A13175, Investigation Into Cause & Consequence of Fort St Vrain Pelton Wheel Incipient Fractures

Investigation Into Cause & Consequence of Fort St Vrain Pelton Wheel Incipient Fractures

ML20085L444
Person / Time
Site:	Fort Saint Vrain
Issue date:	10/16/1974
From:	GENERAL ATOMICS (FORMERLY GA TECHNOLOGIES, INC./GENER
To:	US ATOMIC ENERGY COMMISSION (AEC)
Shared Package
ML20085L438	List: A12714, Public Svc Corp Prenuclear Pelton Wheel Cavitation Rept A13175, Investigation Into Cause & Consequence of Fort St Vrain Pelton Wheel Incipient Fractures A13823, Fort St Vrain Dew Point Moisture Monitor Sys Post-Mod Test Results (RT-355C) A14049, Investigation Into Causes & Consequences of Malfunction of Static Shutdown Seal S/N 3 in Fort St Vrain Circulator C2102 from D Penetration
References
GA-A13175, UC-77, NUDOCS 8310280310
Download: ML20085L444 (100)

Text

.

Sog F. n:m a~

m r - p pj GA-A13175

,,jg'aj d>by $ ?

I'

-e.h

.g.. -

ll\VESTIGATION INTO THE CAUSE Q::.$.

j ...

AND col \ SEQUEL \ CE OF THE 4..%

FORT ST. VRAIN PELTO N WHEEL W; .

INCIPIENT FRACTURES ..

• F:

by N[

~. .n . :

% wg PROJECT STAFF m. -.

p. , 's M

. . fi RETURN TO Q* M I)lRECTORATE OF, REGULATORY OPERATIONS 1l$)s 4 '-

%.;;,c.

e. - r y._..f3;f;_2f.:.,..-  ; ' 0 I '- i ,.

' 1 Ji9 Prepared under E '.~ E . 3

[M. y- ,b s Contract AT(04-3)-633

? I.?.'.. f. s .

for the :7..*., ~

Nh San Francisco Operations Office 1

.s 'N i T4

- + pq U. S. Atomic Energy Commission y,tQ.2['

p .-

l' .:  ;?:[?

.  ;};jk ' y,.

.4,,4

.@.R y y. : .'fr h ;[, ,

DATE PUBLISHED: OCTOBER 16,1974 Med.)  ;., . y ' ;.;f

y :. n C Ri e o J,.

y;gs - .

.u "E_pg -

?~.f.p' *f 2.:.f*W '

U 8310280310 760930 1 'I': l; .$.4 PDR ADOCK 05000267

[F.*k pd

' i. . /;

S PDR . .

((. .Q*;-; -

p p. 4, * :;K :.

sp,; .:

. ..Yh

, .ll:.l_'k Phys ENERAL ATOMI -

GENERAL ATOMIC COMPANY P.O. BOX 81608 SAN DIEGO, CALIFORNIA 92138 i

[

, l

f. 3

V NOTICE This report was prepared as an account of work sponsored by the United States Gove rnment. Neither *he United States nor the United States Atomic Energy Com. "

mission, nor any of their employees, nor any of their contractors, subcontractors, or their e' aployees, makes any wa rranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or asefulness of any infor-mation, a ppa ratus, product or process disclosed, or represente that its use w ould not infringe privately owned rights.

?

Printed in tne United States of America Available from National Technical Information Service U.S. Department of Commerce-5285 Port Royal Road Springfield, Virginia 22151 Priest: Printed Copy $5.45; Microfiche $1.45

GA-A13175 UC-77

~

INVESTIGATION INTO THE CAUSE AND CONSEQUENCE OF THE FORT ST. VRAlb PELTON WHEEL INCIPIENT FRACTURES PRINCIPAL CONTRIBUTORS E.A. BRASS E. L. DOWTY V. J0KSIMOVIC P. C. RICCARDELLA D. I. ROBERTS H. N. WELLHOUSER Prepared under Contract AT(04-3)-633 for the San Francisco Operations Office U. S. Atomic Energy Commission GENERAL ATOMIC PROJECT 900 DATE PUBLISHED: OCTOBER 16,1974

@ ENERALATOMI GENERAL ATOMIC COMPANY l P.O. BOX 81608 l SAN DIEGO, CAllFORNIA 92138

l l

CONTENTS

1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . .... 1

2. TECHNICAL INVESTIGATION. . . . . . . . . . . . . . . . . . . . . 5 2.1. Metallurgical Aspects . ..... ............ 5 2.1.1. Location of Fractures. .. ............ 5 2.1.2. Pelton Wheel Materia'. . . ............ 5 2.1.3. Metallurgical Examination of Failures. . ..... 5 2.1.4. Diagnostic Tests . ................ 21 2.1.5. Conclusions. ....... .... ........ 29 2.1.6. Design Property Determination. .. ........ 35 2.1.7. Opinion of Dr. Paul C. Paris . ..... ..... 35 2.2. Operating History. . . . . . . . . . . . . . . . . . . . . 39 2.3. Experimental and Analytical Program. ...... ..... 42 2.3.1. Fractures in the Bucket Root Area. . ....... 42 2.3.2. Safety Margins After Incipient Fracture. ..... 46 2.3.3. Fractures in the Curvic Coupling Area. . . . . . . 63 2.3.4. Summary. . . . ... ............... 77 2.3.5. Consequences of Failure. .. ........... 78 2.4. Safety Considerations Associated With Reduced-Speed Water Operation of Circulators . ....,..... ... 78 2.5. Water Operation Speed Control. . . . . . . . . . . . . . . 91 2.6. Consultants and Advising Agencies. . ........... 91

3. CONCLUSIONS. . . . . . . . . . . . . .............. 92 3.1. Conclusions From Metallurgical Eraminations. . . ..... 92 3.2. Conclusions From Previous Operating History and Experimental Stress and Safety Analysis Work . . . . . . . 93 3.2.1. Incipient Fractures. . .............. 93 3.2.2. Localized Fractures in the Curvic Coupling Area. . 93 3.2.3. Safety Aspects . . . . ............ .. 94 REFERENCES . ............................ 95 111

FIGURES

1. Fort St. Vrain Pelton wheel. . . . . . . . . . . . . . . . . . 2

2. Fort St. Vrain circulator. . . . . . . . . . . . . . . . . . . 3

3. Fluorescent penetrant indications of Curvic cracking . .... 6

4. Processing history of Pelton wheel castings. ......... 8

5. Fracture face of Curvic crack in S/N 4 . . . . . . . . . . . . 9

^

6. Fracture face of Curvic crack on S/N 3 . . . . . . . . . . . . 10

7. Fracture f ace of Curvic crack in S/N 4 showing bright rim. . . 11

8. SEst photograph of fracture surface in Curvic of S/N 4. . . . . 13

9. Apparent fatigue striations on surface of Curvic crack . . . . 14

10. lietallographic appearance of curvic cracking . ........ 15

11. Meta 11ographic appearance of Curvic cracking . ........ 16

12. Typical grain size of castings . . . . . ........... 17

13. Detail of cracking from Curvic area. ............. 18

14. SEM photograph of surface of typical bucket crack. . . . . . . 19

15. SEM photographs of striations on bucket fracture face. .... 20

16. Fracture faces of Charpy and tensile test specimens. ..... 26

! 17. Comparison of the room-temperature fatigue properties of Pelton wheel casting (heavy section) and published data for Inconel 718 castings. . . . . . . . . . ....... 30

18. Fatigue dat for heavy-section material from Pelton wheel castings at 70*F . . . . . . . . . . . . . . . . . . . . 31 l 19. Fracture face of typical low-cycle fatigue specimen from Pelton wheel casting . . . . . . . . . . . .......... 32

20. SEM photograph showing detail of fracture surface of bucket fatigue tested to 21 x 106 cycles at Convair. . . . . . 33

21. Details of fracture surface. ................. 34

22. Effect of a notch on the fatigue properties of Pelton wheel material at 70*F . . . . . . . . . . . . . . . . . . . .

r 36 l 23. Effect of mean stress on the 70'F fatigue properties i

of the Pelton wheel material . . . . . . . . . . . . . . . . . 37

24. Goodman diagram for the Pelton wheel material. . . . . .... 38

25. Operating history of Pelton wheels S/N 3 and 4 . . . . . . . . .

41

26. Safety margin at 10,550 rpm. . . . . . . . . . . . . . .... 44 iv

FIGURES (Continued)

27. Bucket loading dynamics. . . . . . . . . . . . . . . . . . . . 45

28. Load amplification at each resonant speed. . . . . . . . . . . 47

29. Averaged load amplification. . ................48

30. Bucket bending stresses. ... . . . . . . . . . . . . . . . . 49

31. Safety margin at 9000 rpm. . .... ...... .......50

32. Safety margin at 7000 rpm. . . . . . . . . . . . . . . . . . . 51

33. Bucket fatigue test setup. . . . . . . . . . . . . . . . . . . 53

34. Fatigue crack growth testing Inconel 718 Pelton wheel castings . . . . .. . .... ....... . . . . . . 57 threshold vs "R" ratio . .

35. AK .... ...... .......59

36. Fatigue crack growth in the Pelton bucket. ... . . . . . . . 61

37. Curvic fillet stresses . . ......... . . . . . . . . . 65

38. Fillet stress caused by turbine disc strain. . . . . . . . . . 67

39. Shake table experiment . . . . . . . . . . . . . . . . . . . . 71

40. Finite-element model . . . . . . . . . . . . . . . . . . . . . 73

41. Poiton wheel stress distribution . . . . . . . . . . . . . . . 74

42. Pelton wheel critical crack size evaluation. . . . . . . . . . 76

43. Peak component temperatures following rapid depressurization . 85

44. FSV rapid depressurization at 105% power . . . . . . . . . . . 86

45. FSV heat removal and heat generation rates . . . . . . . . . . 87 f

46. FSV rapid depressurization at 105% power . . . . . . . . . . . 88

47. FSV pressurized cooldown from 105% power . . . . . . . . . . . 89 l

TABLES

1. Pelton wheel material. . . . . . . . . . . . . . . . . . . . . . 7

2. Summary and main features of cracking. . . . . . . . . . . . . 22

3. }!echanical properties of castings. . . . . . . . . . . . . . . . 24

4. Results of stress corrosion cracking tests . . . . . . . . . . 27

5. Thermal shock tests. . . . . . . . . . . . . . . . . . . . . . . 28 f

6. Summary of the Inconel 718 Pelton wheel operating hours. . . . 40

7. Bucket test results. . . . . . . . . . . . . . . . . . . . . . 54 v

f f

TABLES (Continued)

8. Fracture toughness testing of Inconel 718 Pelton wheel castings . . . . . . . . . . . . . . . . . . . . . . . . 56

9. Pelton wheel bucket analytical / experimental comparison . . . . 62

10. Maximum helium temperatures af ter loss of normal cooling . . . 80

11. Events requiring Pelton wheel operation. . . . . . . . . . . . 81

12. Worst-case accidents reanalyzed. . . . . . . . . . . . . . . . 82

13. Summary of safety limits . . . . . . . . . . . . . . . . . . . 83

14. FSV acceptance criteria for consequences of DRA-2. . . . . . . 90 vi i

u

1. INTRODUCTION During a fluorescent penetrant (Zyglo ZL-22) inspection of a cast Inconel 718 Pelton wheel removed from service at Fort St. Vrain (FSV),

4 cracks were discovered at the roots of the curvic coupling teeth and at the j bottom end of the Pelton bucket splitter. Six consecutive buckets out of twenty were found to have similar incipient cracks in the same location.

None of these cracks were visible to the naked eye. This casting was j serial number (S/N) 4. The Pelton wheel casting is shown in Fig. 1. A cutaway view of the circulator is shown in Fig. 2; the Pelton wheel can be seen on the right-hand end of the machina.

A Pelton casting (S/N 3) from the same heat, which was 'used at the Valmont test facility but not at FSV, was inspected by the same technique.

Similar cracks were found in the roots of the Curvic coupling teeth, but none were found in the buckets.

The mcting Curvic coupling of S/N 4 on the steam turbine disc (wrought 422 stainless) was checked in a similar manner and found to be without flaws. There are two other Curvic couplings in the machine which are also wrought 422 stainicas. These were also checked and found to be without flaws. In addition, ten prenuclear Pelton wheels used at FSV, two of which were also cast 718, were checked and found to be without flaws.

Castings S/N 3 and 4 were subjected to extensive metallurgical and engineering examination and testing,_ described in this report, and the following conclusions were reached:

1 t-

l

- -h d' ~~;

)

i u =..

~ -

W,' .

3

• - ;M'

~

?? J $

l -._

,f . .

,~

g p ,1 >

a:,

./ <

74K8081 Fig. 1. Fort St. Vrain Pelton wheel 2

WATER SEAL REVERSE FLEXIBLE COMPRESSOR THRUST TOR 0IDAL ROTOR BEARING WM INLET I TURBINE DUCTING

=r*"4"fLd/Tgh[td'@-= e bE$AP) / '%

y -

fgfyg "i

L -2n!"

SHUTDOWN' SEAL bNhO@k

^"""X .a

[ a j s STEAM LABYRINTH HELIUM 4 .

LABYRINTH s

~ .

o JOURNAL INSULATION x 'N '- i

^' JOURNAL SHIELDING k 8 EARING pp SHIELDING INSULATION CONE BEARING WATER INLET CHECK VALVE Fig. 2. Fort St. Vrain circulator

1

1. All cracking is attributed to high-cycle fatigue.

2. The evidence strongly suggests that the cracks in the Curvic coupling were formed at Valmont and caused by the abnormal steam conditions used at that facility.

