ML20100N533

Intervenor Exhibit I-7,consisting of Suffolk County Exhibits 7,24,32,54-59,66 & 67 Re Cylinder Blocks & June 1984 Design Review of Tdi R-4 & RV-4 Series Emergency Diesel Generator Cylinder Block & Liners

ML20100N533
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	11/01/1984
From:	FAILURE ANALYSIS ASSOCIATES, INC., SUFFOLK COUNTY, NY
To:
References
OL-I-007, OL-I-7, NUDOCS 8412130209
Download: ML20100N533 (36)

Text

go-3290 q _

-r

_ - JM l

l 4 R.ough f- 69 v i SUFFOLK COUNTY, 7/31/84 G

$s q>'1_I L/ s'q UNITED STATES OF AMERICA &?/

NUCLEAR REGULATORY COMMISSION '/ LOC :J3 Y~

Before the Atomic Safety and Licensing Boartd, IIbY 3 U3 > 1 0 ,T'; 7' ':a 0% ""L; ...jc' e

) (P A

g In the Matter of )

)

LONG ISLAND LIGHTING COMPANY ) Docket No. 50-322-OL

)

(Shoreham Nuclear Power Station, )

Unit 1) )

D b SUFFOLK COUNTY'S EXHIBITS TO JOINT DIRECT TESTIMONY CYLINDER BLOCK EXHIBITS

]

gyg g gaseLAT9ET geemtsatts seed no S o . j,_7j,,JgA, _etnad sin.E'/' s I-_

is a s sw e r et - g IDtuftFitt

_ststif te__

)

V etJESTtB_ -

. , . < - ,,,, ,,,, . s St . ' St - / _

b 3"

8412130209 841101 PDR ADOCK 05000322 Q PDR

4,

~

N ll VOLUME 4 g,  ! iOh, O- 's' .' 7 '

CYLINDER BLOCK EXHIBITS v 'Uc .(

4ci; '% ,,j.ecy Xp4

7. DesignReviewofTDIR-4andRV-4SerienEmergencyd,ielseII\

~

Generator Cylinder Blocks and Liners, JLne 1984 3 24. Deposition of Maurice H. Lowery, pgs. 1, 14-16

32. Deposition of Clinton'Mathews, pgs. 106-107 j l

54. Letter from Reis to the Administrative Judges Concerning a l

Morning Report of 4/16/84 J_

~

55. 3/20/84 Morning Report Concerning Con Rod Bearing Cracks and Eddy Current Examination of ti.? Cylinder Blocks Cracks

56. TDI Owner's Group DRQR - Cylinder Block l l

O 57. Deposition of William J. Museler, pgs. 1, 7-8, 14-17, )

43-46, 98-99 '

58. Deposition of Robert Taylor, pgs. 1, and Exhibit No. 1

59. Deposition of Robert Taylor, pas. 1, 39-41, 67- 69-70

,3

66. Deposition of Simon K. Chen, pgs. 1, 129

67. Handwritten Memo to Pratt from Lowery on Cylinder Block Casting - RV's O

3

3 l

l l

b

O -

\. , Ptlin gr,; - - ~ ;i , .

-Q N).w ' N(f,s gy '

s(-

M O \L aAA-84-5-4

?, _ , 9 y, [2 03315A/RKT 9.

~ ~ . -' p .

e, u  :-

O -

, h f C:

' s' Y;~ ,

arygeg-DESIGN REVIEW 0F TDI R-4 AM RY-4 SERIES

' EMERGENCY DIESEL GENERATDR CYLL SER BLOCKS AND LINERS

.O g This report is final, pending confirmatory reviews required by FaAA's QA ope >ating procedures.

Prepared by Failure Analysis Associates O' Palo Alto, California 1

O l

l

^O Prepared for i TDI Diesel Generator Owners Greup O .

June 1984 0

1

. 1 I

] I l

-l t.

