation in the field has been confirmed for the case with 30 percent bypass which envelopes the planned operating conditions. 6.1.3 Preheater Conditions External split flow increcses preheater qualities. Void fractions are increased somewhat, but changes are not judged large enough to have a significant effect. 0855c'/0119c/060383:5 69 6 -l
6.1.4 Preheater Sludge Deposition Differences in potential for sludge deposition are not considered significant in v'iew of the uncertainties involved and the probable dominant effect of operating conditions. 6.1.5 Preheater VaDor Blanketing dec.eause of the slightly higher void fraction for operation with bypa'ss, sligntly earlier vapor blanketing may be expected. For split flow operation, it was judged that the changes in parameters are not of such a magnitude as to significantly affect the long-tenn acceptaoility of this mode of operation. 6.2 Water Hammer The eff ect of the split feed operation on the anticipated water-hammer transients for the ste,am generator was reviewed. It was concluded that no additional water-hammer evaluation must be performed for the steam generator as a result of the split flow modification. 6.3 Structural Evaluation Structural analyses were performed for the main and auxilary nozzles and the upper internals structure for the split flow leading conditions. Analyses of the upper internals included an evaluation of striping. Based on these analyses it is concluded that the most . limiting cross-sections for the main and auxilary nozzles continue to satisfy ASME Code, Section III
. requirements. For the upper internals the following conclusions are made:
L 'l) The central drain satisfies the fatigue usage allowable for both transient and striping conditions. l 0855c/0119c/060383:5 70 4 -2 L
2 A f atigue usage of 1.0 due to transients alone is satisfied for the
~
21 auxiliary nozzle discharge pipe, but is predicted to be greater than 1.0 for the intermediate deck plate.
'3 } - The;f atigue-usage due to striping exceeds a f atigue usage of 1.0 for both the intermediate deck plate and discharge pipe.
The striping evaluations for each of the components incorporated conservative
~
estimates of fluid temperature gradient, film coefficients, and oscillating frequency of the temperature. .These conservatisms reflect the uncertainty in predicting the characteristics of a striping flow field. Based on the results I^ of striping tests, however, it is expected that the affected components are less severely loaded than determined from the conservative estimates. In addition, ' based on past experience, crack propogation analysis is expected to show limited growth of a crack and result in a conclusion that cracking would not be a safety hazard. 0855c/0119c/060383:5 71 43
, a 7.0 WEAR-ASSESSMENT METHODS / ANALYSIS RESULTS Several methods have been developed and applied in the counterflow preheat steam generator' evaluation program in assess.ing tube wear. The various methods employ. operating plant data and scale model test data as input in determining ~ parameters. which, with appropriate coefficients, provide an assessment.of tube wear. This section describes the bases for each of the
-methods and the results obtained. In addition, the results from 'a number of the methods are correlated to provide a basis for the overall assessment of wear.
7.1 Ga Method 7.1.1 Basis ? >c .
- r 0855c/0119c/060383:5 72 -
'I- \
e --
a , c , e. The wear coefficient, K, was obtained by applying the Ga method to a known wear. case for which acceleromater data is available. Wear volumes have been measured in the. laboratory for three (3) tubes removed from a steam generator
- at a non-domestic operating plant. These three tubes had been instrumented with accelerometers during prior operation of the unit. The measurea wear volume, by baffle location, is listed in Table 7.1-1. Applying the wear volume data with gs values determined from accelerometer data measured at or adjusted to the E baffle plate and with plant power history prior to tube removal, wear coefficients were determined. The wear coefficient calculation is tabulated in Table 7.1-2. Note that the K values for the
~ three tubes are a,b,c,4 l-Wear coefficients based on the g6 values at the inlet pass (B-D midspan) have also been calculated for the 04 steam generators. From scale model test data, o, c, -
o 7.1.2 Verification Ga wear coefficients have been calculated for D3 model steam generators using g5 values .for two tubes from Almaraz 1, and ECT wear measurements for the equivalent tubes at Ringhals 3. The resulting wear coefficients and corresp'onding wear volumes are given in Table 7.1-3 along with the D4 values from Krkso. The average coefficient using combinedSRinghals r , c, 3 and Almaraz 1 Both the D3 and D4 wear ECT data was calculated to be( coefficients are the same order of magnitude. Using the maximum and average 03 coefficients with Ga values from Almaraz 1, the maximum and average wear volumes for D3 tubes R49C51 and R49C71 were determined and are plotted in Figures 7.1-1 and 7.1-2. Figure 7.1-3 shows the Ga method applied' to calculate the Ringhals 3 measured scar depths using the 0855c/0119c/060383:5 73 g_g
maximum wear coefficient and the worst case scar volume to depth correlations (see Section 7.4). This assessment is typical of the, method as applied for saf ety calculations and was shown to bound all scar depth measurements. These plots show the Ga wear estimates to be greater than that determined from ECT measurements. In addition, wear _ estimates were plotted using Ga data from tne SSPB f ull scale, 03 model test, at the equivalent of full power (92.5 percent model flow) . These Ga values were used with the maximum D3 Ga wear coefficient. This data showeo that the ECT wear volume was encompassed by the SSPB Ga predicted wear volumes for both Ringhals 3 and Almaraz 1. 7.1.3 Ga Correlations 7.1.3.1 Ga - Velocity Correlation For the Ga-velocity correlation, RMS tube gap velo ~ cities were obtained from the 2/3' scale model 04 baseline testing at prototype full power conoitions. The velocity for each tube was determined by averaging the gap velocities on either side of the tube. For power lev (.ls other than 100 percent, the velocity was determined to be linearly proportional with power level. This is based on flow test results which indicated that RMS velocities are approximately linear with main feedwater power level, for power levels between 70 percent and 100 p arcent. A linear regression analysis was done on E baffle Ga r data foreeight
, *-A - tubes from Krsko at or near the "T" slot which have a dominant [ }responsemode. Data was taken at 100/0, 90/10, 80/20 and 70/30 split flow conditions. I t was found, as a result of this analysis, that(
a , e , c. See Figure 7.14 A linear regression analysis was conducted on all the E baffle Ga data from F
- the 16* baseline model and Krsko. The analysis results, shown in Oi - ure t- Again, only 7.1-5, indicated that[-
r A ,c. *- mode were those tubes at or near the "T" slot responding in a ( considered . The data plotted for the 16' baseline model includes power levels In combining the Krsko and 16* of 112 percent,100 percent, and 70 percent. Baseline data, -it has been seen tnat the data was well grouped and easily 0855c/0119c/060383:5 74 'l-3
- e. , c., e fitteo.
This indicates good correlation between plant and 16* model data. 7.1.3.2 Ga - Turbulent Force Correlation The turbulence induced dynamic forces acting on tubes in the steam generator preheater produce dynamic tube motion. Measurements of the dynamic forces were made in the 2/3 scale D4 model. The dynamic force measurements were
~
' obtain'ed during model flow operation at 110,100, 85 and 70 percent of f ull
-Krsko flow. It was found that the total PMS force over a frequency bandwidth
, * >c. t- c : a- c . *-
-e of( _ j(scale model frequencies 4 ; prototypic) provided a measureofdynamicforcewnichwasmorecorrelatiblewithGathanthe[
forces. , o , , e. After analysis of the total force data for a specific tube at each of the four (4) model flow set points, the data set was curve fitted for RMS force (
, * . t. C jversuspercentflow. The RMS force vs. flow was generally fit with a - 4. ' i '
r RMS (force data for flows other than the model flow test points, such as those, for Comanche Peak, were determined by interpolation on the curve fitted to the data set. (Comanche Peak test flows of 80, 90 and'100 percent correspond to Krsko flows of 74.4, 83.7 and 93 percent respectively.) The 2/3 scale 04 model RMS forces were then scaled to produce full scale plant RMS forces. To determine a Ga-tube turbulent RMS force correlation, maximum Ga values were This selected for each tube location included in the available data base. data base included Krsko 70/0, 70/30, 80/20, 90/10 and 100/0 data, 16* model data for these same Krsko flow conditions, and 16* model data for Coninanche Peak flows of 80, 90 and 100 percent. The RMS force data and the maximum Ga values were then plotted to determine the correlation. This is shown in LFigure 7.1-6. Reference to the figure shows the equation for the curve which 0855c/0119c/060383:5 75 7.q
- e. .