3. The cracks in the Pelton wheel buckets (S/N 4) were formed at FSV and were caused by high-speed water operation, t

4. Fracture mechanics analytical methods supported by experimental testing show these cracks to be self-arresting at all the required operating conditions, and they will therefore not impair the function or safety of the circulators.

4

i i

2. TECHNICAL INVESTIGATION 2.1. METALLURGICAL ASPECTS 2.1.1. Location of Fractures Cracking was observed on the Pelton wheels in two locations:

1. On wheel 4, which was operated at Valmont and at FSV, cracking was observed in the Curvic coupling area (Fig. 3) and in the buckets.

2. On wheel 3, which was operated only at Valmont, cracking was observed only in the Curvic coupling area.

2.1.2. Pelton Wheel Material The Pelton wheels are precision castings of Inconel 718. Applicable specifications and heat treatment are shown in Table 1. The processing history of the Pelton wheels is illustrated in Fig. 4.

l 2.1.3. Metallurgical Examination of Failures 2.1.3.1. Curvic Coupling Cracking. The main features of the cracks in the Curvic couplings of wheels 3 and 4 were generally very similar.- The j fracture surfaces were coarsely faceted and showed clear evidence that cracking had propogated along paths. related to the grain structure of the material (Fig. 5). The crack faces on wheel 3 were completely discolored-(Fig. 6). On the other hand, the fracture surfaces on wheel 4, although discolored over the greater part of their surface area, also showed a bright outer rim (Fig. 7).

5

'-- +

. g ,- r., pyy p et

.

'j'. . . . ' -

v ,

3 4;;:

3-

-p - . 7 - .

s .- , c: ,

• c.1i.,

% j$ i 9 a k hE , ,$

gc, d i .p fi . s , L / .. y

,.a ,

h .4., .' ,

. z t] .

g

.1 ?,?,* g '

{

an ,  ;

M_. - ' .

~

F b: l

.s

- /..  ?,f.

-~ . :_ _

. .E

.n v

-:,,_.~ -

4 ~k .- 2hi;

'?

! (Q

. g :-;-e,7v? '

a

. }p ' :. '

. . . ,e . i r. r ', ',

' y . A;,.c L ,

, ~a

_* ,e l 7 '.11T7 491 l Fig. 3. Fluorescent penetrant indications of Curvic cracking i 1

6

r TABLE 1 PELTON WHEEL MATERIAL Inconel 718: Investment cast Specification: AMS 5383 Heat Treatment Homogenize: 2000*F, I to 2 hr, air cool Solution Treat: 1750* to 1800*F, I hr, air cool Age: 1325'F, 8 hr, furnace cool to 1150*F, 10 hr, air cool

~

.7,

PRECISION VACUUM CAST HOMOGENIZE + AGE AND TEST

+ 2 SETS OF TEST BARS SOLUTION ANNEAL ONE SET BARS 1 r INSPECT CASTINGS 1 r IF MECHANICALS AND INSPECTION ARE OK, SHIP 1 r ROUGH MACHINE CURVIC 1 r AGE CASTINGS + 2nd TEST BARS SET OF TEST BARS IF MECHANICALS OK FINAL GRIND CURVIC 1 r FINAL INSPECTION Fig. 4. Processing history of Pelton wheel castings 8

I l

9

(~

y

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

.f_

2,

g. -

?i

'j i ROOT OF

[b . ,, ; '^ ,

6 2 CURVic TOOTH

).  ;-

['

J bl

{;

j.} n t  :

- 3 t;-

f. . it i i, i f

v ' I )

I' d i' :

• WATER '

SLINGER .

a r- ,

o FRACTURE

g. :

PRODUCED

[;I - IN Opty

. ~; Q_ a f- CRACK

+ ~. a , _ ? :f ',=

f ,

., . ,,.. ,,.%:27

_ .w u s ,

f-' .

-J' f v+ .A+

[m 3;g- m -.. _

[

NA ~a-u , ,

74K8104C Fig 3. Fracture face of Curvic crack in S/N 4 9

1 1

l

. 1 1 l

a s . . m l

,e 3 - . .. ...,.

a r

%_ +

CURVIC TEETH

! l

.. ROOT d

.{

.1 1

~1

-)

1 y

./

):' )^\

~;j

~

WATER I  :

SLINGER <* ",' ..

1 r- '

?. . T',::n , ^ e' t' ' f-i t  ;

q

, _c q ;

t . . s f .ed l

t Q', '

if " 3

),.

/ w.u

:: . :ii1 li

- t. .

74K8106C Fig. 6. Fracture face of Curvic crack on S/N 3. This fracture is not fully typical since cracking in this case appears to have initiated from an original casting defect.

10

, A.

7,.{ :~- ._

w j A, f. l

' R, - .:;E .~

it --' t-

' n. .: ,.

<. , inl 1 -- ,- .

,-2 -

N -, '*

g $

l

.r- [ '_~ , rj'

2. '-j.

t .

__& .:- -r; .

74K8103C Fig. 7. Fracture face of Curvic crack in S/N 4 showing bright rim (arrows) 11

Scanning electron microscope examination of these surfaces confirmed their heavily faceted nature (Fig. 8) and revealed what were believed to be fatigue striations (Fig. 9). However, the fact that the fracture surfaces were heavily faceted and contained numerous traces of the underlying microstructure made it difficult to positively attribute these striations to fatigue.

The metallographic features of the cracks in the Curvic areas are shown in Figs. 10 and 11. As indicated, in some cases the cracks were very straight and unbranched, whereas in other cases the cracks wandered considerably. However, it should be noted that although the cracks frequently followed paths generally aligned in the dendrite growth directions, the cracks were not intergranular in character. Figures 12 and 13 also illustrate the very coarsely segregated microstructure of these castings. It should also be noted that the castings were extremely coarse-grained (Fig. 12) and contained numerous intermetallic phases, including dispersed Lave's phase and Ni 3 Cb needles. However, it was notable that crack propagation paths did not preferentially follow any of'these intermetallic phases (Fig. 13).

2.1.3.2. Bucket Cracking. The fracture surfaces of the bucket cracks on wheel 4 appeared morphologically similar to those observed in the Curvic area. In particular, the fracture surfaces were coarsely faceted (Fig.

14). However, these surfaces were bright. Scanning microscope examination of these surfaces again revealed evidence of apparent fatigue striations (Fig. 15). The complexity of the fracture morphology coupled with the multiplicity of striations made it difficult to unequivocally attribute these striations to fatigue. [However, the close similarity with the test- t induced fatigue fractures (see Section 2.3.2.1) clearly indicates that the bucket cracking was caused by fatigue.]

}

The metallographic features of the bucket cracks were generally similar to those of the Curvic cracks.

1 12

ROOT OF CURVIC TOOTH

~

~

l 3*gh: ~ -- v* .

N7e- -

u .

4 '- 2

: }

p a

f

{

i l .

?

FRACTURE CAUSED IN

^

OPENING

' , . ,s; j E,,8 74K8102 %10X Fig. 8. SDI photograph of f racture surface in Curvic of s/n 4 13

74K8100 2500X Fig. 9. Apparent fatigue striations on surface of Curvic crack

(

)

14

l l

i 7 ,. . . .

• 9. e,t*.n. o,.1 .*

4. N,[f *

, t: )

J.,4 dia, W. .9 0. 3 . ' 'l i * ' * , i a

,r,.;'*

%.. . s. r,

' '"uD , d,;d. *<, 'mYr,,p,3, . -

~4 .

• • 4- _' ,~{.Q?.. **#* [', s* .'. *.."4 L' ."> ., , n r.' $. ' **W. (A

.g

• i ;R 3,\'.?,.,s.yy.>* . *s. ..mg,.a rr,

.$. . ..*\. vy,,a '.f. .t... f pQ p ,. * . :.; * .

1g, .6

.:., ( ge.v -

., r S

.- .m

& ~ e. - .. , t ,,,%g.. ,,.. .,

~;- ,. ; . %t p. ;,i,

.a y '. . + .

\ s4 +.

" g . #4 s1 T. , 4.yv:

e

~a g,. t o.

~

. -w yJa .n w..

h..

F . .,, .,. . . . y' : y,. .r. . ~g)m., y y>:

g.3 y *.(,.

. .. , . .y je yv - ;r 3. *~.o* '

,1 s s  ;.,1 .

e .. \. 4,4

,:s.,:. ,;n. .: n. . y. m,.aI.,:e

• 1 .

~

.r m'

1 .:m w 7 .g, y..

. .s.u o. ,,,.,n

+,-

, ; ,ne v:,.a.;,.

.],y..c ef',r,(( fg;P.t.,4 V 2.o *e*' D'.y,;s ,8i7j f *.,e j *,,,fq

• ,; ... , i,.

g/

• a ,.r* K o*, * ". . ;'*,f .sr '. 1i, .~f *., .l*'s- s. W*4 -

(s.

. ; .... Fw. '.+,y 6

3 e i . . { ., r . . . . .. . t / .e,. -V.* ,f* & ,' 'g s ., e.g4

• ** j,s .,

,'s. y' *

. ,.-,, i . h.. .a. .: . ,; o

• y s.. 6  ; p.

• a

• . y\ ' . .e t. t" ,s.

. , y. - s.

~ ,,f jt -) . .,'.b: , .v.'

,' i, ..J,.

• ,t y'.. .. , .

. . . y $., i. . . l.h.*. .e.s * ,. 3'

' f,.e S.ep, *t . ' .'. %

' .s = g ,t c.I.

1 '

a. m  ; ..'*% .

. .w* -s . . **'- * '

,p. ,%.s ...p f,*,g,,)..?p,.,..e.7,.**j,,g./,.d**

, v .

% . . so .g N .p

--( l . , .

.t6 ... ..$

t

? .>

i j , '.. i s

• Ce -

. , m .> .

3 .* - Q

,,y f. g.,. * ';p *.

. ,,- ,. n , m. ,

.yi,#, .; . .,.

• 5 r.* b < . s. g t .

a'a

• U- .

)'.. . i.f 4 2. s t ,.,

e

't" p ; ,,l,, % <84

'I. 1

<4,- 4,

• • J

,,( o ..w'. (s

.: y, g

~

...,s. <. .

• ', ., f.'t g g* y . p.,*:A t '( tl.- . 3. ,~ . e. h 3, e -

e M..ji 3;

. o . . y.

< ,' .t L(' ", ,.s..t _

., L ,:.s . , ,',, ,.: , e ,,

. >:. k. .

. i, : J.> ,, ;,1, y ., '. ;-Z .i t. 4.Wg, g,'s,q, . ',y ,C, ,a,

. s ...

. y .. s. g

.e q .. .  %> . . r x..

-.a,.n< r e p., ,.-~.l ,..<.. s .4.M.m.. .:n.

~

.l

<t .. ,,m

.- . s.n w

a, .. . y.. ,,. . m. >..;.,...v a .w. s .. ..<

n. ,

- 1.e

, .. .rv /.. i .+.+ , 5

., p . .<4 . -,

n . l + e e ,<,-

.t a

..,e

f. e.

. f. g* ,t*4 v** )

• h ,M,e

, ~ i 's . . *

• as *", ?

M ' ' . . '* l ;' ' 8 ,.' . '} .I>> . .,i

.{~ . . , V ,y..., . ]

• V3,ak{'[.[.*v,

. ie g. ,

ip '

? . '. .g e4 g. . -,j ., e' +.l.'r  ;,* J . p

\/ X , , , .

f

. v .,. ," 4. .

• . . . ') , ..rg ,,e, g ,J'g s*jg ,. .

• ." . , . 9 .. . * '4 '44 -

+4 w *g., e e ! *. '-t

• g Gad .# .

' , , _ _(,;.ss,> e*, ..}. . s?. . .-e,. ,i\ *':.T .. .*\..'

' , ' e / U,. ,.., * * '.

P ') c , '.

.'r l , . . . f."...<..,t 8 ';' ,. . . . fo. ** '*t D.

. e J. :M*

s... , ,. }s

. t .-r . '%' 1 . s- *. r.

. .s . *

,. 4

• - . - e ' s t.~,3. $..%. >;t-

. w s,; ,. .

. <ie. , .d. \.,==,4 ' .cay# .)s . s.)*'I'.3

..'.s l

'

• ds, l' 'J' ' ' [ 4 ,.',s.A, *. .. <v . e . 5.<* *

• V ;/w *.s.

, -( -

.N>< ,. , 1 * ., J...

.. ,.s.<,- >.< ,.

Y9(,.c

..'"..,. A

~' r 4 .;.')/ l N' *, N *

+ . I .

% . ,.'  ;., t ,, p . $. h.~ ;., , J. ,4

. . ~. I, . o. ..

( ),,,. . .~x..-..

/. . .. .

,,,e 1 , ',

• .- f V.v. *4*' .

• i,\ )vt 8. ;s

• jf

, * ' fg .,

', / , f..

^~. r 'p*.;, *,. et.,?",v.4.%,. .-

, . ;.,,f ,,g 4..;;,. j-[g ,,.,a t 74K8088 30X Fig. 10. Metallo<;raphic appearance of Curvic cracking 15

1 1

i I

I i

1

.';.c t. ~:f,ds s e,= -+

gtt m p * ' e- ".+ .

pa ,

.h.*' (=<4.6

a> .\e 1

.d,+y.-

, r; . .v

> +

J e*

c ,.g,

. 3,;, I r,u}

$ 9a s. a .'f:.t.,3.g A;g t$L . v.q .

.? .

e.L.