D STATEENT OF APPLICA81LITY J This report summarizes a structural integrity investigation of the TDI R-4 and RV-4 series engines installed in emergency generator sets in nuclear power stations.

3 3-3 G

9 Q

"D 1 1

i 0 -

O

\ EXECUTIVE SUP94ARY

- This report summarizes a generic investigation of the structural adequacy of'the TDI R-4 and RY-4 series diesel engine blocks. The results are

. based on strain gage testing; analytical models, including several 2-D finite element analyses; and review of field experience.

'O Cracks in the block top region have been identified in the diesel generator engines at Shoreham Nuclear Power Station (SNPS) and in other en-ginas in non-nuclear service. The majoritV of cracks can be classified either 9 as radial cracks, extending in a vertical _ plane outward from the cylinder head stud counterbore, or as circumferential cracks, extending downward from a horizontal plane and outward from the corner of the cylinder liner landing.

The radial cracks are the only type found in the SNPS engines, but both radial g and circumferential cracks have been found elsewhere in non-nuclear service.

. An additional type of cracking identified at SNPS is associated with the camshaft bearing supports. This cracking is unique to the inline engines and is attributed by FaAA to the casting process. s At Comanche Peak, cracks

'O unique to one engine block have been found. These have also been attributed by FaAA to the casting process.

There are three possible mechanisms of crack initiation (acting

'O separately or in combination) in the block top. The first mechanism is low cycle fatigue (LCF) associated with the stress range from each startup to high load levels. The second is high frequency fatigue (HFF) due to the firing pressure stresses. For both LCF and HFF there is a high mean tensile stress g resulting from th9rmal expansion and stud preloading. The sum of mean and alternating components may produce the third mechanism, overload ruptu re.

This is most likely to occur above rated power level (>110".) in blocks with below average material properties.

O All of the three mechanisms are potentially responsible for . initiating cracks in the ligaments between the cylinder head stud holes anc the liner counterbore, and such cracking is predicted to occur by Goodman diagram analy-sis. The only projected consequences of this ligament cracking are possible O

11

O-

. t.

coolant leakage (but not into the cylinder) and greater chance of cracking O between studs of adjacent cylinders. .._ _... .. ___ .....,. _ .. _ .,. ... _

- e"e!! the -e stie- ^- e?!" W "ty O' th: ;i.d a;. ;; 2.e T. ty ff;'d ;;-

~

~

~

f ree_

Cracking betwun stud - holes of adjacent cylinders has ban obsened lO infrequently, but is potentially more serious than ligamenc cracking. This cracking has been cbserved in SNPS engine DGIO3 to a depth of 51/2 inches.

Cracking No adverse consequences to engine operation were experienced.

between stud holes is conservatively predicted by Goodman diagram fatigue ~

0 analysis,' assuming a ligament to be cracked, either in LCF or HFF.

A linear cumulative damage model and the observed crack growth in SNPS engine DG103 were . combined to predict conservatively the amount of crack propagation that might occur during one LOOP /LOCA event.

This analysis indi-

O cates that blocks with ligament cracks (e.g., DG101 and DG102) are predicted to withstand a LOOP /LOCA event with sufficient margin provided that: (1).

inspection shows no stud-to-stud cracks detectable between heads, and (2) the

g specific block material of DG103 is shown to be sufficiently less resistant to fatigue than typical gray cast iron, Class 40.

' t F The block tops of all engines that have operated at or above rated loa should b inspected for ligament cracks. Engines such as tnose at Catawb and

,O Grand Gulf t are found to be without ligament cracks can be oper d with-out additional in etion for combinations of load, time, and n er of starts that produce less exp ed damage than the cumulative age prior to the latest inspection. Engine hat have been operate t or above rated load

O without subsequent inspection the block +l. should conservatively be assumed to have cracked ligaments purpose of defining inspection intervals.