.results fram a method of' least squares linear regression curve fit. ' In:
accition, the _ coefficient of' determination-(R) for each curve is presented which demonstrat'es the quality of-the fit. A w s
)
s i F i.-
-I i
l i- [ s [ . 0855c/0119c/060383:5 76 L. S
.e ,.
j . x .. ,8 - 3s ..i ) 8.;~L () u 4-
- '( -
( i
. Table 7.1-1 x n ,
- 4. -
=
- s. a
-Pulled. Tube. Wear Volume Data .
,%p-. . s 3
' V' ' '
O ~ Wall Reduction Tube Baffle + ,- 1.'6ar' Vol dnO 2' ( i n ) 3 mils- l h; Percent , (..4
-gg.l a h ,
.l -
a'
-. R49C 56 B._ 'w(Q 2 i20 +
.w' i- "f ' i'-[(_
- y s (s 6 :2.5
-~.) , G. 3 N 0.7/
0.61 x-10-4 2 - - 0
\s'\
s s
\
+ V
.I -
G O'.56 x.10-4 3 '\l.3 i N m .
- .} l pg >
r ~ t. ( .
)
, 2.4 x 10-4 5 2.0 )
JR46C56- B ,- ,
- v. r D _.- ,s 0. 5' x 10-4 '- 2 } .0 -
s (* - , 1 y \ .- i K-},yj - . , R49C35- , G .' y 0.l}xi10-4' i 2
1.0
%, 4 %.,*-
'[ ( \ %, ,
. s. .. 4- . ,1
( x .
- 4. ., 1
<ki 4
s (1
., d\ i 3
1
~
b .. - -j \ 7, y, , y = s ,
% )
~ 3 $< .
. I a , s
- i
( \ wy t s'. v.
.L. x ,
.l>
< > .x wnx '
s;, t N s,,1 \ y,r % s
-N
'\ H s
\ ..
y 1 4
- g .
e s s
~ ~ ' -
,N , i.:
. t i
-- x g
LO855c/0119c/060383:5:77 y '} q T s .,. _ ,,g, . ______________s___,31- ,
. .x .
* 'Y e , . r
' d, ' .
[.. , 7 ,. Y i ,
,s ,.
-' i r i I F
-. a,b,c.e Tables 7.1-2 and 7.1-3 and Figures ~7.1-1 thru 4
, 7.1-6 contain Ga data from plant and scale model
.s
- g,5 testing which is considered Westinghouse T 'P'ro'prietarf e , ,
k IW
'*g.
4 '. ,' )$
':% % . r..
\'?,'
4 3 _$
> s'
~
i!*.
't ;
7 h i.
,4 A.
s k' q T, 0020G/FTE/7-28-83 g_7
a { 7 4 '
- 7.2 :lNon-Linear -Flow Induced-Vibration Analyses - Work Rate Method
- j--
-7.2.L -Introduction and Summary
~
F rs c .This section presents 1the ' method based-on the application lof generic analytical tools, originally developed by Westinghouse in support of the Model
'03 program, to t6e Mo,dels D4,L S[and O E steam generatora.. . This method was used to,.obtain1 turbulence induced vibration response for a steam generator tube, it was ' developeo- to , deal with J a non-linear structural system comprised of, a stesn. generator tube and the tube support plates through which it passes.
.This.-system is geometrically non-linear due to the gaps between the tube and v ,. <s th,e boundarg.ofc the holes in the individual plates through which the-tube passes'. -g t '. s
- ' :;,5 <j Th'e' core of t$emanalytical -tools was the . flow induced vibration (FIV) model.
m ;j.
.-" ' This model pro i'ded the capability of determining the tube response to
~
.I
. turbulent excitation in.yterms of the time-history motions of the tube inside
'_ each plate hole',_as ,well/as the force' between the tube and the plates when
. contact occurs / . .
- a., c., e_
Section -7.2.2 summarizes- the conclusions of the evaluation. Qualification of Reproduction of wear rates and , L the FIV model 'is addressed in Section 7.2.3.
< 3- .
. scars compatible with those observed on -the pulled tubes from Krsko is the Section 7.2.5 describes the 04 FIV model and gives subje,ct of Section 7.2,.4.
=,
,a, the bases for both the boundary conditions and the input excitation required
./,
for the#a'nalyses. Section 7.2.6 discusses conservatisms in the FIV model.
/
Section 7;2.7 presents the plan' used in the application of the FIV model in assessing tube wear. . t m/ 0855c/0119c/060383:5 80- 9 - $. F ~ s c.,p+ww- ,, w e--' -- ew t-m=, ++,e+- - t-,- . - - - r r ~e--- '
,1v v -v-e *+m -r-- r-
7.2.2 Synmary The FIV model assessment of work rate and tube wear considered window ano non-window tubes, both expanded and unexpanded. A design basis or best estimate assessment 'showe'd significant work rate reductions for all of the cases considtred'. A limiting. safety assessment applied conservative, upper bound correlations J and input for use in projecting limiting wear rates. 7.2.3 Qualification Experience with the Model D3 design modification program providea a broad basis of comparisons between FIV model calculated parameters and their physical counterparts as measured both in the field and in various test facilities.- These comparisons included tube response parameters, such as RMS displacements and. contact forces, displacement spectra, and the g6 parameter, as well as wear related parameters such as wear scar arc lengths and azimuthal orientations and work rates which led to strong positive correlations to measured wear volumes. These excellent comparisons were obtained through a range of 40 to 100 percent of full flow (power) and for conditions' representing essentially the entire front row of tubes as well at tubes inside the bundle. - Thus, the FIV model ~ and methods which were employed for the Mooel D4 evaluations have been extensively qualified. l-l 7.2.4 Analytical Reproduction of Rrkso Field Wear Scars Nondestructive examinations were made on three tubes removed from a Krsko . Model D4 steam generator. Results from these examinations include both the wear scar azimuthal locations at the plates and their associated wear scar v olumes. Figures 7.2-1, 7.2-2, and 7.2-3 reproduce the reported wear scar locations at the'B,- 0 and G plates for the three pulled tubes, R49C56, R49C35 and R46C56, respectively. These figures represent views from the top of the
. pl ates. The water box is in the positive X coordinate direction anu the fs
.feedwater flow is opposite this direction.
0855c/0119c/060383:5 81 g,g o _
y , , -__. b In; order to'show that equivalent Krsko wear scars were analytically reproducibie with turbulence excitations when- the plate positions- were within Model D4 drawing tolerances, the objective was set to determine a single set of plate positions which would simultaneously produce tube total work- rates and arc-lengths at each of the three tube positions compatible with the ~ wear volumes and arc lengths' observed on the three Krkso pulled tubes. This exercise was completed using measured D4 turbulence force spectra appropriate
'to 'each of the three positions. The final plate positions were restricted to ,
In addition, it be within. drawing tolerances' of the nominal plate positions. is important to note that only the plate positions were varied during this
~
exercise. No attempt was made to superimpose plate hole diameter and/or position tolerances on the plate positions. 4 After experimentation with the Fly model, in terms of' obtaining both pre-dynamic static" solutions and dynamic solutions, a set of plate positions was determined which produced the wear scars shown inside the circles on Figures 7.2-1 through 7.2-3. A comparison of the total wear volumes implied
~by t5e analyses to the measured ' total . wear volumes showed good agreement at both the R49C35 and R46C56 tubes.