,y;;h (Y

h . 4 .3. r .$4:h.?y

.k

'i *

.(

I rw . 7* ' ;

- . y ,4 M, y

, jue" 'y .

i

"'* *' 3 * '/h Ct t.. w h 4.' s-2. . ' .- ,

3 W-; ,a - g b!.'..tk<

+ .., ;. .,4 R Qf}y.? c,W,,"* R...K>.g mg"s

.4. ,,.c.}i. A. . 0 .. ,j:

s

. K.,6.

o , ..,,','

f- ;

V .t S: .g .a 3 ... . p - ..

9 s v , w .G.3 . . . ' ;J. . .,.g*

p'- .. . *.-i....

. p,,",f (y, ;p Q'*?.Q,

; Q 4 2 -J .*%g,.of

e. &. 4

• ,, , ,? .,.. A.J.,,

.- '-3 *.e ma

• var a. ,,, .y'.o* p:rp*.w' q.

.,e Aa'

-". .e. ., s ~a.

. , ^ y e..,

% g.

, ,e 9.

9 i . }- ..> :. ,4 . *

.i s~<r. 9q

. , ,g > ga.,gr .n,7 7-

.gy M

,, -v, .

• . ; m . m.

d g. W; O. ,' ..p. -.- ',: ' * . .d* u. . .a,%-

g, . ,1. .4119asseen r 4

.g }; ,, 1t."ia..;. .= #

f Pd,++s. , .:. < f: .u. .,,,,, ,r. a .

N.

-- .a ..,v s

'4 Y % .,*&,HM . . r ~

fhM .l-l D V ., . '*f. 0l v, p,

y. .d:..z  ;,.. .s , . w . w:y nam: m n i hDD= , .,, h..~ .;' $ bhN!hk,N.#,$$h;a Ao,,, . . 4. m

< s.J '% A e...*..e.1.%-

w av a . . - %

y

• 4.w,e $;

, Ar ny:U.a u C .'Y o p *17.k.5* *

• R r*'t ,f[ f

,,...,.W..n U

' ^ a b y%t6he.-.*

..,a 8 s

., . w-\ .,

g...

'h f.y.Qen:,

.c cv. rh. ,- *.,'~).. d :}..}. ;& :s; n .% t p.aa M sg _ 7 . .., *..' A.. . w* * - b s.F~A._r' .,,.,cg < v ,._

y v

+* g

.5-e#..* y 2.s t.A,. . ~)

. +

>A sm s ( ** 2

• s1 L . '. ; - * %

y ,4 : 6-

'"c p ..

a ..* q.

,. .ia..y ;y~ .. c.!]>r}g. f . -;4 *~.'rd ~C .}.4

..s1ef.

'~ $}r S.. . * .: f* ;g .,* L -(.zVxQ

• • • • • 6 a. '

i' 'U .

• y%..* ; ** f. .AU-W .T .

L.s>*).A w.s 1

~ t . . af.a.

..n.

Q,..? -

s , ..Cf d,, ,.LQ :p. . .c g l~,

,. -,'  ?.:, ,+.& W r..'

J .,+,, 42N 4.

. . . * -. -- n.

c .4, gy .., .2 C.L.t e V g; q d. {.,.4 ,, . h.

hmmt - ,e4 , .

',,y ,f.Q,1,

,7,s..% e ,jn gm W .*g

g. <s 4.

<m v' tp. . J.* .... .. .. .. -i, . %*,. h '.:e/ <*fd y . ... .% .

qg  :

2Na..,&.

y

f . . .J Ay vC f 'c'

1:;,, ; s ,e g V

-w-M A.,g

%.- , 4,

,v.z. n -

g * * - - ~ , . >* * . . . . .

. {".}~ l C.'

N-d /;N *:.h .i;,;k'][)(*4 u,.

e- . .w t -

~ .* $ h '",'*,'..w.  %.A e t ' \ *,[3 vh>[i 4, t <%.{

t ,; , s. . . .r. .

.M

= >#

j . +4 sp.c E .. .. //.4.}. p S )4 Qa

,, t Eff 'M

.- > A +u, fo . . ~- - - .

g ..d .t.y g 7 _'.T,, - , .;

. p:gn

..p . ung~ 3:

n x ,2 N. w,;

3 .

$,i, x - 9

-g,.

• • *1 U y , ^ ' ,,, , .

4m. ..

'

• i g .j wp. ~ . . , "'t x,wq. g.%g >m.. ~./.'arv== - . 4, ) $f @f~',y.q'MI., fY -

.. t $t + g+11f : f.,-4--

h* m. _. $ : a,. L.. pe w -.

.: 6 p ., .

w.-.%

2m:.md V%y

-.4 M*,y ?

s

. w;;* --..+

f7';~ * ')**

Elsw a, q pr .. . .' t j , s 1

-**v - k) f *"*" . i 1  :,.mns$

w

% s,

> Y s(n'r:;,r ; , , ...V% '

l i

f )?v ,,%e m a L, k ..-...+..-g .

a*.L..L e*

\

9- .r *V M as a . 1am:n ..

K'~c'*dr S P'-W.4 '

74 KM 03 7 20X Fig. 11. Metallographic appearance of Curvic cracking 16

I 9

f' *

- ..: i4 ,,; *, '

{ ,

i il . AE.'.

i i ,

l  : :

L Oj_g.

\

'g.  !

I

. A. ; t;

( '

-;i , . ,

l

!! % g g i I $ I. .

_p, X L - , J u . I

• - t.

} l

's s .g  ;

W N '

n[.:..I4L  % g a, b ' , ' ' '. ':st- _'..,

• *'s. ' . ! . *. ,. ','..

b; , ','p w. y - - -

';__ 'f ', ,,[ .

?

- . ; d, ,

, C S : .';

.. r C ' ,'l. : , , ,".-. :~ .

1 r

?,  ! -

3

,sa i

ki! E i v:

5 y

7*. . s * '

L ' .- <

+1 l

1

-L 74K8095 Fig. 12. Typical grain size of castings 17

n e ,. .,- - ,... r ,, ~ ,,- ,.- ,. g _

.w .g-. ..,,,.,-.,-n

. ~

,g

. c. . ,

4

. s.

j

^

=*

1 g * .~.6 ..

4

. g . - W.

. .4

,1, r . . ,

s m.s ,

v c

- .\. g ., er. .

e m t- ,, .

t

• j d

a 7 V ,.s r.4 - ' -r-  ;

~-

, ;y j e -.

q . g. 7 . . 7 .

. y 3 x  ; .- ,

g. , . N. . n.

, .4 f N - j ,; ,

' , ' f. C  %; .

I

.s.':. ..

. .i

=~ .;

t- . ,'. , y- .3

, .g* g. .-

... r - .,  ;. [.

g , 4.- p' O - *

, - j e

e > . +

s' . -

. ;e s

..q,F --..  %- . *,. , ,9 I .} .}

-8

+

-:%+., < ,. .'.t;

(. <

m -

. K ,. .

(, . ' q[ ' '

i n j,Q

• -_,o i -y

'A

1. . '.'S . ,.

f:_ ,

+ ,Faj , g4 ' 'Q g f t -; p.

L 4 O , -,

n:rc - ', ; j ? , ,~ m.Qge Q . s G

(: . y. ,

jf. 1 g

g:  := ,

,s . ,  :.r -e ;. j c- . .p e.

.:.- -. . ra - r

/,

3 4:

.> j - ,. ,

.s .

,F .

goo

.. ' +-

,# 2+,

, ; 1 . .: ~ , ,

7 ;- ,, ,

3].

g  ?.r . - .- -

g f.

,

....s E. ,

.( =y. n?i

<:+4 l- .+ - . , . .

.w .<.,4:- .

s-

• s,.,.,

4 A, L e -

-"op?g v ~

v.

i -

. k r,3 a. Qf b dj t

.,.'..m h ,h mf'

.n

..p, , .,

.-t ww w L- .

74K8090 750X 1

Fig. 13. Detail of cracking from Curvic area 1

1 18

\

I I \ s. .

. ., m

.c i ,

i

-= s

k. . T ,. 2

. LABORATORY-INDUCED

. FRACTURE N e-

~

s , .

o . r

. /

. . \ .

f _ '% , ,

t

/

- ~Y ' ^

l

..\ g

\

, '/',

m ..

i

/f.

. .< g.

6 Ks

~ '

, .,\ \ (. 'y:, V~-

, ,. ,s /. .k i, Al J T y Kh,

• f.. '(,]f /

l Y , ' ; ,~ ,

"welp ' ' *Waw, 3 Sg K ,.,e u

4 l

c yh. .. . )

l _

i .

/ ..

r . .

, . k--

l ,

~

v 9

\ ,,

Y k. .E

,i i

g'3 1,

74K8097 %25X Fig. 14. SDI photograph of surface of typical bucket crack 19

)

Fig. 15. SUI photographs of striations on bucket fracture face l

20

i 2.1.3.3. Summary. The main features of the cracks are smnmarized in

Table 2.

2.1.4. Diagnostic Tests 2.1.4.1. Cencral. On the basis of the foregoing, it was concluded that the fractures were most probably the result of high-cycle fatigue.

Ilowever, the influence of the complex metallurgical structure of the castings on the fracture faces made it difficult to confirm the presence of fatigue striations with absolute confidence. Accordingly, it was not possible to completely eliminate consideration of other mechanisms . 4 which cracking could have occurred, such as

1. Propagation of original casting defects.

2. Mechanical overloading.

i

3. Thermal shock.

4. Stress corrosion cracking.

5. Low-cycle fatigue.

Therefore, a series of diagnostic tests was performed to establish the possible relevance of these mechanisms.

2.1.4.2. Original Casting Defects. The detailed manufacturing and inspection records for castings S/N 3 and 4 were examined, and it was

, concluded that it is almost inconceivable that the Pelton wheels' contained the observed cracks prior to Valmont testing. Five spare castings were removed from storage, examined,'and found to be free of surface cracks. In addition, the coincidence of the Curvic cracks with the teeth roots 21

- .-. . _- . ... _ . - - .. . - . . - - _. _=.

l f

i l

TABLE 2 SU!CIARY AND MAIN FEATURES OF CRACKING l'

Unit Crack Location Main Featutes S/N 3 (run at Valmont Curvi Dark-colored, heavily faceted only) fractures I Curvic l S/N 4 (run at Valmont  !!ost of fracture faces dark and FSV) colored; however, innermost edge

j of cracks is bright. Faceted k identically to S/N 3.

I Bucket Bright fracture surfaces, faceted j identically to Curvic cracks.

,l i

4 22

presents a strong case against the cracks being original defects since the Curvic teeth were machined into the casting after casting and heat treatment had been completed. Based on this information and in conjunction with metallographic evidence, it was concluded that the in-service cracks were not associated with original casting defacts.

2.1.4.3. Mechanical Properties. The Pelton wheel castings were specified to Aerospace Material Specification (AMS) 5383, which calls for separately cast specimens. These separately cast tensile test specimens were tested upon receipt of the original esstings and the completion of the heat treatment. Additional tensile and impact test specimens were machined from thin (water slinger) and thick (cone) sections of Pelton wheels 3 and

4. The results obtained are summarized in Table 3. As indicated, the separately cast test bars produced with these castings and the test specimens from the thin section exhibitad properties that exceeded minimum requirements. The tensile and yield strengths of the specimens from the heavier sections of the castings fell below minimum specified values. This does not mean that the castings were out of specification. A review of these tensile test results with specialists familiar with Inconel 718 casting confirmed that the observed disparity between the properties of relatively heavy _ections and those of separately cast test bars is not uncommon. Therefore, the 718 Pelton wheel castings were judged to be normal and in accordance with AMS 5383.

To confirm that the Pelton wheel castings had been propcrly heat treated, sections from the heavy portion of one of the castings were subjected to a complete reheat treatment and mechanically tested. As indicated in Table 3, this reheat treatment had relatively little effect upon mechanical properties. In addition, examination of the microstructure of the material before and after' heat treatment showed no significant change. These results confirmed that the Pelton wheel castings had been corectly heat treated.

23

TABLE 3 MECHANICAL PROPERTIES OF CASTINGS CHARPY V UTS (ksi) YS (ksi)  % Elongation (Room Temperature)

Separate test bars made with castings 154 116 22 Tests from castings Thin sections 125 to 139 101 to 113 11 to 17 Heavy sections. 115 to 119 88 to 97 7 to 10 14 ft-lb Reheat treated heavy so.ctie.1 from 124 104 12 5 casting Minimum specified values on separately. 125 110 5 cast bar (AMS 5383)

Note: It was also observed that fractures on all these mechanical test specimens have no resemblance to the service failures.

It should be noted that the fractures observed on all tensile and impact test specimens (Fig. 16) bore no similarity to those produced in-service on the Pelton wheels. It was therefore concluded that the in-I service cracks did not result from simple mechanical overloading or impact. '

2.1.4.4. Stress Corrosion and Contamination Testing. Extensive analyses were performed using microprobe and Auger spectroscopy technioues in an attempt to identify possible contaminants present on fracture surfaces. The results of these analyses showed that contamination levels were too low to be able to attribute the fractures to stress corrosion and confirmed that the dark-colored material on the fracture surfaces in the Curvic area was an oxide film approximately 2 microns thick. An oxide film <

was also observed on the fracture surfaces in the bucket area, although in this case, the film was relatively thin (1/2 to 1 micron).

In addition, stress corrosion cracking tests were performed on specimens cut from the Pelton wheel casting. The results of these tests are shown in Table 4. As indicated, the only cracks produced occurred in hot concentrated caustic, and these cracks were entirely intergranular and bore no morphological similarity to the in-service failures observed in the Pelton wheels.

It was therefore concluded that stress corrosion cracking was not likely to have been a factor in these failures.

2.1.4.5. Thermal Shock Testing. To evaluate the.effect of severe thermal shock, individual buckets and a complete Pelton wheel were subjected to a number of thermal transients of varying severity. As indicated in Table 5, these tests failed to cause any cracking.