For blocks with know r assumed ligament c ks the basic approach to O

assuring reliability i nspection and material evalua n. The absence of tween stud holes should be establishe eddy current i detectable cracks en heads and at the ends of the block before r rning the inspection b eng:ne emergency standby service after any period of operation oth than l 0

iii x[ 1 O

3

9

(

D ,

1.0 INTRODUCTION

This report presents a generic analysis of structural integrity of

' cylinder blocks and liners for TDI R-4 and RV-4 series diesel engines. The

) integrity of any particular cylinder block depends upon several plant-specific variables such as firing pressure and temperature, assembly clearances, cylin-der head stud configuration, and material properties.

1.1 Service Experience 9

Two types of cracks have been found to occur in cylinder block tops of this design: cracks in the radial / vertical plane at ~the stud holes, and circumferential cracks in the liner counterbore at the liner landing ledge

[1-1]. Figure 1-1 depicts the potential location of block top radial and circumferential cracks. In addition, for the in-line engines, cracks have occur' red in the cam gallery above the cam shaft bearing supports. The survey of industry experience with TDI R-4 and RV-4 engines summarized in this sec-

) tion has not been independently confirmed by FaAA, and therefore is not sub-ject to FaAA's usual quality assurance procedures.

1.1.1 :ihorei.fy tbclear Power Station, J Shorehain has three TDI OSR-48 diesel engines designated DG101, DG102, arc.1 DG103. As of April 30, 1984, the engines had operated between 1091 and 1270 hours0.0147 days <br />0.353 hours <br />0.0021 weeks <br />4.83235e-4 months <br />. A significant percentage of those hours was at or above full load, as shown by Tables 1-1 through 1-3.

)

As part of the engine requalification program after the crankshaft replacement, each engine was operated for 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> at or above full load and was then disassembled and inspected. During these inspections,

)

radial / vertical cracks were discovered in the blocks of all three engines.

Crack maps for DG101, 00102, and DG103 are presented in Figures 1-2,1-3, and 1-4, respectively.

)

1-1

) h

O  ;

I

. ,I

' No circumferential cracks were found in any of the engines. However, O each block had radial / vertical cracks between the cylinder bore and the stud hole. Sixty-seven percent of the ligament cracks were between 1 and 1.1/2 inches deep. A typical example of a cross-section of a radial / vertical crack

-through a ligament is shown in Figure 1-5. As depicted in the figure, none of O 'the cracks extended below the corner fr .ned by the counterbore and the coun-terbore landing. ' This demonstrates the apparent arrest of radial / vertical cracks that occur in the ligament region. In addition, when first inspected, the engine block from DG103 had a crack that extended between two adjacent studs on the exhaust side of Cylinder Nos. 4 and 5 as shown in Figure 1-6.

O After inspection, DG102 was operated through 100 starts to loads great-  !

er than 507, and wa- then reinspected. Review of inspection reports before and after the 100 starts showed no crack extension discernable by eddy current

O examination of the stud holes and liner counter bores.

In order to allow calculation o"f the growth rate for the crack between stud.. holes in the DG103 block, a strain gage test of the block top was per-f rmed, as described in Section 3.0. After the strain gage test, LILCO con-iO tinued with qualification testing of the DG103 engine. While operating the engine at full load, the plant experienced an abnormal load excursion. During this event, the power demand exceeded the diesel capacity and, over a period i 23 seconds, the diesel slowed to around 390 rpm, at which time the load das

O dropped. The diesel continued to run at low load for 10 minutes before it was manually shut off. Upon restarting the engine and continuing with qualifica-tion testing at 3900 kW, a crack at Cylinder No. I was noticed, and the test-ing was stopped. At the time the crack was noticed it was reported that the O

engine output parameters were satisfactory.