One further plate shif t (within hole to.erances) l provided a very good match for the R49C56 tube. Here the analytical volume was about 1.5 times the measured volume. . Field _ and analytical spectra are compared for tube R49C56 on Figure 7.2-4 and ifor tube R49C35 on Figures _7.2-5 and 7.2-6. From Figure 7.2-4 it is seen that both the field and analytical spectra for tube R49C56 show(
~
- s e
~ ~ fn,c.,igure7.2-5the F
~ ' analytical spectrum is compared to the upper E p1 ate field spectrum for tube
}a ,b, c., oOrthogona R49C35. Both spectra show a[
spectra for tube R49C35 are given on Figure 7.2-6 where the upper D plate field spectrun is of interest. ( o .1. . Based on the data presented above, the f ollowing conclusion was considered appropriate: . Measured D4 turbulence excitations, applied at three different tube-positions with in-tolerance plate positions, produce wear scar volumes
- 0855c/0119c/060383:5 82- rp g
and work rates compatible with the observed Krsko pulled tube wear scars and wear volumes. 7.2.5 Single Tube FIV Model Description A three dimensional model of the tube and support plates from the tubesneet up
'to the top support plate was constructed with the WECAN compt.ter program.
This finite element matnematical model is referred to as the tube flow induced vibration (FIV). modei. {
+, c, e j A-schematic of the tube dynn.c model including the details at. the plates is shown in Figure 7.2-7. A discussion of the boundary conditions and the input excitations for this FIV model follow.
7.2.5.1. Preheater Model and Boundary Conditions Initial conditions and boundary conditions for the FIV model were based on static analyses which addressed the salient normal operating thermal loadings. Theseincluded{ y,<.< 7.2.5.2 -Turbulence Force Input This Section provides the bases for obtaining turbulence excitation measurements and the sequence of data reduction steps used to obtain force time-history excitations for input to the non-linear dynamic (FIV) model of a single steam generator tube.
' Turbulent- forc.es acting on Model 04 preheater tubes were measured in the 2/3 scale D4 model. Force gauges installed at the top and bottom of simulated 0855c/0119c/060383:5 83 p.7/
t-
e
.. a tubes-(rods) measured forces in two normal directi ons lateral to the tube and in the plane of the support plates. In this configuration, the rods became spatial pressure integrators which monitored the net resultant static and dynamic forces acting on a tube in that position due to pressure variations in the surrounding fluid flow field.
Time-history force measurements' from these gauges were recorded for a
'relatively long time period and then converted, via a mathematical From transf ormation called a Fast Fourier Transform,. to the frequency-domain.
these flow test force results, four plots of force data in the frequency domain were obtained for a given tube position. Each plot represented the data recorded by one of the force gauges on the rod. The plots were root-mean-square (RMS) Force Spectra and characterized the average energy of the random fluid forces and their distribution with frequency. Note that a RMS' Force Spectrum is the square root of the more familiar Power Spectral Density plot where the engineering units on the vertical scale would be pounds squared per-Hertz. The RMS Force Spectra of the 2/3 Scale. model measurements were then converted to full scale'RMS forces in the proper form to provide excitation input to the FIV model for each 'of the tube ocsitions and power levels tested. This conversior. producea a synthesized force time-history 'with statistical characteristics very similar to the original set of random force signals but without the 60 Hz line noise and the natural frequency response of the measuring rod. Thus, apolication of the synths. sized forces to the FIV model for short times led to three dimensional tube vibration responses that accurately simulated their long time physical counterparts. The synthesized force time-histories obtained from the test data were then applied as input to the FIV non-linear tube model as time-histories of force.
.These-four force histories, representing the four test force gauge measurements, were applied simultaneously but with zero correlation in the cppropriate x and z model directions. To obtain a physically realistic distribution over the length of the tube, equal portions of the forces from
- the top gages were applied at nodes located /8 and 3 /8 below plate D.
0855c/0119c/060383:5 84
Likewise,. equal portions of the forces from the bottom gages were applied at nodes t loc ated /8 and 3 /8 above plate. Here, represents the vertical distances.between plates B and D. 7.2.5.3 Drag Forces The steady flow .(drag)' forces applied to the FIV model were obtained from two
. sources, the first, a correlation between the drag force and local tube velocity measured in the 2/3 scale model tests, and the second, by direct measurement in the .95 scale air model.
The .first method was based on the classical . equation F D = 1/2 M DLD C where a velocity relationship was substituted for the coefficient of drag (CD ) au CD = A(Re[. Drag forces for the D4 FIV model ,were calculated using the local velocity 4 distribution for the tube in question. Results were then scaled to prototypic conditions. The second source of drag forces employed a pressure integrator at ten elevations'en the tubes in the first pass of the preheater ~in the 0.95 scale air model of the 04 steam generator. These measurements provideo a direct drag force input to the FIV model. 7.2.6 FIV Model Conservatisms 7.2.6.1 Effect of Friction The first order effect of friction on work rate based wear time estimates was
$6,E 0855c/0119c/060383:5 85 ry , f 3
..; ' :o ; .. '
.v - my C ,9,_
- To quantitatively assess.this effect, zero and non-zero friction coefficients
~
were considered in the FIV'model evaluations. 7.2.6.2 .0ther Conservatisms in the FIV Analyses In' addition to the friction eff ect there were two other eff ects which were judged to have significant-conservative influence on wear time estimates. The first of- these dealt. with scaling of the 2/3 model turbulent forces. To obtain' full . scale forces, the. following scalir.g rule was used F, o,V,2A, V" oV Z p pp Ap
' For equal Euler nebers o,V, =oVpp which reduced the present s'caling to
. F, A p = 7m = (2/3)2 =g 1
P P and Fp = 2.25 F ,. a , e_ 4 s L_. '
*F --f orce, o = mass density, V = velocity, A = area, v = viscosity,
~
D = length, and subscripts m = model and p = prototype (fullscale) . [ 0855c/0119c/060383:5 86 fj,g
_%E [ l t
~~ __
- 7. 2. 7 FIV Model Evaluation Plan
. The FIV model provided the flexibility for assessing the relative eff ects of various- f actors related to the flow induced tube wear issue. The effect of the tube support configuration was evaluated by conside ing in the assessment expanded _ and unexpanded window tubes, and expanded and unexpanded nonwindow tubes. Cases were run for_ various friction coefficients to assess the eff ect discussed in Section 7.2.6.1. A number of series cases were completed to consider the eff ect of' support conditions (tube / baffle gaps) changing with time as a' result of wear.
Finally, on the basis of the previous studies, safety cases were completed, applying the equilibrcting forces discussed in Section 7.2.6.3 to provide
. limiting wear assessments for expanded and unexpanded window tubes. For this l_
limiting assessment, tne friction coefficient applied was well below the minimum friction coefficient neasured in AECL wear tests. Maximum turbulent force loading was applied. l t
.c 0855c/0119c/060383:5 89 g_;g
o .. b '
,..-,Q.
~~~~% \ s x'
/ ., -
,,,.-}
'N N hW ,
i N: '
\1 1 i I i O- ,
I
- 5 l l 2 IB
.Di I i 1
/
l
% \. 8 ,.
. '/
j i N '
/t.