Thus, it was concluded that the Pelton wheel failures were not induced by thermal shock.

25

______ _ _ . . ~ _ __ , . - . . -

l I

l l

l s -

?

r. '~ % .; . -

,,3 . $

1,

's t .

'g ,' $

(a) 4X m, rpyv, 4 ,. ,,,. s q;g .3,

, >.3, , ,

l l

(b) 4X Fig. 16. Fracture faces 6T Charpy and tensile test specimens: fa) room-temperature tensile specimen; (b) room-temperature Charpy V-notch specimen i

4 1

l 1

26 1

1

I l

l 1

TABLE 4 RESULTS OF STRESS CORROSION CRACKING TESTS l Environment Time Results 90% Na Oli, 350*C 3-1/2 days Intergranular cracking, unlike fractures on Pelton Wheels MgCl , 154*C 2 weeks No cracking 2

Molykote G, 350*C 2 weeks No cracking Results of Surface Contamination Studies Environment Results Curvic fractures %2-pm-thick oxide, no signifi-cant contamination Bucket fractures N1/2- to 1-pm-thick oxide, no-significant contamination 27

TABLE 5 THERMAL SHOCK TESTS Buckets Heated to 550' (four times),1100* (once),

and 1600'F and quenched in cold water; no cracking Complete wheel (unloaded) Heated to 550*F and quenched in cold water four times; no cracking 28

2.1.4.6. Fatigue. Axial push-pull fatigue tests were performed on cylindrical specimens machined from the heavy sections of the two failed castings (S/N 3 and 4) and a spare casting (S/N 1) from the same heat.

, Initial results indicated that the fatigue properties of these castings were generally similar to published data on cast 718 (Fig 17). The tests performed extended from the strain-controlled to the load-controlled regime (Fig. 18).

Examination of specimens fractured in the low-cycle (strain-controlled) tests showed that the fracture surfaces were quite dissimilar in appearance to the in-service failures on Pelton wheels (Fig.19). This tended to eliminate low-cycle fatigue as a relevant failure mechanism. On the other hand, the fractures produced at the high-cycle end showed evidence of faceting and were generally morphologically more similar to the in-service fractures.

In addition, examination of the fracture surfaces of bucket specimens fatigue-tested at Convair (see Section 2.3.2.1 and Figs. 20 and 21) and the crack propagation rate specimens tested at Del Research Laboratories showed that specimens cracked in the very-high-cycle regime (106 ) exhibited fracture faces very similar to those observed on the in-service cracks in the Pelton wheels.

2.1.5. Conclusions On the basis of the foregoing metallurigical examinations, the  ;

following conclusions were derived:

f

1. Based on the oxidation of the fracturc surfaces, sit is concluded that the cracking in the Curvic area on S/N 3 and 4 wheels was initiated during operation at Valmont.

29

t 100 4

't 80 -

A  ;

i e

~

O U e

- 60 -

o A

= e a: A

.n

• e

e. e Q 40 --

'8 5 i"

4 20 -

e PUBLISHED DATA A PELTON WHEEL DATA t

I l 10 10 5 ;g 6 10 7

CYCLE LIFE Fig. 17. Comparison of the room-temperature fatigue. properties of Pelton wheel casting (heavy'section) and published data for Inconel 718 castings

_. -  ;- W-' - - - -

6 _

g _

j _

F _

0 7

t a

s g

n i

t s

5 a i 0 1

c l

e e

h w

n o

t l

e P

m o

r f

E l

i 0 1 F a I i L r e

C t I

a L m C

Y n C o i

t c

L O

e R L s O -

T R y N

O T v C N a O

C 3 e N

I i 0 1

h A D r R A o T O S L f a

1. A t a

d e

u g

i t

a F

2 .

t 0 8 1

1 e g _

3

- 0 i _

3 2 1 0 5 F _

0 0 0 0 1 1 1 1 O7!J $u; ED$$4 U

s '

f??{eh?$ v. . $$ s"h?h.k-

.. h:.,,

..:lY$ .  ! b.

^'

6 k...gy.$$%.g-+

=- ., * $.;. hW . - k. . ..?

-o  : :+a' .-;?.' #',2

..N

-p ..

.' ', ,.h, ,+,

s ;N:

eg!. ..,', p,4Y V,..,m,

g. ,

?.g.,. . '. ;.g* ~

,, ,' o_ 3y[;.. ,l. . w g.y l -, -,- .,

,'~

a. a

' b

~ , .

. 2 s .: .

<. J.

4X Fig. 19. F acture face of typical low-cycle fatigue specimen from Pelton wheel casting 32

\

r s g

0

ae,g%ffN '

. * , 3

. 3

.* 2

.qa s

$;. c.4 ak' ,

M,;4. y+' ~

vf,,, '

4wnMm {

• s* *

$s ,g #. 1 N <

,\

k g5 s.w* L t p:"  :

r e Ayy:w';R sf'- ug- 3M ?',3;A.'

.r u

s n.

M',D ' -

ti j  ;, t t,

\',(y,i..

s j 9~.:.- .

x 'a ...

y

.-r e&; ' ;~

y

~

.n .s mm

4,,9 x _8a:w.,,*{:

r..;0'y

.M' c. q . ,Q , ,.,,$

. ., c. ' '

Y qA^' . ,N MM;am';i. * '

i

g. ,

c a

1 (( +

d

[b#_v%Q-a.[ c 75'. . d *g u

@,,yb,. wy:c::.1 r

-

• s

+

• C.* '+

gi ;6 okr:~.',,y:. 3

." t i,

,,.f',ij..

i

' - f'? , "

a,;t w :*

h '~

.e

• f,

'* * , i . Y'l9;t;y v

gsg;y{.;d,.en=j',.f .

'c. 9;4***m'v. JyM.

%..G{ ': t 6

4

'. i a

w?.!e.f~0@7

,1 ,

- t

, - 3 .$

l 3

e R  ; sy.,ph&,;teh,

c *t gv E

An7M@$ W,

.^

y#(y'ln;h.

s. i .

do h

3 [c ,_..J- , rT i. ,.

g M$.o + ~

. 1 s -

l R:. T.

y.

'g, ~

- 1

. $p:g,. 4  ?.Qw.4

t

t. .

a

- 3 4,,

(4'( p c . A,. y .

. l, ~_.. :;e f' : ie s

• vg.,';e.3f '* , %

l

'r ,

c S c l'

,u.

a.-

. 7#-

. ,>.. .,pz 4. Q ;* +,h 9

6

,;: ,, . ,. yAc .

0

. 7Q&;[

) -

i

.;t -

,o 2

, ?;; .. .

.y

.j. t s ,v 7,Q,K. .; W '

k  ;

( 3 3 i s

v O

I

": \

l l l ,

l l

l l

i

, i l

M".

% 4.: ?Y,$

1000X s

. / ' -

..A h l'~ *

^

- =

, ./ ;

'% .< ,s

r .

u. / ,

74K8050 10,000X Fig. 21. Details of fracture surface 34

l

)

2. Some limited propagation of the Curvic cracking occurred during operation of the S/N 4 wheel at FSV.

, 3. Owing to the lack of heavy oxidation of the fracture surfaces and because it was confined to the S/N 4 wheel, which had been operated at FSV as well as Valmont, it is assumed the bucket cracking took place at FSV.

4. All the cracking on both wheels appeared similar in character and was concluded to be caused by high-cycle fatigue.

5. The mechanical properties of the material were typical and within specification.

2.1.6. Design Property Determination As a result of these conclusions, further attention was directed toward more accurate characterization of the fatigue properties of these castings. In addition, work was undertaken to determine the properties of the castings relevant to the fracture mechanics analysis of residual life.

l (The latter data, which comprised K id determination and crack propagation rate studies, are described in detail in Section 2.3.2.2..

The additional fatigue properties determined on the castings are summarized in Figs. 22 and 23. As shown, the material did not exhibit an abnormal notch sensitivity; however, it was noted that the fatigue

]

properties of the material were unusually strongly influenced by mean k stress. This is summarized in Fig. 24.

2.1.7. Opinion of Dr. Paul C. Paris All the physical evidence and related documentation were made available to Dr. Paul C. Paris of Del Research Laboratories for the purpose of obtaining an independent assessment of the probable cause of the fractures.

35 i-_., _ .

- - ' - ' - i' I -

3 10 AVERAGE CURVE FOR PLAIN SPECIMENS O 10 2 _

U

,r 0

u G

E C

I u

• 5 2 1

< 10 -

1. NOTCHED DATA (K = 3,8) FREQUENCY (2 cps) 0 e i ,

10 2 3 5 6 10 10 10 10 10 CYCLES TO FAILURE Fig. 22. Effect of a notch on the fatigue properties of Pelton wheel material at 70*F w ~_ _ _ _ - _ _ _ _ _ _ - _ _ _ _ _ _ .

8 0

' ~ ~

1

- l a

i

%NN  %

t a

l r

e m

e e

h 7

0 w

%N I

1 n

o t

l

' ~ P e

e h

t f

o a s e

i t

6 r 1

1 0 e p

o r

E P R

U e L

u g

AI F i t

S S O T

a f

E R S E F A T S L

• C 0 N y 7 A c E

M 1 5

o e t h I t S

K n 0

o 6 s s

I e S r E t R

T S s S

S E n N T R a A S e E m M N A f O

R E

M 1

4 o o E t _

Z t _

I S c K e M' 0 3

f f

E 3

2 g

i p

' - 3 2 1 0

0 0 1 10 1 0 0 1

1 Cyc n$-

s 55 r1$54 u"

\ i t

90 80 -

i 70 -

60 -

2 0

3 sn 504 g' 105 cyctcs E

w 5 40 fc 4

30 -

6 10 CYCLES 20 N

\

N N 107CYCLCS N

10 N --- -._.

N ,

10 CYCLES -------

o

' ' , , i ' , i P o -2o 30 " $ g, go 100 130 - 120 nga stress (KS1) 7 0 Fig. 24. .

Goodman diagram for.the Pelton wheel material. (Note that the 10 and 10 curves are obtained by extrapolation using the conservative assumption that this material does not exhibit an endurance limit.)

Dr. Paris is in complete agreement with all the conclusions given above and in addition believes the following: -

1. The cracks in the Curvic coupling area were driven by a vibratory disturbance emanating from the steam turbine disc.

2. Torsional excitation from either the steam turbine disc or the Pelton wheel can specifically be eliminated as a cause of the crack propagation in the Curvic coupling area.

3. For the purpose of establishing permissible operating stress limits, the basic material should be considered as flawed, whether or not it is cracked.

2.2. OPERATING HISTORY The operating history of all the Inconel 718 Pelton wheels is summarized in Table 6. In addition to S/N 3 and 4, in which fractures were found, the tabulation includes the operating times for the four Pelton wheels presently installed in the plant and the two Inconel 718 prenuclear (large) Pelton wheels. These were used during the water-operation-only phase of the program. The earlier prenuclear wheels were made from cast 17-4PH material. These are not shown in the tabulation.

Figure 25 summarizes the operating history of the S/N 3 and 4 wheels on a comparative basis. Wheel 3 was installed in the prototype circulator used during the development and qualification program conducted at the

) Valmont test facility. Wheel 4, after operation at Valmont, was used during the plant check-out program at FSV.

39

TABLE 6

SUMMARY

OF THE INCONEL 718 PELTON WHEEL OPERATING HOURS Valmont Fort St. Vrain Water Steam Water Steam 5000 5000 5000 5000

<- > < to > < to- > < to >

-to Serial Heat 5000 7500 7500 5000 7500 7500 5000 7500 7500 5000 7500 7500 No.- No. RPM RPM RPM RPM RPM . RPM RPM RPM RPM RPM RPM RPM Remarks

-- - - - - - - 13 28 110 0 0 0 Prenuclear 11

'2 0 148 300 0 0 0 Prenuclear 3 VH164 100 .20 15 100 350 80 - - -- -- -- -

Cracked curvic 4 VH164 31 22 25 73 3 40 58 0 1/2 81 2 0 Cracked curvic o Cracked buckets 5: VH164 3 3 2 46 -3 16 0 0 0 0 0 0 In "C" circulato.

0 0 0 "D" circulator

'7-VH215 13 4 7 93 24 110 0 0 0 In 8 VH118 6. 15 -5 45 8 44 111 16 5 73 17. O In "B" circulator 9 VH118. - - - - - -- 0 0 0 0 ~0 0 In "A" circulator Additional information:

S/N 3 underwent 15 thermal shock cycles at-Valmont.

S/N 4 was run at 37% overspeed'(13,050 RPM) with steam at Valmont.

S/N -7 was run at 34% overspeed (12,750 RPM) with steam at Valmont.

S/N 4 was run at 10,000 RPM for 1/2 hr on water at FSV.

S/N 8 was run ati- 10,000 RPM for 3 hr on water at FSV. ,

I I

S/N 3 HISTORY AT VALMONT kk% S/N 4 HISTORY AT VALMONT

[

b(((((((( S/N 4 HISTORY AT FSV

< 5000 m f///////////////////A 5000 AT VALMONT, S/N 3, 120 HR ABOVE SELF-TURBINE SPEED PELTON TO AT VALMCNT, S/N 4, 80 HR ABOVE SELF-TURBINE SPEED WHEEL 7000 AT FSV, S/N 4, 58 HR ABOVE SELF-TURBINE SPEED SPEEDS 7000 (RPM) TO 9000 h

> 9000 l (1/2 HOUR AT 10,200 RPM)

I I

<5000 --==

STEAM TURBINE 5000 I P S p 700l AT VALMONT, S/N 3, 525 HR ABOVE SELF-TURBINE SPEED AT VALMONT, S/N 4, 115 HR AB0VE SELF-TURBINE SPEED

> 7500 my AT FSV, S/N 4, 81 HR ABOVE SELF-TURBINE SPEED I I I i 1 1 1 1 0 50 100 150 200 250 300 350 HOURS OF OPERATION Fig. 25. Operating history of Pelton wheels S/N 3 and 4

The following factors concerning the operation of wheels 3 and 4 appear to have a direct bearing on this investigation:

1. Wheels 3 and 4 were subjected to the abnormal steam condition at Valmont.

2. Wheel 3 was subjected to repeated thermal cycling associated with the rapid transition from steam to water and water to steam.