! Inspection of the block top revealed cracks between stud holes with depth of 1 1/2 inches similar to those shown in Figure 1-6 at three loca-iO tions. At four other locations, between-stud cracks developed along the top surface which did not extend to measureable depth! down the sides of the stud hole. At one location a crack that previously extended 0.8 inch radially from one stud hole towards the adjacent stud hole grew to a maximum depth of 3.9

~

t iO inches. In addition, the original crack between Cylinder Nos. 4 and 5 had ex-1-2 O -

5

)

I' tended to a depth of at least 5 1/2 inches. As shown in Figure 1-7, the D. ligament cracks had also grown approximately 1 inch. Figure 1-8 is a crack map for DG103 as reinspected.

i I

Inspection of Blocks at Other Nuclear Power Stations N -1.1.2 D Catawba Nuclear Power Station has operated its la emergency die enerator approxirrately 810 hours0.00938 days <br />0.225 hours <br />0.00134 weeks <br />3.08205e-4 months <br />. The la diesel has been inspected for ock to cracks, and none have been found. The load history for the la e ine is

'shown n Table 1-4. -

D Riv Bend has two TDI diesel engines of the DSR-48 esign. Each

- engine has a roximately 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> of factory operation only. Engine logs show that both engin were run at 100% load. To date one e ine block has been inspected by the en gnetic particle method. No cracks re found in the block g

top.

Comanche Peak St m Electric Station h inspected both of its TOI

- DSRV-16-4 engines. The en nes have been op ated for approximately 90 hours0.00104 days <br />0.025 hours <br />1.488095e-4 weeks <br />3.4245e-5 months <br /> 3 at the site. Subsequent ins ction of t block top region revealed several indications that are coasiderab diff ent from radial / vertical cracks found at SNPS or elsewhe e. The two lar indications, illustrated in Figure 1-1, have baen metallurgically e::ami d nd weru inr.tifisc as interdendritic 3 shrinkage or porosity resultir. rom the casting process.

Grand Gulf Nuclear ation has inspe ed the block top of the Division 1 engine after 1,397 h rs of operaticn, incl ing 338 hours0.00391 days <br />0.0939 hours <br />5.588624e-4 weeks <br />1.28609e-4 months <br /> between 80% and l

100% load, and 14 h rs at 110% load since Nove er 1981 [1-2]. No indica-tions were reporte . The load history is shown in le 1-4 l

l 1.1.3 Non-N ear Service T experience compiled for engines in non-nuclear ervice tends to suppor the observation of the apparent arrest of ligament crac at the depth of e liner landing ledge (1/2 inches) when cracks between st holes are 1

t present. The motor vessel Edwin H. Gott has been operating fo at least b i l-1-3

I . .

e .

l Hole for cylinder head stud _

Ligament .

/ area Counterbore Between stud hole area- }

_ O l( ""'

[Counterbore 7 Cracks in TDI

& //

D SR V- 16-4 engine at g Radial /

crack Comanche Peak

'"IY '

i F

Pilot ' ,

! Circumferential I""*I*'

I crack l

i Iigure 1-1. Location of cracks.

t.

t . .

Exhaust 1.25' .

1 .2* 1.0* 1.1" O.65*

' /.2* l / a /

O o 00 00 00 00 o o

! I C yl. 4 8 j j 11 C yl. + 7 Ji Ji Cyl. # 6 JJ i j C yl. 4 5 JI l l C yl. + 4 j j l l C y l. 4 3 }l l l C y l.

• 2 l l 1 1 C yl. # 1 l 0

x \ \.5" x0.5" \1.2" O .6* '7 1 O.5" intake ,

i Di'mensions indicate crack depth figure 1-2. SNPS DG101 crack map.

A I

i a

R e f: TER Q-308. TER Q-329, Q-06

y ,

O 0 -

2 y 1

+

2

' - l

. n e

y

• o C 7 P.

t 0

. f O o U * *

. 2 a

8 1 +

O

/ - l C

y C.