*~.
x, - s s ~% .../'
< >\
Note: Field Data Outside Circle f Analytical Data In Circle Figure 7.2-1 Reported Krsko Wear Scar Sites and -t-Analytical Scars for Tube R49 C56
.5. ,
.- -O~.
p-- K r g \ ..
.p s. s
\
,s'.
%. \
./ %. .
A 4 4 O \
/ \
'o /
/J u c t /
I
,l "z 1
} T
\' / :i -
\ B
/
k -
\
G '/ f' (- % g.s " ' s [ .~.,...e-i
,L ,
s' Note: Field Data Outside Circle Analytical Data in Circle I i Figure 7.2-2 Reported Krsko Wear Scar Sites and Analytical
\ Scars for Tube R49 C35
+ ,
~
9 .#.- fl
.f /
f^f,/ / < ^j ! Lt *
. fr [?
'r D ;
e
- r E
- O {' :
k I . 5 / S p' w, e, * ,
~
G s [
/ .- 'I .
h ._s , , G - g p ..
- (-
Note: Field Data Outside Citcle Analytical Results In Circle Reported Kre.ko Wear Scar Sites and Analytical i , Figure 7.2-3 Scars for Tube R46 C56
L .,. : s. 4 1 4 a,b,c.e Figures 7.2-4 thru 7.2-8 contain details and results from the FIV analytical model which are
- considered Westinghouse Proprietary.
. 1 l
l i f i- [. l. I
-0020G/FTE/7-28-83 ']-l$
7.3 Linear' Flow -Induced Vibration Analyses 7.3.1 Summary and Conclusions
-The dynamic response of row 49 tubes 'to flow induced excitation in the Model D4 Steam Generator preheater was analyzed using a linear .two dimensional mathematic al- model . The results were used to evaluate the effectiveness of proposed modifications for reducing tube response caused by the fluidelastic vibration -mechanism and to assist in the interpretation of field and model tube vibration data..
Average cross-flow velocities for input into -the analysis were determined from tube gap velocity data from the 2/3 Scale Water Flow Test Model supplemented
-by thermal hydraulic computer analysis.
Good . agreement between the calculated tube response frequencies and Krsko field data was demonstrated for a support case in which the tube was [ .- e ,c, e. u Tube support Case 2 This is designated as tube . support ' Case 1 in the report. represents the tube expansion modifications where the reduced hole clearances at the "B" and "0" plates were-modeled as supported elevations. - a ,c., 7.3.2 Model Description TThe lumped mass dynamic mathematical model of a row 49 tube is shown in Figure.,c,e_ 7.3-1. 0855c[0119c/060383:590 9 -L O
' , c-;f Tube natural frequencies and fluidelastic stability were evaluated using the z
linear vibration model. . Cross flow velocities in the preheater region were input based upon base , case gap velocity data from the 2/3 Scale Water Flow Test.Model . supplemented by. thermal hydraulic analysis.
-- 4 a,c,e
, a , c, t-
.The most act i ve tu bes, however, exhibited a significant{ 3 response over a relatively wide range -of powar levels. The presence of a strong and distinct
, => c, o
. response in the vicinity of( } suggests that the tube response can be
- represented approximately in a linear manner at that frequency over the particular ~ range of power levels. Consequently, a linear tube vibration model was developed which adequately represents this mode. Linear model results were!also consistent with test data at other well defined response peaks.
- 7.3.3 Natural Frequencies and Mode Shapes Nat' ural frequencies and mode shapes were obtained using the linear model for several tube support conditions that correspond to measured tube response frequencies for expanded and unexpanded window tubes and for non-window
. tubes. Selected mock shapes for two cases are plotted in Figures 7.3-2 and o ,c,G.
For tube support Case 1, 7.3-3. =, c, a- . b the corresponding f undamental freque.ncy is For tube support Case 2, s,6,c,e
- r. athe e, fundamental mode L.
)does not involve significant tube motion in the preheater.
0'd55c/0119c/060383:5 91 7-3 l
2- . 7.3.4 Fluidelastic Evaluation Fluidelastic stability ratios (the ratio of the eff ective velocity to the critical velocity) were calculated to evaluate the potential benefits of the
' design modification .regarding fluidelastic excitation. Incorporation , ,.
of bypass flow reduces preheater velocities to approximately( {pe,c.,4 rcent of the unmodified two loop plant values and provides a corresponding improvement in the stability ratio. The expansion of tubes in the areas of highest inlet pass velocities also improves the stability ratio. For tubes located it
. . a c e.
R49C53, the stability ratio is improved by a f actor of approximately( !as 'a result of the effective tube support configuration of Case 2( a . c . 9-as opposed to case 1 -
, c, .
- a,c, + -
. The stability' ratio of tube R49C53 is calculated to be ess than for the modified 04 four loop plant.
Stability ratios were also calculated for unexpanded window and non-window model 04 tubes at the following locations: 2nd row unexpanded tube R48C41, 6th row T-slot tube,- back of T-slot tube and a tube near a flow buster. The results for two sets of parameters indicate stability ratios of:
~~
- o., c. A
~'
) 5 These support conditions represent the condition f ound in 16* model pla e searches or -in the plant data. [
. [ Tub s at similar . locations in Model 02/D3 generators, however, have not indicated significant wear by eddy current tests.[ .
] his support condition h h not been observed in plant testing and is considered to have a low liklihood of occurrence.
*Pettigrew provides a damping equation that gives a value of [
In Pettigrew's formulation damping is inversely [* Hz for model 04 parameters. proportional to frequency. 0855c/0119c/060383:5 92 9-22
The effect of excitation' mechanisms acting on tha. tubes as reflected by tube vibration data and tube wear data is incorporated in the wear assessment meth od s . Gs, values from plant and model data include values fran tubes calculated -to have stability ratios greater than unity. Ga correlations utilized in the assessment of the design modification also include tubes
' having stability ratios exceeding unity.
A good correlation has been found between Ga and turbulence force over a wide
, range of Ga values. The non-linear model has been found to conservatively predict the wear of tubes pulled from a Krsko steam generator which include r W' tubes that are in the high velocity area and vibrate inn, modes. While it cannot be conclusively determined whether or not fluid elastic excitation is a significant contributor to tube response, the methods applied for wear assessments adequately predict measured wear and tube vibration including tubes having calculated stability ratios exceeding unit.
7.3.5 References (1) H. J. Connors, "Fluidelastic Vibration of Heat Exchanger Tube Arrays", Trans ASME, J. Mech. Design,1-00, April 1978e pp. 347-353. (2) M.. J. Pettigrew and D. J. Gorman, " Vibration of Heat Exchanger Components in Liquid and Two-Phase Cross Flow," AECL 6184, May 1978. ( l l t 0855c/0119c/060383:5 93 9-13 i
~
?
- a,b,c.e Figures 7.3-1 thru 7.3-3 contain details and results from the linear dynamic analysis mode) which are considered Westinghouse Proprietary.