3. Wheel 4 was subjected to a 30% overspeed at Valmont during steam operation.

4. The most severe water operation was experienced by wheel 4 during high-speed operation at FSV.

2.3. EXPERIMENTAL AND ANALYTICAL PROGRAM A systematic program of analysis and experimentation was instituted to establish the causal mechanism for initiating the fractures and the operational safety margins that exist if the fractures occur.

2.3.1. Fractures in the Bucket Root Area

2. 3.1.1. Causal Mechanism. The metallurgical evidence presented in the previous section supports the a priori judgement that the bucket fractures found in wheel 4 resulted from an alternating bending stress, probably caused by the high-speed water operation at FSV.

(

2.3.1.2. Basic Stresses. The analysis of the bucket root stresses were reviewed at the 10,500 RPM design point with the following results:

1. The uniform centrifugal stress was 5,000 psi (tension).

2. The bending stress induced by the force of the water at the tip of the splitter (point of the crack initiation) was 19,000 psi.

42

3. The steady bending stress caused by the displacement of the l bucket cross-sectional mass centers from a radial plane was 1,000 l pai (compression) at the splitter.

The 19,000-psi bending stress caused a 38,000-psi double-amplitude cyclical stress (assuming full reversal after the impulse load was removed) around the 4,000-psi net (5,000 pai uniform tensile minus 1,000 psi bending compressive) steady-state stress. Figure 26 shows this point on a modified Goodman diagram of the 718C material for 10 8 cycles. As will be shown, the free vibration of the bucket occurred at the rate of 6,300 cycles /sec, or 2 x 107 cycles /hr, requiring the use of endurance-limit (assumed to be 108 cycles) criteria for judging safety margins.

2.3.1.3. Dynamic Considerations. Each bucket is excited by the driving water twice during each revolution. Possible amplification of the bucket stresses due to superposition of pulses, as illustred in Fig. 27, was investigated by 1.- Expermentally establishing the first flexural resonant frequency of the bucket (6300 Hz) and the log decrement for the material (0.0016).

2. Formulating the equation of motion for the bucket in the first flexural mode modeled as a one-degree-of-freedom, lightly dampe'd strueture, as follows:

" ,"n g + ( X = F(t) ,

where F(t) represents the skewed double-ramped periodic forcing-function, w,and 6 are the flexural-resonant' frequency and log decrement, respectively, and.the derivatives are with respect-to time, t..

43 t

l

_ _ _ - _ _ _ _ . + ,- L~ - - ,< ~

Q%

NN h\

\\

\ // '\ .

'l

/

N 30 -

/*

CENTRIFUGAL STRESS = 5,000 PSI

_ 20 BENDING STRESS = 119,000 PSI E *

+ m s ?rp A s

._ s s x '

s

's D s , t

~

10 '

CAST 718, 108 CYCLES

, J %

,l  %

NO CRACK INITI ATIONs' ,.

- > ;n - ':', , ~

s ~

0 ,

0 50 100 150 7 mX10~3 ' PSI Fig. 26. Safety margin at 10,550 rpm

i N

O _

I .

T _

A _

T O

R 2 .

l

/

r 7

3 s

c i

m a

n y

d g

_. n r

7 i

_ 1 d a

o

~ l

_ ) t k

e c

- u B

- 7 2

p g i

2

- F I

/r 7

b J

I' -

0 E

C R

O F

D

l l

3. Determining the dynamic response of the bucket through the speed '

range by electrical analog compt;ation.

The amplification of each speed resonance is shown in Fig. 28. Figure 29 shows the averaged amplification at each speed obtained by calculating the total response over a 5% bandwidth, assuming that the time at each ,

speed within the bandwidth follows a normal distribution. Figure 30 summarizes the bucket bending stresses as a function of speed with and without the averaged amplification given in Fig. 29.

Operation at higher than 9000 rpm at FSV could be expected to initiate fracture, as shown by the modified Goodman diagram in Fig. 31. In contrast, Fig. 32 shows the satisfactory safety margin of appcoximately-2 that exists at 7000 rpm, indicating that the stress would not be sufficient to initiate a fracture.

2.3.2. Safety Margins After Incipient Fracture The program to establish the safety margin for operation at 7000 rpm, assuming the buckets-have experienced an incipient fracture, proceeded according to the following methodology:

I

1. Experimental determination of data points derived from actual bucket specimens subjected to extended, controlled-load fatigue testing.

r

2. Crack growth rate characterization of the material (da/dN versus

%) .

3. Analytical verification of the test interpretations by correlation with accepted fracture mechanics principles.-

4. Use of'an analytical model based on the experimentally determined' data points to predict useful life of the. damaged buckets'at

'46

-l

0 0

I 0,

0 1

0 0

I0, .

8 d

e e

p 0 s 0

I 0, t 6 n a

n o

s e

r h

0 c 0 a i 0, e

5 t a

n

' o l i t

i M a P c I R i f

i l

p m

0 a

0 d 0, a 4 o L

8 2

g i

F l

l i

0 0

0, 6 5 4 3 2 i 3

8U5 5 C$ t 3I<

O O

O.

meme m

O O

- O O

O O

- O OT 8 a e

=

.S O

=e4 er4 ed O

O E h

c0

- n O. CC g N CO O

rA 9

4) 60 to O k O e O

E <

O N

_ g ,

m m

I O

O a

., s O

l l l l O O

O O O O

• O og O O

_g n N

= e. . . o. m P M N amme H10lMONV8 %5; B03 B013V3 N011V31311dWV 039VB3AV 48 I

I I

i a

30,000 4

4 4

, WITH DYNAMIC FACTOR 4

20,000 -

G S.

E a

t; E

E 5

i " 10,000 -

WITHOUT DYNAMIC FACTOR t

0 I e i e i i 3,000 4,000 5,0C0 6,000 7,000 8,000 9,000 10,000 RPM b Fig. 30'. . Bucket bending stresses-

'49.

l i

40 AT 9,000 RPM CENTRIFUGAL STRESS = 4,000 PSI BENDING STRESS = 14,000 PSI AMPLIFICATION FACTOR = 1.85 30 -

G n.

m WITH DYNAMIC FACTOR e

o 20 x

~- WITHOUT DYNAMIC FACTOR rn b* <

O >

,i s, .+:mi!

10 Mis ,  ;;g;7

-iji!!G

' f Y

,e '

" iissssssamassissi~.asssicaing:.s  : - -

' /

JNO CRACK INITIATIONC

'S -^30?f:5:s50M:5:0$ny:::g:?:As tysj:is ' , < ,

sig,,,* is .

^

4

^-~

0 0 50 100 150 cr X 10 -3 PSI m

Fig. 31. Safety margin at 9000 rpm

0 5

1 E

D U

T I

L P

M I

A S F 0 P L n 0 m 0 AN 1 p 0I 1 HO r 0S . - I 5, P I ET 0 NA 0 20 = OI 0 0 T 7 R

=

S8T 5O NI ON I S t S C N I

P a E = A I K n R F GC TS i RA ~ g SSN AR 0 r EO MC 1 LRI a ATT I R X m GSA SO y U C PF ,. r m FGI - c t I NF 0S j9 e

MRI I 0S .

g%

f PTDL 0, RE a RNNP ' S EEM 0T 0CBA 1 S '

0 .

0, . ' .!{

0 2

3 7 . ' lg ij 5 T - , _ ,

g A - n i l F

- %N3 lE O

,wI.

%T.

~ ,

eAI

,s .

T.,

- _,Ns Iw

- I

. ', , K, C.

- l ' A.

R.

g s C. ~

- O p N ~,. s, i =

$s:' '

f:

- w {

s _

0 0 _

0 '

0 0 4 3 2 1

_E ?o x b u

various loadings and, in particular, the level of loading that predicts infinite life.

2.3.2.1. Experimental Program. The experimental program to provide test data points for crack propagation in actual buckets employs a controlled loading arrangement that subjects the test specimen to cyclical i bending stresses (near zero to maximum) at the rate of 30 cycles /sec. The apparatus is shown in Fig. 33. Two of these units were in continuous operation, developing crack growth data points for use in the safety margin analysis.

During the course of testing, the specimens are periodically heat tinted to mark the progress of the crack at specific intervals. This allows close determination of crack growth rates for specific phases of the test.

For the most part, specimens began the test series with an artificial flaw produced by electrical discharge machining (EDM) in a 0.010-in.-wide by 0.020-in.-deep slot. This procedure eliminates the time it would take to induce a natural crack. However, when fractures were produced in uncracked specimens, the cracks were in the same location and were similar to the fractures found in the service unit.

Table 7 summarizes all the testa completed or in progress as of October 4, 1974. Test series 6 is of particular interest. A large crack (approximately 0.25-in. deep) was propagated at 23,000 psi. At this point, the fracture was heat marked and the loading continued at 15,000 psi for 7 6

x 10 additional cycles. A microscopic examination of the fracture showed no crack propagation from the heat mark.

These tests clearly demonstrate the self-arresting characteristic of the bucket cracks when the applied loads are below the levels necessary to drive the crack into the principal load-bearing regions of the bucket cross-sectional structure.

52

~ m m M

'OY W.:,f d#P. ,_ - - -, -

. A ,

. g

-(

~

) - .

1. j

. l e e V- A. .,

. . , .s h

I'... -

; .. )

g 7* . ' .: . .

f  :

0

.7; ,. ~. g t

i mogpag 53

l TABLE 7 BUCKET TEST RESULTS Specimen Stress Initially No. (PSI) Cycles Slotted Remarks 6

1 46,000 1.3 x 10 No Total failure 6

2 46,000 1.7 x 10 No No crack initiation 3 23,000 2.2 x 10 Yes 0.27-in. total crack extension 4 37,000 1.6 x 10 Yes Total failure 5 15,000 1.2 x 10 Yes No crack growth 6

18,400 2.0 x 10 Yes Crack growth to 0.09 in.

15,000 1.2 x 10 Yes Limited crack growth 7

y 6 23,000 1.0 x 10 Yes Crack growth to 0.25 in.

6 15,000 7.2 x 10 Yes No crack growth 7 23,000 3.7 x 10 Yes Crack frowth to 0.32 in.

6 8 18,400 7.6 x 10 Yes No crack growth 6

23,000 7.2 x 10 Yes Crack growth to 0.22 in.

15,000 1.0 x 10 Yes Test continues

_ _ n_

I 1

l l

The application of well established principles of fracture mechanics

, to determine the limiting loads for zero crack propagation is described in later sections.

I 2.3.2.2. Fracture Mechanics Properties. Dynamic fracture toughness (kid) data were obtained for cast Inconel 718 through a series of instrumented, precracked Charpy testa performed on specimens fabricated

] from a spare Pelton wheel casting and were tested at Effects Technology, Inc., using their Dynatup 500 instrumented impact system (Ref. 1). A summary of the significant test parameters and results is given in Table 8.

The material exhibits a fracture toughness of approximately 60 ksifiT.,

which is relatively insensitive to temperature in the range tested. This behavior is consistent with the published static fracture toughness (Kg) behavior for wrought Inconel 718 in this temperature range (Ref. 2). It is not certain at .his time whether the lower measured toughness of the cast material is caused by strain rate or microstructure effects, but the

] difference is not severe, and it can be concluded that the measured

! toughness is a reasonable characterization of the Pelton wheel meterial under service conditions.

j The material was further characterized through a series of fatigue crack growth tests performed under various environmental conditions. A j total of six reduced-thickness compact tension specimens were fabricated j from two different locations in Pelton wheels 1 and 3. The spec 1.. ens were tested at Del Research Laboratories, with emphasis placed on obtaining data in the very low (da/dN = 10-8 in./ cycle) crack growth range (Ref. 3). The l test matrix and resulting da/dN versus AK I data are summarized in Fig. 34.

f The test matrix was set up to permit isolation of the effects of temperature, steam environment, R ratio (omin/ cmax), and material variability. The data points with left arrows in Fig. 34 denote the observation (in all specimens) of a crack growth threshold, or AK1 level, below which cracks do not propagate. Once the threshe.td was observed and i

55

TABLE 8 FRACTURE TOUGHNESS TESTING OF INCONEL 718 PELTON WHEEL CASTINGS Experimental Conditions (a) Test Results Frac Gen Integrator Crack Dynamic Specimen Temp Load Energy Time Dial Load Yield Energy Length Fracture Toughness No. (*F) (lb/div) (ft-lb/div) (psec/div) (ft-lb) Pp (lb) Pg (lb) E (ft-lb) a(mils) bd( ' '"*

4(N 600 1000 5 200 10.2 3300 -

9.5 0.079 kid (p=0.010)=62.0 1 600 500 5 200 5.8 1800 --

5.1 0.164 60.1 6 600 500 2 200 4.4 1130 -

3.8 0.214 56.0 2 300 500 2 200 5.2 1680 --

4.8 0.181 63.6 3 300 500 2 200 6.2 1900 -

5.8 0.159 61.4

$ 5 72 500 2 200 4.9 1750 -

4.4 0.184 67.7 7 72 500 2 200 6.3 2180 --

6.0 0.146 64.4

• Impact velocity (v,) = 16.95 ft/sec; system response time (T R) = 62 usec.