1 1

o -

U 3 l

• . y

• C 3

[O 1 h p o t p a o e m U

• 1 NO j/O d k

• . 4 k c

5 0 a

_ + c r a

t s

1

/ l C

y o

c r

o 2

0 1

c u t G a k a D s c U h o i S x n i d P E I N 5 ' i n S

+ -

3 s .

l n 3 y -

U lo C u 1

n e r

V e u

• O o m i g

4 iD f 8

"3 1

+

• 5 1

/ - l C

y o

1 U ,

o "5.

7 O '

O #

• 5 U

/ - .

l C

y 1

o[*

2 0

2 8 '

1 + 1 U

/ - l C

y 1

O

> ,i U.

l; l l l l ll l1 .

C .

t 0

1 3 1

U X + -

.

• l

• Q 5 y 5 R 1 C 1 E T

. O O  :

f e

R U 2 "5. l y 5 1 O C 1 U O 3 .

4 d 4 e

r 8

l /

< "5 3 y 4 "5 u 1 1 *. O C ". 1 s 1 a /

3 O h e t

p m f o

U e d o t

s 4 o n a d

j k c h p a

• l "5 a t p m 5 y "0 r t O C o 1 1 c e k s 1 e D c u e t r

a a k a .

O h x

d t a

i c n o 3 c

E 5 n d i

t 0 f

i 4 i n a 1 G

c D l

' s i d

"5 "5 "5.

1 1 1 ,O f C y

E *7 S. * "5. i 1 s i n

o I n S P

N s e S U \( O n c O e a .

i mf r 4 O 6 -

D us 1 f e p r "5

l y o u g

T 1 C 1

• i F

U O "5

7 l

y O

W_3 "3 1 O C 9

9 4 5 5 2

1 8

l C

y O

A1

• 2 "5 ,iE7!.Y.

1 "9 O -

, l

b .

Cylinder j $tud

" ' ' tUo'u'S crae ligament Counter J M

/ / bore d

- d

/ fCounter

=

af bore j landing

= - = :.:-

3 Figure 1-5. Longitudinal section between adjacent cylinders.

) '

)

Fe A A-84-5-4

3

( '

D

~

3 3 y

~

y x -

~

~ &

~ ~ f y

• ~

3 ~N /_

) V l

^

] k O

~' ::

O i Figure 1-6. Longitudinal section between adjacent cylinders for SNPS DG103.

O FaA A-84-5-4

D J

) - -

l

_- ~

f -

_A

% ~

~

D

~

J kw n_ s

) (~ _ _

m-D

) L Figure 1-7. Longitudinal section between Cylinders 4 and 5 on exhaust side of SNPS DG103.

3 Fa A A-84-5-4

)

k Exhaust .

1.5" 2.5* 1.5* 1.5" 1.5* 7 1 5"

.3" 5.5" *

• 1.5*

1."

1 2 1 1.5*~ 1 .5" 1

/.5" /.5" 1

/.5" 1 /.5* /.3* f.5* /

b M ( M ' M,/ 1.5"% Cyl. 45 X do R Cyl. d4 i : l Cyl. d3 I I l l C yl. 4 2 1 l '

n/

Cyl. e t l l l Cyl. 48 l l l l C yl. 4 7 j i 11 Cyl. 46 j l l l 1 1 I l l

. 1.5" ~/

2.5" 1.5"

/

1.5" 1.5*

f.5' 1 4.4*

1.5"

• 3.9" 2.5" 2.0" 1.5* 1.5*

2.5" 1.5" Intake Dimonsions indicato crack depth

' Top surf ace indication. Na depth to cr'ack measurable down stud hole figure 1-it .ri.'S DG103 crack map as of 4/23/84.

i i

noi. o-4..

A real ase, the ligament crack is only about 1 1/2 inches deep, so less ress O is actu ly transferred to the stud-to-stud region than is calculate by the

. model.