-- W>
J v t e s. 6 1 f
~
- 0020G/FTE/7-28 -
- -,. . , . . . . ~ . , . . . _ . -
pr '7 t
- . Wear Depth to. Volume Relationships 7.4.1 Sinale Scar Depth to Volume-7.4.1.1 ' Calculation Method Relationships between single scar volume and maximum scar depth have been
~
determined for a t'ube passing through a baffle plate hole. { 3 The same approach was used for the calculation of the wear volume versus wall scar depth for expanded tubes, the only addition required being the local value of' the expanded tube radius as a function of length along the tube. This -local relationship of tube radius to axial length was based upon measured va10es for tubes expanded by the hydrau-lic expansion process. 7.4.1.2 Correlation of Calculated and Measured Results The single scar wear ' volume'to depth curves developed are shown on Figure 7.4-1. The unexpanded tube design basis curve is for a t'ube having a .750
%%T-e inch diameter baffle hole and havirg an angle of
. inch diameter tube-in a( % **'-
r The scar volume to depth for this inclination of hole and tube axes ofi . tube is an excellent fit to the measured volume and scar depths determined from tubes reaoved from operating steam generators. The volume depth curve for..he expanded tube was developed for the same initial gap and inclination angle as the unexpanded case c but ,sc with aa " flat"
,- ..c e expansion length of with a final diametral gap of( } The " flat" is
-0235c/0119c/060383:5 94 rf 36
that axialilength of the expanded zone over which the diameter varies by no more than .002 inches. - r P Ae-- The-upper bound curves'were calculated using an initial diametral gap of [ { inches and an inclination' angle of[ } For the expanded tube the same o.,t e_. conditions applied except that the expansion used was a{ " f i at"
,Sc,e-exp anded - to a[ jfinal diametral gap.
Figure 7.*-2 shows these curves superimposed upon measured single scar wear volume and dep' t h data for tubes removed from D3 and 04 steam' generators. The best fit to the measured scar data is shown as a dashed curve. The wear data from the operating plants is bracketed by the analytically developed curves. 7.4.2 Sinole Scar Volume to Total Volume 7.4.2.1 Calculation Method Curves of. single scar volume as a function of total tube scar volume were developed using the non-linear analysis model (Section 7.2) for both window and non-window tubes. These curves are shown on Figure 7.4-3.
'The curves were developed from iterative calculations made using test model .
data as input to the nonlinear analysis program, yielding maximum single scar volume and _ total tube scar volume for all the interacting baffle and tube intersections. The curves in Figure 7.4-3 are the best power curve fits to the calculated data. Based on the ~ analytical results showing differences between window and non-window tubes for the single / total scar volume ratio, the 02/03 measured data were re-evaluated to separate windw and non-window tubes. The separate data were fit with a power curve and showed differences very similar to the analytical results. The 04 analytical results showed qa.A E 7 L 1 7.4.2.2 . Design Basis and Upper Bound Curves
.The. range of wear volume data measured on tubes removeda ,c, o from D2 and 03 steam based on 2a uncertainty generators led to the selection of a f actor of 0855c/0119c/060383:5 95
').-2 6
levels: to b'e applied to the 04 design basis curves to give the upper bound l rel_ationships, shown. on Figure' 7.4 4
~
These curves, used' in conjunction .with the. single scar volume to scar depth ~ curves presented on Figure 7.4-1, allow the determination of the total tube wear volume as,~a function ~ of . maximum scar ' depth. Table 7.4-1 lists the single scar and total tube scar. volumes for. the design and! upper bound cases for
~
window - and non-window tubes. . 4 y. r I i r
-0855c/0119c/060383:5 96 g 3 '/
p d - g , , - --e ,--- - a e-.,w - s,..,i, -
.g.e--.;----+'- -"wWw g = - - - * * -' _
6
'~ a,b,c.e Table 7.4-1 and Figures 7.4-1 thru 7.4-4 contain analytical. wear sear data which is considered Westinghouse Proprietary.
J i.
/
{ l'. f-f i l_ 0020G/FTE/7-28-83 7,.29
7.51 Statistical Evaluation of R49C53 Preload Forces 7.5.1 Method Description
^7.5.1.1 Variables Required For' Analytical Model To determine-preload forces generated by tube-plate interference, a dynamic finite element model was constructed. The model input required the definition of tube , sheet angle, plate hole position relative to the tube ~ sheet centerline, and the radial gap defining the position of the tube centerline with reference to the hole centerline. These values vary due to manuf acturing tolerances, manuf acturing procedures and assembly, and are also aff ecteo by pressure ' and temperature at -operating power conditions. To account for tnese effects, a study was made to evaluate the statistical distributions for plate
_ hole tolerances from blueprints and manuf acturing procedures. Included in this study were actual measured preload forces on plates from tubed steam generators and gap measurements on expanded tubes from the pressure controlled expansion qualification test. In addition, for nonexpanded tubes, gap The compilation of measurements were obtained as discussed in Section 3.1.3. this data resulted in a statistical distribution for each variable input to
' the analytical model .
7.5.1.2 Distributions of Plate Position Variables The distribution used to establish the tubesheet angle was a function of the individual hole position on the primary side of the tubesheet relative to the hole position secondary side of the tubesheet. The tubesheet angle was geometrically calculated from a random sample of the tubesheet primary and secondary side hole positions. Plate hole distribution was, as for the tube sheet hole position, a function
. of manuf acturing procedures and tolerances from the plate blueprints.
Gap distribution f ar the expanded tube, Figure 7.5-1, was the result of defining a mean _ and standard deviation from data from the pressure controlled a, b, c., t., mean with expansion qualification test. The values calculated were 0855c/0119c/060383:5 98 7-29
. a,b A t
- a standard Ldeviation of[, These values are for both B and 0 plate elevations. The non-expanded tube gap distribution, Figure 7.5-2, was obtained ~ from the ANCO probe measurements.~ Manuf acturing data was also considered. Considering .the data from all' sources, the mean' gap value for the
- s.b.c
- non-expanded glaje elevations was(
[with a standard deviation of Tube bow was not considered in the analysis.
- ..i 7.5.1.3 The Dynamic Finite Element Model The parameters required for input to the dynamic model were based on
- statistical random-samples of the distributions from Section 7.5.1.2 using the
" Monte Carlo" methodology. Four conditions were considereo in the analysis:
- 1. Expanded tube at B and D level with baffle plates 1 free (not wedged).
- 2. Expanded tube at B and 0 level with baffle plates wedged.
- 3. Non expanded tube with baffle plates free -(not wedged).
4 Non expanded tube with baffle plates wedged. The baffle plate hole distributions were representative of the radial offset from true position. For each of the samples taken, an engle of orientation was also sampled from a uniform distribution of 0* to 2n radians. These values were then converted to rectangular coordinates for the plate position input format. 7.5.2 Analysis Results The results of the analysis are tabulated in Table 7.5-1 and represent the mean and standard deviation of the sixty sets of preload normal forces measured at the B, 0, and G plate levels. An evaluation of the samples for condition No. I was made to determine the It was found point where the mean and standard deviation values converged. This was further l that sixty samples was an adequate sampling of the data. 1 i I 0855c/0119c/060383:5 99 r),30
verified by computing the standard error of th2 mean t:hich shows the computed
~
;mean value should not vary more than + 4 percent.
~
7.5.3. Statistical Analysis Conclusion , 5 a 4 1 0855c/0119c/060383:5 100 7-3l
g-a,b,c e 3; Tab!e 7.5-1 and Figures 7.5-1 and 7.5-2 contain data from the statistical preload evaluation which are considered Westinghouse Proprietary. k 6 0020G/FTE/7-28-83 r;,3 ?
, . ._ . - , _ . . . _ _ _ _ _ . _ _ . . ~ . . _ . _ _ _ _ . _ . _ . _ _ , . . . _ _ . _ _ _ , . _ . . _ _ __
s 6 8.0 SELECTION ~0F TUBES FOR EXPANSION / MODIFICATION WEAR ASSESSME This section' summarizes.the data base for selection of tubes for. expansion, the selection of typical generic expansion zones for_04, 05 and E Models and
. the wear. assessment for tubes in the modified configuration.
8.1' Data Base f or Selection of Tubes for Expansion The vibration data base used for. selection of tubes for expansion includeo:
'a. Krsko data - Ga values were determined for 16 tube locations including tubes ,in both steam generators,
- b. 16* Model data - Ga values were determined for about 50 tube locatio the 16' sector. .