( Specimen was not precracked.

- a

SYMBOL

'" MATERIAL

' 10 ENVIRONMENT NUMBER (oF) (R= cr M IN/ a MAX )

e I A RT AIR 0.1 X 2 A 212 STEAM 0.1 A 3 A 212 STEAM 0.8 D 4 A 600 STEAM 0.8 5 8 600 STEAM 0.8 E 6 B RT AIR 0.1 100 R = 0.1

\" xx x" x

lic * ' a* gs xx x, a x

x*x

. %y"x x% so '

a

-y 10- -

$ xt .g "x4x 4 X X' + - ,.A

-?

g AA R = 0.8 i Ak

• A  :; T v

I I i 1

-8 -6 10'9 10 10'I 10 10'b da/dN (IN,/ CYCLE)

Fig. 34. Fatigue crack growth testing Inconel 718 Pelton wheel castings

}

l 57

s confirmed, the load was increased and crack growth was monitored to yield the complete crack growth curve. Comparison of the data from various specimens shows that although the effects of temperature, steam environment, and material variability are secondary, increased R ratio results in a pronounced reduction in the threshold and the crack growth curve. Accordingly, two separate trend lines have been drawn through the data, and lower-bound threshold levels of 8 ksivin. and 5 ksivin! have

~

been established for R = 0.1 and R = 0.8, respectively. These data have been extended over the R ratio range of interest (-1 to +1) using the standard extrapolation procedure shown in Fig. 35.

2.3.2.3. Fracture Mechanics Analysis of Bucket Fractures. The data in Table 7 have been used to develop an analytical model for prediction of crack growth and residual life in cracked Pelton wheel buckets. The accuracy of this model is being corroborated by ongoing fatigue testing of precracked buckets, and the present model reflects several adjustments and corrections which have been incorporated to provide better analytical / experimental agreement.

Crack growth has been assumed to occur in the maximum bending stress cross section of the buckets in two distinct phases, as illustrated in Fig.

36. Phase I refers to the period when the crack growth is restricted to the splitter portion of the bucket and no significant shift in neutral axis occurs. Phase II refers to the later stage cf crack propagation when the crack is growing into the body of the bucket and a shift in neutral axis is expected. The standard equation for stress intensity factor (Kr) in a beading stress field (a )*

b .

at - o, x,jwa- ,

was used to compute stress intensity as a function of crack depth (a) in both phases. The difference in neutral axis behavior between Phase I and Phase II was accounted for by using different bending correction factors 58

S S

E D N L H O G H U S O E) T R 4 ES H 3 RL T UO G

DI TR EI CT R( AN U RO

, SA C AT EA MD 0 1

oO 5

1 0

)

D K W

L - /

• O i). o N "

Hl Si I

M EV RI

- - 0 K

(

d l

HS - _

- 0 o O h k (K 0 0 I s e

A 0 1

5 T A r

- R h t

R K

_ A 5 5

_ 1 3

0 R E _ -

g S i E L E

L C

Y

_ F _

MT A S E C F

G O _

E L N A T -

R R R E A 0 O V P I E 1 S E -

D V LM I OO T

_ HR I SF S E O RA .P HT A S bDN4U E D

N O L I

FC DEN N ERI Y O S L I ANKN T

C BI AO

_ U D

E R

_ - E lll 1l

(Mb ) fr m the literature (Ref. 5). The results are presented in Fig. 36.

Actual bucket loading in service is completely reversed (R = -1), however the compressive portion of the loading cycle is generally not considered to produce significant crack growth, so the curves of Fig. 36 can be considered as plots of AK versus crack length. The extrapolated threshold 7

AK7 of 8 ksivinI. for R = -1 is shown as a horizontal dashed line in Fig.

36. A series of stress intensity curves is shown for various bending 1

stress levels between 0 and 50,000 psi. The corresponding speed (rpm) and Y

e laboratoty test load (lb) are also given for each curve. The most significant feature of these curves is that as the stress (and correspondingly speed and test load) decreases, the stress intensity factor decreases to the point where it eventually falls below the 8000 psi /in!

threshold AK7 level, and therefore no crack growth would be expected. This implies that there is a threshold stress (speed / test load) below which the buckets should have infinite life, regardless of incipient cracking.

Quantitatively, these curves show that while the original Pelton wheel design speed (10,550 rpm) is above the threshold, a reduced operating speed of 7,000 rpm would be below the threshold by more than a factor of two.

As noted above, fatigue testing of precracked buckets has been performed and continues to verify the accuracy of this analytical model.

The first phase tests were performed at 1500, 1200, and 750 lb, yielding the results given in Table 9. The analytical model of Fig. 36 has been used in conjunction with the fatigue crack growth data of Fig. 34 to predict crack growth and failure in these bucket tests, and as evidenced by j the comparative figures in Table 9, the agreement is quite good. In most I

cases the analytical predictions are very accurate, and where there are significant deviations, the model is on the conservative side. Two d probable sources for this conservatism are the use of a lower-bound trend curve for fatigue crack growth (Fig. 34) and the fact that the analytical i model used tends to overestimate neutral axis shifting in the Phase II 1

region.

i i

60 l

. ~

_ 8

_ 1 7

0

_ t 3

1 6 e 0 _ i 0

k c

u

_ b n

) o S t T l F 5 e I

H ) i P

S B 0 f~ e

) 1 t L

. 1 ( /

N S a h I

E I D t

( S X O A

a AH A O I

T n L i P L A

A H T R T 2 P

E D

0 i

R T

U E

N D E

/

S T

e 4

0 TP H

E h

t w

o

)

CI (

M D r C P K g

A R C R (

C 0 0 A k D 0 0 R c E 5 2, C a E 0 0 3 r

)

P l 1 5 0 17i c S / , 7 5 0 T / 0 0 / / 3 F _ ) 0 0 0 0 / e I

H I

S 6, 0, 0 0 0 0

u g

S P 6 5 7, 9, 0,

( 1 I 8 7 i

S / 7 t I I X

' S

/

0 0

/

0

/

0 / a E A

_ S 0 0 0

0 0 0 0 F 1

i S E R 0, 0, 0, 5, i 2

0 A L T 6 7 3 5 0 H A S 4 3 2 1 9 .

P R 6 T

U E

N 1

3 g

O i N

( F 1

t .

0 NK kl I C

- A 0R -

2Cj1 I l g

0. RE _

OP 0 0 0 . 0 0 4 3 -- 2 1 5 '

C _E- E 5y' Uj*E h S

1 l

t TABLE 9 PELTON WHEEL BUCKET ANALYTICAL / EXPERIMENTAL COMPARISON 4

1

! Test Predicted Actual No. )

Load (1b) No. of Cycles of Cycles 0 0 1500 1.4 x 10 1.3 x 10 (Failure) 6 6 1200 4.0 x 10 8 x 10 4 (Failure) 750 (Tint) 13.6 x.100 17 x 10 0 (No tint) 2.5 x 10 0 4 x 10 0 500 = _7 x 106, no crack growth l

)

9 l

62- .j

The favorable analytical / experimental comparison lends a high degree of credibility to the analytical model as applied to Pelton wheel service.

Further testing at 500 lb has confirmed that this load is, in fact, below the crack growth threshold and yield infinite life.

2.3.3. Fractures in the Curvic Coupling Area The Pelton wheel is provided with a face spline which joins with a mating spline on the steam turbine disc. The face splines are of a special geo.netry and are produced on special-purpose machine tools built by Gleason Works of Rochester, New York. Gleason identifies its proprietary design under the registered trademark of Curvic coupling.

In addition to the highly effective torque-coupling characteristic of the Curvic design, the curved-tooth geometry provides precise registration of the mating parts. In their usual form (and as used on the FSV circulators), the Curvic coupling teeth interlock, totally preventing relative radial motion. As a consequence of this feature, circumferential (hoop) strains are essentially the same on both parts in the immediate vicinity of the coupling.

The Curvic coupling must be maintained in an axial preloaded condition to develop the transverse stiffness necessary for shaf t dynamic considerations.

2.3.3.1. Causal Mechanisn. The results of the metallurgical examination suggest that the cracks were propagated by a vibratory stress, i probably emanating from the steam turbine disc. Furthermore, since the vibratory stresses were not likely to be of appreciable magnitude, a high steady-state stress component would be necessary for crack initiation and propagation. As a result, the investigation into the causal mechanism concentrated on i

63

1. Sources of significant steady-state stresses in the tooth fillets.

2. Probable source of a vibratory disturbance from the turbine disc.

! 2.3.3.2. Steady-State Stresses. A comprehensive two-dimensional finite-element model of the Curvic tooth geometry was constructed and analyzed for the following basic loading patterns (as shown in Fig. 37):

1. Compressive stresses applied to the flank of the coupling teeth produced by preloading of the assembled parts.

2. Tensile hoop stresses produced by axisymmetric loads, such as centrifugal stresses, and uniform strains transferred from the t steam turbine disc through the Curvic coupling.

The finite-element computations were used to derive an influence relationship between the local tensile stress in the tooth fillet and the two basic loading mechanisms. The results of the computation provide the following relationship:

of = 0.54 og + 3.25 2

• where o g = principal tensile stress (parallel to the surface) in the tooth fillet radius, o g = compressive stress applied to the flanks of the teeth, 1

02 = uniform hoop tensile stress.

The coupling preloading procedure develops a total axial load of 56,000 lb, corresponding to an average tooth compressive stress of 42,500 pai.

64

1 i

J l

c-) s / 7; I

i of = 0.54 oy + 3.25 c2 i

T N

v "f i

4--- '

+

+--- -.->

4--

02 M 4--- G2

---+

4--

e --->

j 4--

, ~ --*

4---

2 i

Fig. 37. Curvic fillet stresses I

l 65

The most significant hoop stress is caused by the forced axisymmetric l strain transferred by the centrifuga11y loaded steam turbine disc, as l illustrated in Fig. 38. At 10,500 rpm, the induced hoop stress is equal to 11,100 psi.

I Combining these two basic loads in accordance with the influence {

formula given above results in a tooth fillet tensile stress of 59,000 psi.

1 Other stresses that would be additive to this have the following origins: l l

l l

1. Overspeed.

2. Thermal effects.

The records indicate that only Pelton wheel 4 was subjected to an overspeed. At the 13,000-rpm overspeed, the resulting tooth fillet stress increased to 83,900 psi.

Possible stress abnormalities stemming from transient or steady-state temperature distributions were considered, especially with regard to changes in the preloading level and forced hoop strains developed by longitudinal temperature gradients.

Although the bolt effective temperature lags behind the temperature of the clamped components during heating transients, the change in preload is mostly ameliorated by the offsetting disparity in the coefficient of thermal expansion between the bolt and the shaft-disc combination which comprise part of the clamped aesembly.

Thermal modeling of the assembly is subject to error in treating the ,

conductivity across the Curvic coupling interface. Analysis indicates the possibility of sustaining no more than a 125*F temperature differential across the coupling during a severe thermal transient. The difference in 66

=

=+= = = = = _

j

/

/

S S

E _ = = == = == n i

I S T R a r

P S 0T 0E 1 L

, L

> t s

c s

6 I i ,

3F d e -

_ n

. i b

r u

t y

) b d

e s

- u a

c

- s s

M e P r R t 0

0

'- _ t s

0, l e

0  !

l 1

L AT I

D A > i F

Al t R0 8 l

3 4T 7C .

0E g 0L i F .

F

0. E 0D 1

- f e

c l

the coefficient of expansion between the 422 turbine disc and the 718 Pelton wheel reduces the thermal strain so that at the 125'F temperature difference (600*F disc to 475'F Pelton wheel), the added stress in the tooth fillet would be about 10,000 psi.

Therefore, although the added fillet stress is difficult to quantify, the analysis indicates that an additional fillet stress of about 10,000 psi

! due to thermal effects might be possible.

To ensure that some other serious thermal effects were not being overlooked, a standard, preloaded assembly was subjected to several cycles of slew heating to 650*F followed by slow cooling to room temperature.

When this procedure did not produce cracks, it was repeated, except that the cooling portion of the cycle was replaced by oil quenching of the Pelton wheel. No fractures developed.

2.3.3.3. Vibratory Stresses. The probability that, the steam was involved in generating the vibratory stresses is supported by the following facts:

1. There were no failurer on the prenuclear Pelton wheels.

2. Wheels 3 and 4 experienced abnormal steam ~ceration at Valmont (where the fractures were concluded to have initiated).

3. The most rigorous water operation on wheel 4 occured at FSV.

Although the influence of the steam primarily acts to produce thermal }

effects, the idantification of high-cycle fatigue damage suggests that the steam operation could be responsible for the vibratory component as well. .

Accordingly, the system was reviewed for strong resonators that could

1. Be excited by steam operation.

2.- Produce significant streseas in the Curvic coupling area.

68

The fundamental (zero circle) flexural mode of the turbine disc was experinentally and analytically found to be resonant at 405 Hz.

Plate bending analysis and a finite-element computation established that a 0.015-in. deflection of the turbine disc rim would develop a Curvic coupling hoop stress of approximately 3000 psi, equivalent to a fillet stress of 9800 psi. An alternating stress of this magnitude applied to an area subjected to a high steady-state stress would be sufficient to initiate and propagate a fracture.

A review of the probable steam power spectral density uncovered several important differences between Valmont operation and normal plant conditions:

1. A considerable amount of the Valmont operation on the S/N 3 and 4 units involved choked flow through the steam nozzle.