~

As in Sec on 3.2.2, the thermal stress is obtained y scaling the g ' stress measured at ge No.13, using the model in Figure -11 to obtain the scaling factor. The imary assumption in this proce3 is that the thermal stress acts in the plane f the block top, analogous o pressure. The result is 17.3 ksi = 9.7 (10.85/6. for 100% load.

O 3.3 Discussion of Stress State t Crack 5 es 1 (Ligament) and 2 (Stud-to-Stud)

The stress shown in Table 3-2 an divided into mean and alternating O components for fatigue analysis. or low cy e fatigue caused by startup plus load change from 0% to a pa icular load 1 el, the relevant alternating stress range is the peak str ss at that load leve upper pressure band) minus the p' reload. This rang may not be substantially 'ferent for starts from O hot standby or cold st ts, because the stress differenc between cold preload and the lower press e band for steady state running at zer load is less than 1 ksi, as shown Gage No.13 in Figure 3-6. The mean stres is the preloac plus half of e range, while the alternating stress is half th range. For high frequedy fatigue caused by the firing pressures, the relevant lternat..

O ing str s range is that caused by firing (the band on each curve f m the stra gages). The mean stress is the preload plus the therrral range plus ha ference between average (mean) stress at load and the median stress in th stress band from firing. The alternating stress is half tne range.

g

' The results are shown in modified Goodman (Smith) diagrams: Figure 3-13 for ligament cracking and Figure 3-14 for stud-to-stud cracking given cracked ligaments. The curves are derived from the minimum ultimate tensile strength 6

3 in thick sections, the minimum specified endu-ence limit (>10 cycles),and ,

the stress for failure in 100 cycles (from the lowest curve in Figure 1-16).

In both cracking locations, the stress state is outside the Goodman (Smith) curve for either HFF or LCF and for any load level of 90% or higher. The D implication is that initiation of ligament cracks in minimum strength material 3-5 3

l .

) .

( '

l is predicted, and given a ligament crack', initiation of stud-to-stud cracks is also predicted. Initiation could occur in less than 100 load excursions from

~

6 0 to 90% power or above and/or steady running for more than 10 cycjes (agg 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br />) at 90% power or above if the minimum material properties are as.

At 110% load, overload failure could occur in both locations with

) 'sumed.

minimum strength material since the peak total stress is 33 ksi compared to 32 ksi minimum thick-section ultimate strength. The fact that few blocks that have run at 110% load have cracks at both locations is indicative of higher-than-minfrum material properties and/or conservatism in the analysis.

)

The stress components in Table 3-2 are believed to be best availab1'e estimates from stra'.1 gage readings or conservative analytic scaling to key locations from gage readings, except for the preload. The preload gage read-

) ings (Gage Nos. 3 and 13) were used directly without scaling, even though the stress due to preload is probably higher near the stud hole in Crack Locations 1 and 2. This unconservative preload estimate partly compensates for conser-vative adjustments to the thermal stresses and for the conservatism inherent j' in t'h'e analysis of cracked ligaments with a plane strain model. This analysis is also conservative for engines that operate at lwer temperatures and/or pressures.

Other than determining the scaling factors S get from gage locaticns

). to crack initiation sites, the analytical models wera used only for insight.

Reasonably good agreement between all available experimental and analytical results was obtained for several 2-D finite element and hand calculation models. However, in such a complicated case, with interacting effects of

) clearance gaps, 3-D geometry and loading, friction, component-to-component distortion interactions, and relatively uncertain material properties, the experimental results are judged to be more reliable than the models.