- c. 2/3 Scale Model data - Turbulent tube excitation forces were measu about 40 tube locations.
These analyses utilized the 2/3
- d. Nonlinear model tube vibration' analyses model force data to estimate tube wear at key tube locations and aided the identification of tubes for expansion, Measured Ga values were correlated
- e. Ga vs. turbulent force correlations with turbulent force data to obtain G6 estimates at tube locations not included in the 16* model.
i
- f. Ga wear correlation - Ga values were related to wear based values and wear measurements for. tubes removed from Krsko.
The above data base, described in previous sections of this report, was used to identify the high vibration level tubes for tube expansion.
. 0866c/0120c 060883:5 38 ,g
, \ sV *\ '
\
, y * *
~
i s
.The objectives for the modificatioc. have been d(scribed i'n Section 1.2 with the Ga value objectives shown/in -
r figure 1.2-1." For selections of the tubes for expansion, the Figure *1.2-1 Ga' objective was ap clied on the~'basis ' - of expanding a.' c.' e those tubes- having estimated Ga values exce(eding t _ 1 E
. \ is .,
i'
' I. 1W s it was recognized, that a f ew tubes s havingy v Ga \vgit.es similar t
, 'ta' the ,'Tig values ~ of Ficure 1.?-l could occur in 'a modifiebr gene".ator. J O4\'
qq s The general boundaries of the expanded tube regions were established based on - < , ~ Ga distributions. The GA distribution for columns 43757 was defined by direct
~
measurements within the 16* -model, f ull length tube region. For regions
-outside' the 16* model full length region, the Ga vs. tuhbulent force i_
\'
correlation-.and the turbulent force distribution from the 2/3 hcale model I we utilized. For key tubes at the boundaries'ob/Se expansion zon'e, Ga wear time n : b estimates were compared withw', nonlinear}nodehe$t'imates a to further assess tthe s -
- 3. Sl expansion zone cutoff tube' locations {or
, s specific s , ss row boundaries.
3 t w t , y a (, i ,
-t .- '
5 46. h * %. orrelation of For a Ga value oK about t- the Ga .vs t turbulent f'ctce 3 a, c%e. ; y - f< q l, 1 5 Figure s 8.1-1
\,
Figure 7.1-6 indicates an fMS Force L shows turbulent force versusx colsnn numby. We force data shc.pn here are d evaluated data based on fittirigd power curve 'for flow dependence to theii ata to minimize the influence of, experimental uncer,tainty on'tf1hi!1."Fom ga .7:.j og s- - this y t L ps exceeded f r [ g 3 1, t s-figure, it can be seen that th'e force guideline of . . -ya- ...-
't
.g'
( *
, 17. , .
i ! , s, c l
'e !
i N.o s *,, s t c e ,
?
~ ' , a , c. , o ,
' 1 ,
- s. .
s ' . ' \ ', To provide' addNichal input to the expansion None sel.ection, non-linear mode analyses sere perf ormed for key tube locations. Trese results are compared with the Ga: estimates in Table 8.1-1. b ' s , e J , 3 0866c/0120c 060883:5 39 4
.h, h ,
)
o
,- y . . 7: . M; +r -
, > s, 7
*c . __, Q%
i_ 4 p i-
. q , 3 j. .
, NY. 'f _,C ., I , ( s t'
, t ; ,) i
./
.! ( p
\1
, f i' _ ! !
(' -
. -\..- .
5 Ve 5
>/J'
' d s' ,.
s
. _ \"
s
,t .
. k
~t i
~
-- a,b,c.e
. Table 8.1-1 and Figures 8.1-1 thru 8.1-3 contain tube wear and expansion assessment data which is i
considered Westinghouse Proprietary.'
] (, .
k( -% f l g ,) )
.i' I~
( > . . :. t . . ,f" - <
^ h, L. .a . gf q' '\ 4.
L
^\g ,
A23 _ d ;.(4- f . .' \, [- y%' .
/'
's i{4
+
l gp* . c N 5., . . NN1 . _f, ,.s. i e ji.a-vit... ' ' ' L. c ,7l- , . p. t f.
- i l
f \
,. Q
-) .
- t!*- ,
f 2-F ~ i V, .
, 0020G/FTE/7-28-83 6 32 x.4 t ,
, .., _ _ . . . . _ . . . , _ _ , . , , . . . _ . . , . , _ . m.m..._,... . - , , ,_,-.,m_,_,,, , , ..,,_._ ,_. . - . - . . _ _. _ , _ . , . , _ , , , _ , _ , . . ,
.)-.
b
'} :. -,- j g .
,.g /A
..a . f *.
- 8.2 . Selection of Tubes for. Expansion
. f.' /"-
u, ,
, ., /.
The dati of Figures 8'.1-2 'and 8.1-3 and the analysis results of Table 8.1-1~ provided the . basis' for selecting the tubes ici/Ixpansion.. This section
-identifies tubesito be expanded'in Model D4' generators and the basis 'for the ,.-
W ' g 9 05 and E }seldction. F
-s .
p 8.2.1 .D4 Tube'Expansio'n = , c ,e.
-r J
====
. ;- l N
, c:,. . " .?
f
- l. ,s t
a- , t 4 0866c/0120c 060883:5 41. 61
.__..c.-
y .,
^
s e D4 ' Tube Expansion Conclusion , Based on' the above assessment, Figure 8.2-1 shows the tubes selected for typical for a
- expan; ion in a 04 steun generator. This expansion pattern '.s
- generic 04 generator. Plant specific assessments based on specific flow or operating conditions may result in minor s ie., s.-
' differences from the pattern shown in Figure 8.2-1. Overall, about(' htubes were selected for expansion.
4 From-Tabla 8.3-3, the following conclusions can be drawn. L. ' ' ' - the nominal conditions'used in the model have probabilities - of e-
*ic lower than
%.. r percent.
calculated preloads at each plate in the range ofl- a a 1
- b. The safety' case conditions 'used in the model - .,c,t, have preload probabilities percent at each plate. The of occurrence of the order of .
combined likelih6od of occurrences of the safety case would be the order of- the produce of the individual plate probabilities which implies a very
" low likelihood of occurrence.
- 0866c/0120c 060883:5 '43 8,es
.. .2
.C. For unexpanded tubes, the probability of(
a ,c, % respectively. For expanded tubes,
'b these pNbabilities reduce to appro$imately{'o'rces would fur
- . respectively. Incorporation of fluid ( }f signifIcantly reduce the likelihood of{
a ,c, e. -
; b p d. .The plate'shif t cases for R49C56 have a low likelihood of occurrence ~
- compared t'o the nominal case.