2. Valmont steam piping had overly abrupt transitions bounding short runs of pipe.

These two factors indicate that the cteam acoustical power spectral density at Valmont contained higher-than-normal frequency activity, particularly at the low-frequency end of the spectrum (see Refs. 6 and 7).

1 The response of the steam turbine disc to the influence of the steam excitation would be in a resonant mode at an amplitude depending on the steam power spectral density distribution. Higher energy content at i

frequencies near the resonant mode frequency would increase the amplitude of the response. Therefore, the alternating stress in the Curvic coupling area was most likely generated by the steam turbine disc vibracing in its first umbrella (zero nodal circle) mode, excited by the abnormal steam conditions that existed during operetion at the Valmont -test facility.

69

A very small extension of the Curvic coupling cracks occurred during the operation of wheel 4 at FSV (this was detected by the difference in degree of oxidation on the fracture surfaces). This observation is not incongrous with the judgment that the steam conditiens at Valmont were the basic cause of the cracking. During the 81 hr of steam operation at FSV, the turbine disc would have generated over 108 cycles of vibration. Fven though the stress amplitude at FSV is significantly reduced from the Valmont amplitude levels, this number of cycles would be expected to cause _

the slight crack extension that was observed. As shown in the following section, the crack quickly moves into an area of reduced stress, which attenuates all crack growth regardless of the magnitude of the driving vibratory disturbance.

A preliminary attempt to produce the Curvic coupling cracking by exciting the turbine disc on a shake table, as shown in Fig. 39, was unsuccessful. Spurious modes due to fixture problems and errors in _

simulating the sl. aft connection to the disc gave lower stresses in the Curvic coupling than expected. The absence of the normally high steady- y, state stress (caused by the steam turbine disc centrifugal loading) necesritated exaggerating the vibratory component in order to induce the initiation and propagation of the cracks. At the higher vibratory amplitudes, the simulated shaft, which was clamped to the turbine disc ,

rather than joined through a Curvic coupling as it should have been, underwent severe lateral displacements. This activity, combined with inadequate coupling of the unit to the shake table top, prevented a pure excitatien of the disc umbrella mode at a sufficient amplitude to induce the crack initiation and propagation. .

2.3.3.4. Upper Limit on Stress. The Curvic coupling does not permit -

relative motion of the mating components. Therefore, the stress exposure experienced by the Pelton whccl in the vicinity of the Curvic coupling would essentially be the same as that experienced by the turbine disc.

70 N_

.. . . . _ . . , ., . . . , _ . , .. . . , , _ . , _ . - . . . . . . . , c ._. . . . , , . . . . . _ .

2*:?'O:$Y

. hl 'rj ,'- . .

.h ., % '3 .,)F

~'

4.',4 >. ; 4 ,

~

' ? '[ W l k_ Y

_ f ._Y'h .

+

~u ( c[ . , I' ,' .

M ' ~ k'i .,

R-l s.,cf ,'

.zf f '-(.L .

. . . g;,. ; . . . . ;

-v -1 4. ,- Z4; ,'

-4

. . .. ': r p'.;:-l.

4 ,

h ' . ;'

m n ,emRes';

me. D -f ' [. .j " '

hye ^Q.[ y b .. .

Ji ). %

.. I Vd. #Yg[,

A s;-t_[;t:, -Q,,," &( f ,

y  ; , D g h; }h

.esffy . s. x-i. .

65)

wy .

~%

.. ' {j;h,

, af WM** 'd i G A , t Ak _

RO $g, au .,. ..

e - -o . .

.m e sg*,,3MAgg;%l; '

. v. .;f- g ,

W - Mg , yiWy <. .

y,, M enn LQ@ y:!% *f, v.s r%.7

- 'l .. ,-j 9

_ . m . s disi M [g

- - - O'9:a .m  :

Fig. 39. Shake table experiment 71

Although the stress / cycle exposure was sufficient to cause crack initiation and limited propagation in the cast Inconel 718 Pelton wheel, the same experience failed to produce cracks in the wrought 422 turbine disc material. This sets an upper bound on the stresses that existed at the Curvic coupling during the period in which the cracking developed in the Pelton wheel Curvic coupling.

2.3.3.5. Fracture Mechanics Evaluation of Curvic Coupling Fractures.

l Several analyses have been performed to evaluate the effect of the observed cracking on the ability of the Pelton wheel Curvic coupling to sustain service loads under steam and water operation. A finite-element model of the Pelton wheel / turbine disk / shaft assembly has been set up (Fig. 40) and run under various loading conditions to simulate circulator operation.

Centrifugal stresses in the assembly at a base speed of 10,000 rpm were determined and are plotted as the steady-state component of stress in Fig.

41. This curve shows a stress of 9600 psi at the curvic coupling, which attenuates to approximately 2000 psi in the central region of the wheel before increasing again near the bucket hub. Based on the premise that steam excitation of the fundamental ficxural mode of the turbine disc was _

responsible for the high-cycle fatigue cracking in the Curvic region, a ---

second case of i0.015-in. dispiccoment at the tip of the turbine disc was analyzed. The resulting stresses are shown as the " alternating" component of stress in Fig. 41 and are approximately 13000 psi at the Curvic coupling, but die out very rapidly to zero shortly beyond the slinger region of the Pelton wheel. This is consistent with the fact that none of the observed Curvic cracks progressed much beyond the slinger region since at that point there would be no cyclic stress remaining to propagate them.

It should be noted that the stresses shown in Fig. 41 do not include any stress concentration effects due to the Curvic root radius or any mean stress effects of preload and temperature, making the magnitude of the near-Curvic stresses not meaningful. However, the trend of rapid 72

o o

' SHAFT o 1, o o

/CURVIC

/ COUPLINGS

,~c  :

TURBINE DISC -

( -

/ -,

p il Ip di il it di it ip ip

^

.6, -

L '

lAl

.' . (- . -

./ . f : f _; " ;_ I PELTON WHEEL

=

T I

_ \

t

= d r

_ k_ -

o o .

r 1, o

=

Fig. 40. Finite-element model ,

73 1

h

(. .

//

15.0 t

C

_U M 10.0 --

5 STEADY-STATE ETRESS COMPONENT 8 TO 10,000 RPM ROTATION S

a ALTERNATING STRESS COMPONENT DUE W TO 0.015-IN. AXlAL DISPLACEMENT

• AT T!P OF TURBINE DISC

[

l I  !  ! I 1.0 2.0 30 4.0 50 DISTANCE FROM CURVIC COUPLING (IN.)

Fig. 41. Pelton wheel stress distributicn 74 s

attenuation of stresses away from the Curvic cracks should persist regardless of magnitude. This attentuation suggests that the observed Curvic cracks are at or near their maximum length and that continued crack propagation is not expected.

The second subject which was addressed is the length at which the Curvic cracks begin to affect the capacity of the Pelton wheels to sustain service loading. A conservative fracture mechanics model was set up (Fig.

42). In this model, a constant hoop stress of 5000 psi was assumed to act over the entire length of the Pelton wheel. Approximating the Pelton wheel cracks by an edge crack in a flat plate (as shown in Fig. 42), the stress intensity factor has been plotted as a function of crack depth. The measured fracture toughness of 60 kai/in! is shown as a horizontal dashed line in Fig. 42. The intersection of the fracture toughness line and the ctress intensity curve can be interpreted as a conservative estimate of the critical crack size for Pelton wheel failure.

The following is a surumary of the conservatisms in the above model:

1. A constant 5000-psi hoop stress was assumed over the entire length of the Pelton wheel, while the calculated stress is less than 2500 psi over the main body of the wheel.

2. The crack was assumed to have a free edge at the Curvic coupling end even though the mating Curvic coupling on the turbine disc would act to restrain this edge and inhibit failure.

) 3. The slinger disk was assumed to be completely cracked even though none of the observed cracks propagated into the slinger at all.

4. The critical crack size determination was performed at 10,000 rpm, although the proposed limit speed for Pelton operation is 7000 rpm, which results in a factor of two decrease in hoop stress and a corresponding increase in critical crack size.

75

1 n < >

f ~

1 = 7 IN. 2

~

l 4-- M i 4-- M

()

w = 10,000 RPM cr = 5,000 PSI MEASURED FRACTURE TOUGHNESS, K id 1

1

! l l .

50 Is l

>- 40 -

U

~_  !

J l P l

'd 30 -

lCRITICALCRACKSIZE

1 5 l E

20 - MA IM h OEMD l g, CRACK DEPTH I C (S/N 4) l

_ !i l 10 -

l -

I i

l I

I I I I I I l.0 2.0 3.0 4.0 50 CRACK DEPTH a (IN.)

Fig. 42. Pelton wheel critical crack size evaluation 76 l

It is noteworthy that even in light of the above-mentioned conse rvatisms , tl 3 calculated critical crack size is still about a factor of four greater than the largest observed crack (and the maximum depth to which any of the cracks is expected to propagate).

2.3.4 Summary 2.3.4.1. Incipient Fractures in the Buckets.

1. The cause of the fractures is the high-speed water operation experienced at FSV.

2. At 7000 rpm, fractures will not be initiated.

3. At 7000 rpm , buckets with incipient fractures have an essentially infinite life.

2.3.4.2. Localized Fractures in the Curvic Coupling Area.

1. The fractures were initiated by a vibratory stress, most probably emanating from the steam turbine disc.

2. The steam conditions at Valmont produced a choked flow situation conducive to exciting the turbine disc at an amplitude sufficient to generate the damaging citernating stress component.

f 3. The total stress experience that existed at the Curvic coupling, although sufficient to initiate fracture in the 718C Pelton wheel, was within the strength capability of the wrought 422 material.

77

4. The fractures will be self-limiting and confined to the immediate vicinity of the Curvic coupling and will not impair the capability of the unit to function satisfactorily.

2.3.5. Consequences of Failure In the unlikely event that a bucket fracture progresses to a failure, the physical damage would be limited to the Pelton wheel cavity. The bucket would be contained within the inner steam casing. ,

Multiple bucket failures could cause appreciable rotor unbalance.

When the failures progress to this stage, the rotor displacement instrumentation will detect the malfunction. This was demonstrated during the development program when, in two separate occurrences, several of the heavier pronuclear Pelton buckets separated owing to advanced cavitation damage in the base of the bucket. In both instances, circulator damage was minor, involving a slight bearing rub in addition to the damage in the Pelton wheel cavity.

A total, rapid failure of the Pelton wheel structure could cause considerable damage to the entire lower section of the circulator, primarily as a result of the unconstrained turbine disc. However, the primary coolant boundary or any other safety-related feature would not be involved in the accident. (The steam turbine disc and Pelton wheel are not within the primary coolant boundary) .

2.4. SAFETY CONSIDERATIONS ASSOCIATED WITH REDUCED-SPEED WATER OPERATION OF CIRCULATORS -

To evaluate safety aspects of the FSV operation with reduced rpm of the Pelton wheels, the following approach was adopted: (1) from a list of events requiring operation of the circulators on the Pelton wheel drive, 78

f the events imposing the most severe requirements were sclected; (2) these events were reanalyzed assuming restricted operating conditions of the Pelton wheel drives; and (3) the reanalysis was evaluated with respect to safety limits and AEC accident-dose guidelines.

A list of typical events defined in the FSAR as requiring circulator operation on the Pelton wheel drive is given in Table 10. The FSAR text describing each of these events indicates that acceptable temperature conditions were maintained throughout the transients. Elsewhere in the FSAR, events requiring Pelton wheel operation of the circulators, assuming specific initiating events, are discussed. Each of these events is identified, and all are listed in Table 11. From the specific sequence of events identified in the FSAR, an envelope event from Table 10 was identified. (For example, the loss of off-site power plus turbine trip event, Section 10.3.1 of the FSAR, results in availability of steam turbine circulator drive for 25 min after scram before Pelton wheel operation on condensate is required. Thus, this event is enveloped by case B1 in Table 10.

As a result of this systematic identification of events, the DBA-2, Rapid Depressurization Accident, was identified as the event which imposes the most stringent requirements for forced circulation. For pressurized conditions, the SSE or maximum tornado was determined to impose the most stringent. cooling requirements. Therefore these two events were selected for reanalysis. The conditions defining the basis for the reanalysis are given in Table 12.

Next, the safety limits which define the bounds of acceptable component temperaturas were determined. The safety limits of the critical components are given in Table 13. These limits were used to evaluate the transients resulting from the reanalysis. The fuel limiting temperature is in accordance with FSAR 3.2.3.3. The other temperature limits have been established by GA based on structural considerations.