Overall, these results support the adequacy of the safety case as a limiting condition not likely to occur .and the plate shift analyset represent. a reasonable lower bound on wear times expected for expanded tubes. d r 0866c/0120c 060883:5~ 44 5-e
- . . _ . . ~ _. . _ . _ , _ , ,,, __.
o' . 8.2.2- 05 Tube Expansion The 04'and D5 steam gen'erators are very similar relative to flow induced vibration consberations a previously discussed in Section 4.2.2.1. The main differences are , a.s w as shewn on Figure 4.2-23. As noted in Section 4.2.2.1 and shown in Figures 4.2-24 and 4.2-25 as well as Table 4.2-1, the velocity differences between the 04 and D5 models are negligibic. This comparison includes diff erences[ [
'The only velocity difference of significant magnitude oetween the 04 and 05
}a,c, c.as shown in Figure 4.2-25. The modelsis(
tubes' in this region would be expanded for both the 04 and 05 so that an increase in the 05 expansion zone due to( [is n'ot indicated. r ;p As seen in Table 4.2-1, the measured velocities fory tubegnearL a
.,sThis reouction for 05 a - a.,c, e_
also applies for the 05[ as shown in Figure 4.2-24. Based on the above considerations, the expansion zone defined for the 04 model was considered applicable for 05 model. The typical, generic D5 expansion zone is shown on Figure 8.2-2. 8.2.3 E Model Tube Expansion The diff erences between the.04 and E models relative to changes in velocity have been discussed in Section 4.2.2.2 and shown in Figures 4.2-27 to 4.2-29. These figures show the generally more uniform velocity distributions of the E model compared to 04 The data also show that the E model velocities are.a l . , *A% 4-rgt row average velocity of( j ,b'e., A. l jlower than the D4 with [ ft/sec f or E . compared to{ } ford 4. i 0866c/0120c .0608'13:5 48 L
The inlet water L6x geometry for the E model is very similar to that for tne 04 with the E model dimensions slightly larger corresponding to the larger E model shell diameter. Based on the inlet area geometrical similarity and the f act that the E model tubes are, [ii' was' anticipated that the E model turbulence levels at the first Row of tubes (Row 48) would be no higher than the corresponding levels at Row 49 of a 04 model. Thus the inlet pass tube excitation forces for an E model can be well approximated by the 04 modified forces. . All tubes of interest in the E model are non-window tubes having support at the E and H plates. In Section 7.4, the relation between single scar and From Figure
' total wear volumes was developed for windw and non-window tubes.
7.4-2, the allowable single scar volure for an unexpanded tube at 40 percent wall depth is( .lNom Fi: pre 7.4 4, the expected total tube r f scar volume fqr , ne : ire.m v no - it scarvolumeod'a
%c.. -lfor a non-window tube.
[for a vndow tube :r d } s ortportional to the total wear The time ' Mr to C perc Mt depti v olume . Th us as sim. ig t h at' t ne s arm! exc.tation force results in the same tube vibratic9 le" e' ( r sx.s.,4 c.e 'l Cuv, the ti.tc to 40 percent depth for E relative to j The asse.iption of the ..ame excitation force resulting in (. D4is{ the san taba "ioration levei can ce s.mpo-ted by the 16* model, non-window tube tsirim fu :@e M9C56. As shmn ir the Ga vs. turbulent force correlativ of F ,93m , 4, the reasared maximum Ga for tube R44C56 as sire v T-slot) turbulent force, f all plotted with tne R *.. n > ms ' essentially on the Des.; fitr 1. 6.' of.rtee correlation. Tube R44CS6 data shown 1 s.: - hich was found for only one tube at mode w inthisfigurewerefora(. essentially one plate position among more than 80 plate positions searched r , afor
, c.' e.,
the response'on 12 tubes. 4
}
mode for which Ga values are lower than obtained response . Therefore the assumption of the same vibration response in Model E Thus the as obtained r for 04 at,a the e c .. tsame excitation force is conservative. above noted( . jfor E model wear times relative to 04 is reasonable an conservative. 0866c/0120c M0883:5 49 gg
O E s , c. , 4_. for E model wear times can be applied to the Ga objective . The{ .c. 4-- to obtain values or to increase the( .. acceptable guidelines for defining the expansion zone. _
- F.R-P
] l 1 a The balance of the planned exocnion zone for the E model is essentially the same as for 04 as shown by the comparison of Figures 8.1-1 and 8.2-3. As c , = ,c., o shown in Figure 8.2-3, the Model E expansion region includes about) } tubes. 8.2 Design Basis Modification Wear Assessment Table 8.3-1 sunnarizes the data evaluations described in previous section that support the design modification. Collectively, the data base supporting the modification is a substcntial body of field, test model and analysis data. In general,' consistent trends ' relative to tube vibration are shown by the data: between field and 16* model data for vibration levels; between 16* and 2/3 model ' data for the vibration level distrib9 tion over the tube bundle; and l L between the Ga and nonlinear model methods for wear assessment with the applied Ga method based on maximum vibration levels being typical of
. conservative, off nominal results from the nonlinear model.
l l l \ . 0866c/0170c 060883:5 50 gnq u . _ .
e - z
' : Table 8.3-2 shows estimates for wear times to 40 percent-wall.; depth based on
~
nonlinear model and Ga methods. Tubes in the modified configuration show' wear
] a e,4for conservative estimates and near(
. times aenarally exceeding [ 1
.<.< } wear'.. . is expected to: envelope most of j f or : nominal ' es timates .-
the tubes in the- expanded zone with the possible exception of tubes near the T-slot corners in Row'49.. For T-slot corner tube R49C56,~a. conservative wear time estim' ate is{ }mi c.tlt can be noted from the results for tube R4
- that. the 'nocif ication (tube expansion plus bypass) improves the wear times for
- expanded tubes. by 'about a f actor of v-The statistical-. analysis for mechanical-(fluid ( }n,otincluded)preload forces on' tube R49C53 described in Section 7.5 can be used to estimate .the probability that the preload forces on the tube would be less than the values used in the nonlinsar.model . analyses.> Since higher preload forces minimize the 'liklihood of tube: impacting, the probability of . lower preload forces than calculated can be used as a measure of the liklihood of the tube lifetime
-being less than calculated. Table 8.3-3 gives the tube preloads at the B and
- O plates for_ calculations described in Sectio'n 7.2, the calculated wear time to 40 percent depth'and the probability of Igwer preloads;than c'alculated.
wh'ile the statistical estimates.of Section 7.5 are approximate for a given 9enerator due to limitations on estimating tube alignment conditions, the -
. values'of Table 8.3-3 can be used to approximate the -potential for tube wear times .less than calculated.
8.4 1.imiting Tube. Wear Assessment A. limiting' tube. wear assessment is the basis for evaluating the perf ormance of ' the modified steam generator from a saf ety standpoint. The limiting wear assessment utilizes: 7 - o conservative wear volume estimates from wear assessment methods presented earlier in sections 7.1 and 7.2. o limiting single scar wear volume to total tube wear volume correlation ? 0866c'/0120c 1060883:Sc 51 6-10
~ . . _ _ . _ _ . _ ._ 2 . _ _ . . _ . _ _ m. _ _ __ _ - _ _ _ _ . _ ., _ ,.
o' limiting single scar wear volume to depth correlation o- conservative wear coefficient o tube wall loss safety limit With'this input, the limiting (minimum) time before exceeding the tube safety limit is determined. This limiting time is the basis for establishing the initial plant operating interval before eddy current inspection. Subsequent operating intervals will be evaluated on the basis of eddy current wear assessments. - . r The limiting wear volume estimate for the modified steam generator was made
'from results from the non-linear flw induced vibration analytical model for limiting expanded and unexpanded window tubes. The Ga methods also appliec for unexpanded tubes.
- , a., c, t.
discussed in The nonlinear model utilized the{ J c qs c, C Section 7.2.6.3 which effectively eliminates all tube supports. AL 2 friction coefficient was used which is well below any measured in the AECL wear test program. With the limiting total tube work rates egrmined from the analytical model, the wear coefficient of g iwas applied to provide the limiting wear rates. Using the upper bound curves from Figures 7.4-1 and 7.4-4 together with the tube wall loss saf ety limit. of 65 percent, the limiting safety case is established. .
, a , c, '
.h jwith the highest turbulent forces in the bundle, was chosen as the limiting case for expanded tubes. For this case, the limitingtige to 65 percent estimate was datermined to be approximately{
wear depth, 4 ( generator.
}a, c 4-was chosen as bounding all unexpanded tubes in the modi For this case, the .limii.idg time estimate was determined as{
a,k4.wi h the G4 w
}a'ik4h the nonlinear model and{
method. These time estimates provide significant margin relative to full cycle lengths in determining the initial plant operating interval te an eddy current inspectie. 0866c/0120c 060883:5 52 9-Il
__ ~ _ _ - . . . . _ . . . . - -. __ _ i-
,- - a,b,c.e 4
. Tables 8.3-1 thru 8.3-3 and Figures 8.2-1 thru 8.2-3 contain tube wear and expansion assessment
, data which is considered Westinghouse Proprietary. b I. t 4 0020G/FTE/7-28-83 f-/L
-a o _
9.0 SAFETY EVALUATION This section assesses the safety implications of plant operation with the counterflow preheat steam _ generator modification imolemented. The hydraulic performance and structural adequacy of the modified steam generator are reviewed from the standpoint of safety. 9.1 Hydraulic Performance Evaluation . The safety requirement governing the hydraulic performance criteria for the modified counterflow preheat stesn generator is that tube wear due to flud induced vibration not result in tube wall reduction in excess of the safety-limit for tubes in service. The safety limit for tube wall reduction is the amount of wall loss the tube can sustain and maintain integrity under the most severe accident conditions. For preheat stesn generator tubing, this limit has been determined by analysis and test to be a 65 percent wall reouction. The allowable wear rate associated with this safety requirement is dependent on the existing wall loss and the iteterval between tube inspections. Two . cases are considered; a tube with existing detectable wall loss, and a tube with no detectable wall loss.
'For the first case, the maximum wall loss permitted for a tube lef t in service is dictated by the tube plugging limit defined in the plant technical t specifications. For plants with preheat steam generators, the tube plugging limits currently in effect are 40 percent and 50 percent. For a plant with a 40 percent plugging limit, a tube with 39 percent wall loss could be lef t in service. For a plant with a 50 percent plugging limit, a tube with 49 percent wall loss could be left in service. To determine an allowable wear rate for such a tube, the allowable additional wall loss is determined with This consideration for uncertainty in the method for measuring wall loss.
allowable additional wall loss, considering an eddy current uncertainty of 10 percent of tube wall thickness, is 15 percent for a 40 percent plugging These limits, criteria and 5 percent for a 50 percent plugging criteria. along with the 10 percent eddy current uncertainty and the limiting existing 0855c/0119c/060383:5 102 f-l
wal1 loss for a tube left in service would limit the total tube wall loss to
- 64. percent. This is- below the 65 percent safety limit.
Given' the allowable additional ' wall ~ loss, . the interval between inspections
~
will determine the allowable wear rate. . In accordance with Regulatory Guide 1.83, Revian 1, Position-c.6.b, the interval between inspections should be not more than' 24 calendar months. The allowable wear' rate for this case without imposing shorter intervals between inspections _is therefore 15 percent per 24 month interval for a tube with an existing wall loss of 39 percent (40 percent plugging limit) or 5 percent per 24 month' interval for a tube with an existing wall loss of 49 percent (50 percent plugging limit). . The 24 month interval used for the first. case is appropriate for a tube with canl existing wall loss since Regulatory Guide 1.83 requires that each tubing inservice inspection include all nonplugged tubes with existing, detectable
. wall-loss. Twenty four months is the maximum interval between tubing inspections.
The allowable wear rates specified above are based on the longest interval-The safety between tubing inspections permitted by Regulatory position c.6.b. requirement that tube wall reduction not exceed the safety limit for tubes in service can be met for higher wear rates prov'ided the interval between inspections is shortened. . F6r th'e second case, the maximum undetectable wall loss is assumed. For the
' absolute. eddy current method empMying-a wear scar calibration standard, the maximum undetectable wall . loss is 9 percent. Without an indication that the steam generators in a plant are performing differently, tubing inspections may
~
be limited to one steam generator on a rotating basis. For that one steam generator, the Regulatory Guide would dictate that tubes in the preheater area, an area of. " potential problems" (Reg. Guide paragraph C.4.c), be
--included as part of the overall inspection sample.
From these considerations and a 24 month interval between inspections, the maximun interval between sample inspections in the preheater area of a given 0855c/0119c/060383i5 103 Q - 7.- 1
.o- a.
steam generator would be 48, 72 or 96 months for a 2, 3, or 4 loop plant, respect ively. On this basis, the allowable wear rate will be determined using the 96 month interval. For an existing 9 percent wall loss and 10 percent eddy current uncertainty, an allowable additional wall loss of 45 percent would limit total wall loss to 64 percent. The allowable wear rate is thus 45 percent per 96 month. interval. As for the first case, the safety requirement can be met . f or higher wear rates provided the interval between inspections is shortened
-accordingly.
While a modified counterflow preheat steam generator represents a significant improvement with respect to tube wear over the base case, the modification does not preclude tube wear in excess of the plugging limit during the operational period of the plant. As a result, some tubes in the preheater area of the counterflow steam generators may require plugging, due to wear. Based on the allowable wear rates determined above, however, any preheater tube wear due to flow induced vibration would progress slowly. This allows use of a routine steam generator inservice inspection program for detection and follow of any tube wear. This inservice inspection program, in combination with preventative plugging criteria, allows safe operation of the plant. Verification that the nodified steam generator meets the hydraulic performance criteria is based on use of a qualified, non-linear, flow-induced-vibration (FIV) ' analytical model; scale model testing, laboratory wear testing; and field ~ data from operating units. The application of each of these elements in assessing hydraulic performance is discussed in Section 7.0. In addition to verification of the FIV model with full scale model data, FIV model results have been compared to field data (tube dynamic response, tube wear mag'nitude and tube wear distribution) and verified. Additional details on verification of the FIV analytical model are presented in Section 7.2. Westinghouse has also evaluated the increase in tube to baffle plate clearance due to wear. Since the allowable wear rates do not preclude some tube wear,
-0855c/0119c/060383:5 104 7-d
o . The the tube / baffle clearance on some tubes could increase witn time. evaluatico 4 this eff ect shows that as wear progresses, there is a trend to distribute the total tube work rate at additional support plates, thus reducing the worA input at any one plate. This trend appears in eddy current test-results from field units and was verified in series analyses using the non-linear FIV model . A check on the wear coefficient determined from laboratory testing was made by back-calculating a wear coefficient for tube wear in operating plants. Knowing the amount of tube wear and the plant operating his' tory, a wear rate was estimated for selected tubes. The work rate for these tubes for the case case configuration was determined using the FIV model. The wear coefficient was then calculated.for the tube wear measured at operating units and compared to the laboratory value for the same type of tuce motion. The two wear coefficients were compatible. As a conservative measure, additional confirmation of the hydraulic performance of the modification will be provided by monitoring and inspecting operating steam generators after macification. This program includes the use of instrumentation installed inside the steam generator tubes to provide an indication of tube response in the modified steam generator for comparison
-with tube response determined analytically and measured in model testing and used in the verification of nodification perf ormance.
In addition to monitoring tube response, a conservative post-modification This inservice inspection program will be developed for each plant. j inspection program will include a conservative scope of tubes to be inspected such that tubes in the regions of maximum response (determined from mooel The inspection testing) will be sampled and bounded in the inspection plan. interval for each plant will be determined on the basis of any existing wear Tne condition in the plant and the projected wear rate for that condition. inspection interval will be selected to provide adequate margin between As projected tube wear and the safety limit for tube wall reduction. operating experience is gained and tube wear rates are better established, the inspecti- interval can be adjusted accordingly. l ft' 0855c/0119c/060383:5 105 R}}