F 79

TABLE 10 MAXEiUM HELIUM TEMPERATURES AFTER LOSS OF NORWL COOLING Maximum Core Outlet He Temperature (*F)___

Case Cooling Conditions Active Core Hot References in FSAR Average Zone

-- Normal full-load operation 1470 1515 A 1 circulator, steam driven; 1495 1545 Fig. 4.3-3; Sections 4.3.1, 14.4.1 no loss of helium pressure B 1 circulator, feedwater driven; 1510 1570 Fig. 14.4-1; Section 14.4.2 no loss of helium pressure B1 1 circulator, condensate driven; 1535 1690 Fig. 14.4-2; Section 14.4.2.1

, .no loss of helium pressure o

B2 1 circulator, fire-water driven; 1550 1835 Fig. 14.4-3; Section 14.4.2.1 no lor.s of helium pressure C. 1 circulator, steam driven; 1495 1545 Fig. 14.4-4; Section 14.4.3.1 with helium depressurization

.D. 2 circulators, feedwater driven; 1505 1560 Fig. 14.4-5; Section 14.4.3.2 with helium depressurization E 1 circulator, feedwater driven; 1510 1570 Fig. 14.4-6; Section 14.4.3.2 with helium depressurization F 2 circulators, feedwater driven; %1550 %1900 Fig. 14.11-11; Section 14.11 with helium depressurization

I TABLE 11 EVENTS REQUIRING PELTON WHEEL OPERATION Event Envelope Event

1. Loss of electrical supplies

a. LOSP + TT (FSAR 10.3.1) Case B1

b. LOSP + TT + D/G failure (FSAR 10.3.2) Case B1

c. Loss of condenser vacuum (FSAR 10.3.4) ' Case B1

2. Sudden loss of steam supply

a. Pipe ruptures (FSAR 10.3.3) Case B

b. Less of three BFPa (FSAR 10.3.7) Case B1

c. Trip of two circulators in one loop Case B operation (FSAR 7.1.2.4)

3. Natural phenomena

a. Maximum tornado (FSAR 10.3.9) Case B2

b. SSE (FSAR 10.3.9) Case B2

c. SSE with single failures (FSAR 10.3.10) Case B2

4. Steam inleakage

a. Subheader rupture and wrong loop dump Case B1 (FSAR 14.5.3.2)

b. Subheader rupture and moisture detector Case B1 failure (FSAR 14.5.3.4)

5. Depressurization

a. Maximum credible accident (FSAR 14.8) Case D or E

b. DBA-2, rapid depressurization (FSAR 14.11) Worst case (F)

?

l l

.81

., . . . . . . . . . . .....T . 4 .

TABLE 12 WORST-CASE ACCIDENTS REANALYZED

1. Design basis depressurization accident (DBA-2)

Event results in Rapid depressurization Unavailability of steam supply to circulators Operational Feedwater drive to circulators

2. Safe Shutdown Earthquake (SSE)

Event results in LOSP + TT l Loss of deaerator Loss of all three BFPs Loss of all condensate pumps Loss of auxiliary boiler Operational Fire-water pump available for circulator drive and steam generator coolant supply

.y 82

[

i TABLE 13 SU) DIARY OF SAFETY LD11TS Component Limit (*F) Potential Consequence Fuel 2900 Appreciable fuel (FSAR 3.2.3.3) particle coating failure Thermal barrier Side wall 2000 Coolant flow blockage 7

Steam generator-Average helium 1950 Coolant flow blockage inlet 1

i

~

r l

l l

83

i i

i l

The rapid depressurization event was analyzed assuming an initial power level of 105% with a 5-min delay to start up one circulator driven by feedwater to the Pelton wheel. The peak component temperatures as a function of flow or single circulator speed are given in Fig. 43. At 7000 rpm, critical component temperatures are below their safety limits. Figure 44 shows salient temperatures during the course of the transient, assuming a circulator speed of 7000 rpm. Figure 45 shows the heat generation and heat removal rates as a function of time for the case of one circulator at 7000 rpm. Heat removal exceeds heat generation after about 1-1/2 hr in the l average channel.

In order to further illustrate the effect of reduced rpm operation of the Pelton wheels (7,000 rpm versus 10,550 rpm), Fig. 46 is presented, which indicates salient temperatures during the course of the transient, assuming a circulator speed of 10,550 rpm.

Following an SSE or maximum tornado, cooling is accomplished by one circulator driven by fire water to the Pelton wheel. The transient conditions for this case, assuming a 5-min circulator start-up delay, are given in Fig. 47. Note that temperatures remain below the maximum values allowed for continuous operation.

From the analysis performed, it can be concluded that the safety of the FSV reactor is not compromised or diminished.

Table 14 shows acceptance criteria for accident consequences. The SSE or maximum tornado appears to be inconsequential. No release of radioactive material occurs, and no damage which could increase the potential for release takes place. The rapid depressurization event, assuming one circulator operating at 7000 rpm, does not result in component or fuel damage which would increase the release of fission products beyond ,

1 the level already reviewed by the AEC via the original FSAR analyses.

84

3,000 777f77 INITIAL POWER = 105%,

5-MIN START-UP DELAY MAXIMUM FUEL C

MAXIMUM TB E

5 AVERAGE STEAM GENERATOR INLET

$ / / / / / / SAFETY LIMIT

@ CIRCULATOR ISOLATION VALVE FULLY OPEN 1,000 -

t t t t i i I 5 10 15 20 25 30 35 40 COOLANT FLOW (103 tgjgg) i I I I l 0 2 4 6 8 10 EQUIVALENT SPEED FOR ONE CIRCULATOR (10 RPM)

Fig. 43. Peak component temperatures following rapid depressurization

i IlI ) l 0 .

1 T _

E T L I 9

L T E N E E I L L T

N T U U R O

O O M T U E E A _

M R R _

R E I

O O N 8 X

A C C E I

r M E G e U

M G M

w o _

A M

I R A E

p -

E X

V T  % -

A A S M 5 E I 7 0 G 1 A

R t E

V a A

M n P

R o

i 6 i 0 t 0 a 0Y z 7A i TE L ) r AD R

H u

( s RP 5 s e

l OU E T - M r AT I p LR T e UA d CT RS I

d CN i I I 4 p EM N - a O5 r V

_ S F

I 3 4

4 g

i I

2 F 1 I

- - - - - - O --

1 -

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 4 2 0 8 6 4 2 -

2 2 2 2 1 1 1 1 -

Co_ IEaE E -

~~

n

L A

V I C O N I M O E I R T T A R

A E E N H E G I 9 T

A E

H _

l 8 s e

t a

r n

o I

7 i t _

M a P _

R r _

e 0 n

%0

,0 e g

57 0 1 6 1 T t A a

=

R e

0RG h EET )

d ZWA R n

I OL H RPU 5

( a U C l SLR E l SAI EI C M a RT I

T v PI E o ENN m OI O e r

l 4 t

a e

h V

S l

3 F 5

4 l 2 g

i F

l i 0 0 0 0 o

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 7 6 5 4 3 2 1

$ W $ "oC P I y E O

ll l  ! l !

T _ m E

L l o

T t U

O E

R O L C E U

M F I 9

U M M M I U P X M R A I M X 0 A _

5 M 5

I 8 r 0Y 1 A e L w TE o AD p RP  %

OU 5 T - T 7 AT E I 0

LR L 1 UA N CT I t

RS I R a CN O I

T E T n EM L A o N - T R I 6 i O5 U E t O E N a E G z R i O M r C A E

)

R u T H s E

S

( s G

A I 5 e R E E M

r E G p V A R

I T e A E d

V A d i

I 4 p a

r .

V _

S F

I 3

6 4

g i

I 2 F d

1 1 -

- - - - - - - 0 0 0 0 m 0 0 0 0 0 a 0 0 0 0 0 0 0 0 m 6 0 2 0 8 4 2 8 2 2 1 1 1 1 1

{o Eaggb

@e r

e s

a a

E 4 M

i 4

4

, h

MAXIMUM FUEL ONE CIRCULATOR DRIVEN BY FIRE-MAIN WATER 1700 - (109,000 LB/HR AT 900 RPM),

5-MIN START-UP DELAY

/

1500 c C 1300 -

NOTE: TEMPERATURES REMAIN BELOW MAXIMUMS

- ALLOWED FOR CONTINUOUS OPERATION E

E 3

g 1100 -

5

@ MAXIMUM CORE OUTLET 900 -

700 -

CIRCULATOR ISOLATION VALVE STARTS TO OPEN AT 91,000 LB/HR AVERAGE CORE OUTLET AND IS FULLY OPEN AT 109,000 LB/HR 500 -

AVERAGE STEAM GENERATOR INLET I I I I I I I I I i 0 1 2 3 4 5 6 7 8 9 10 TIME (HR)

Fig. 47. FSV pressurized cooldown from 105% power

~,

4 i

TABLE 14 FSV ACCEPTANCE CRITERIA FOR CONSEQUENCES OF DBA-2 Release Doses Design coolant-borne activity Below 10CFR100 guidelines

+ fraction of plate-out activity ,

Corollary DBA-2 consequences (one circulator,10,500 rpm) Both acceptable g- DBA-2 consequences (one circulator, 7,000 rpm) No change

>;s .--

_ . _ _ - _ . . _ . _ . _ - - - . _ - _ - - - - - u . - _ -- . _ _ _ _ ___. _ . . . _ .

2.5. WATER OPERATION SPEED CONTROL Since the Pelton wheel is presently capable of attaining a speed of 11,500 rpm durinC depressurized conditions (driven by the boiler feed pump), the control system will be adjusted to keep the speed below the newly established maximum of 7,000 rpm. Additional signals taken from the protection system sensors will be used to monitor the speed of the circulators to ensure that the speed does not exceed 7,000 rpm. If, in the unlikely event, the speed does exceed 7,000 rpm, the protection system will cause the Pelton wheel drive to revert to manual control from automatic control. The operator will be alerted to the malfunction by an alarm. In addition, an overspeed trip set at 8000 rpm will be included.

2.6. CONSULTANTS AND ADVISING AGENCIES In addition to drawing on the many individuals and organizations within GA for their assistance on this program, opinions and advice were obtained from the following experts in related disciplines:

1. Messrs Cordovi, Clark, Muller, Smith, and Ward of the International Nickel Company.

2. Messrs Carlson and Almeida of the Gleason Works.

7

3. Dr. Paul C. Paris of the Del Research Laboratories.

4. Messrs Krasner, Freche, IIallford, Coy, Chamlis, and Brown of NASA-Lewis.

l

5. Dr. Donald T. Klodt, Consulting Metallurgist to PSC.

6. Mr. J. V. Neely, Consultant to PSC.

7. Mr. R. Adset of the Convair Division, General Dynamics Corporation.

91

3. CONCLUSIONS On the basis of the foregoing technical investigations, the following I general conclusions have been reached:

1. No new limitations are required for steam operation of the helium circulators.

2. The present cast Pelton wheels could be safely operated for the planned or postulated accident conditions by imposing a lower limit on maximum Pelton wheel speed.

3. Plant safety is not compromised at 105% power conditions with an imposed' Pelton wheel speed limit of 7000 rpm.

These general conclusions are based on specific facts and evidence obtained from (1) review of the operating history. (2) metallurgical invertigations and analyses, (3) experimental results, (4) fracture

]

mechanics analysis, and (5) plant safety analyses associated with reduced Pelton wheel maximum speed.

Specific conclusions are summarized below.

3.1. CONCLUSIONS FROM METALLURGICAL EXAMINATIONS

{

The following conclusions have been derived based on metallurgical 1

examinations: I 1

1. Based on the oxidation of the fracture surfaces, it is concluded that the cracking in the Curvic area on the S/N 3 and 4 wheels was produced during operation at Valmont.

92

2. Some limited propagation of the Curvic cracking occurred during operation of the S/N 4 wheel at FSV.

3. Because of the lack of heavy oxidation of the fracture surfaces and since bucket cracking was confined to the S/N 4 wheel, which had been operated at FSV in addition to operation at Valmont, it is assumed the cracking took place at FSV.

4. All the cracking on both wheels appeared similar in character and was concluded to be caused by high-cycle fatigue.

5. The material and physical properties obtained from the heat treatment are within specification.

3.2. CONCLUSIONS FROM PREVIOUS OPERATING HISTORY AND EXPERIMENTAL STRESS AND SAFETY ANALYSIS WORK 3.2.1. Incipient Fractures in the Buckets

1. The cause of the fractures is the high-speed water operation experienced at FSV.

2. At 7000 rpm, fractures will not be initiated.

3. At 7000 rpm, buckets with incipient fractures have an essentially infinite life.

3.2.2. Localized Fractures in the Curvic Coupling Area

1. . The fractures were initiated by a vibratory stress, most probably emanating from the steam turbine disc.

2. The steam conditions at Valmont were abnormal and conducive to '

exciting the turbine disc at an amplitude sufficient to generate-the damaging alternating stess component.

93

i

3. The total stress experience that existed at the Curvic coupling, although sufficient to initiate fracture in the 718C Pelton wheel, was within the strength capability of the wrought 422 material. .

4. The fractures will be confined to the immediate vicinity of the Curvic coupling and will not impair the capability of the unit to function satisfactorily.

3.2.3. Safety Aspects

1. The SSE or maximum tornado events appear to be inconsequential; i.e., no release of radioactive material occurs, and no damage which could increase the potential for release takes place.

2. The rapid depressurization event, with one circulator operating at 7000 rpm, does not result in component or fuel damage which would increase the release of fission products beyond the level already reviewed by the AEC via the original FSAR analysis.

l 94

l l

l I

1 I i l

REFERENCES

1. Server, W. L., D. R. Ireland, and R. A. Wullaert, " Strength and Toughness Evaluations from an Instrumented Impact Test," ETI Report i TR-74-29, March 1974.

2. " Technical Support Package: Materials Data Handbook: Inconel Alloy 718," NASA Technical Brief 67-10282, August 1967.

3. Paris, P. C., et al. , " Extensive Study of Low Fatigue Crack Growth Rates in A-533 Steels," in Stress Analysis and Growth of Cracks, ASTM STP 513, September 1972.

4. Pook, L. P., " Fatigue Crack Growth Data for Various Materials Deduced From the Fatigue Lives of Precracked Plates," in Stress Analysis and Growth of Cracks, ASTM STP 513, September 1972.

5. "ASME Boiler and Pressure Vessel Code, Section XI: Rules for Inservice Inspection of Nuclear Power Plant Components," 1974, Appendix A.

6. Horse, P. M., Vibrations and Sound, McGraw-Hill, New York,1948, Chapter 23.

7. Lighthill, M. J., " Jet Noise," AIAA Journal 1 (1963), p. 1507.

i l

95 f

h;#> $e#

4Ae

+*@

L_ _ _ ___ - - - - - - - - _ - _ - _ - - - - - - - - - - - - - - - - - - - - _ _ - - - - - - -