)

)

3-6

) -

(

_ )

I

.O

'l TABLE 3-2 iO CONSERVATIVE ESTIMATES OF STRESS NORMAL I

TO CRACK FACE AT CRITICAL LOCATIONS

O .

lO Stress (ksi)

Pressure Range Load Preload Thermal Level Location Experimental Experimental Experimental Analytical 5

g 4

i TDI Gage No. 3 8.2 ,

-- -- 5.0 100

O 10.5 3.1 -- 90 l 5.0

~

L1gament TDI Gage No. 4 10.5 -- 3.3 100 10.5 3.6 -- 110

.',i ,

( ___

o

~

Block top at 8.2* 14.1 5.9 -- 90 i stud hole 8.2* 14.9 -

6.3 9.5 100 i Location 1 8.2* -

18.8 6.9 -- 110

'O 9.2 3.4 90

4.3 --

Between FaAA Gage No.13 4.3 9.7 3.7 6.06 100 l

studs 4.3 12.2 4.2 -- 110 l (for cracked 4.3* 16.4 6.1 90

• 'g ligament) Block top at --

stud holes 4.3* 17.3 6.6 10.85 100 Location 2 4.3* 21.8 7.5 -- 110

• unconservative 0

3-9 0 .

N

O

~

/ l

O 5 l- 1 E

• O O

~

% J i

O Q .

c i

w s.e_. _

t _ , =

!O 7 _N h f e

. [ T .

i a

O I

. O , '

O i-a$

E E

4 = "I =

5 y =-

= c O i W

5 I 5 ZG a l

O .

O 5 h

lO I

s ' .; -

[

I Q .. O g -z. .

j i s

!o O  !  :

e O

i li=

J I l o

i l

l

' FaAA-84-5-4 O

l .- -___- - ._ - _. _-__--_ _ - _ - . _ - . - _ _ _.

t , ,

O

) .' -

e i a a e a i i e a i i i o 0 (max.)

2 .

Q

) 16000 -

- 0 (min.)

l 3

1 Transverse +

! 12000 -

Gage 13 -

- e - s .

i O _

'g

) E 8000 -

uJ <

! E 2 -

M ,Ft _

^

4000 -

Longitudinal * -

C.; Gage 11 a (max.)

a (min.)

O _

D

-4000 - -

i I

f f f f f f f f f ,f "i I I t o o o o o o o o o o o o o o o o o o o o o o e o o o o N v e N v @ O C o e o e o

• N M

• N O " " N N O o T Torque on Torque on Cylinder +5 Cylinder +6

Engine of f  :  : Engine on r PRELOAD (ft-lbs) LOAD (kW)

Figure 3-6. Principal stresses vs. load for Gages'11,12 and 13 (located between studs.

p

• Principal stresses are located within 15' of gage axis. '

! Fe A A-84-5-4  %

l

~

D . .

r.

O 2c -

3 I O '

n Q

]

n W -

O J3 L

J e o },!'

s s

1. b O

, s' 55 E  %

~

i U$ "

1. 0 S

• 3" Tu v EfK g

~h*

I r f

),_ - A, >

I a g .

es

. 38, I 9 aa be

.s \ ,, _

OM

\ I e

e 1 1 \ __ y VA Q/ -

I m

D

, - a O

a S

O h ,

?a C

O L D m Fa A A-84-S-4

lI l:liljl ,i llJ t .

d F' F

e H12 I . '

0 3 .

g i

n F %a

, ~ k c

d F 0- a a H 01 ;- r c

lo d t.

a d n

% o a e 0 l l

o m a

1 g 1  %

0  % i l

- F 0 0 C 1 9 ) r ELF F 0 i

s f o

I C F' 2 k

(

L H m.

i m S r g

' 5 S a d E i F%a C0 R d L9l o T )

S h t

N i m

A S b )

s E (

al e M n vc a t y m

) d uc c s o o

f e 0 G F 0 I

l C 0 ec vy 1

1 .

L( r u c 3 1

c9 -

0 3 F1 F> e H( r u

g i

r

~

- - - - - - - ~ - _

0 4 0 6 2 8 4 O 2 2 1 1 n 5 ,mnEF* u oz F<zEuF:< t m5 eaa b-ll l llI ' lll . ll

lll e

. - - - - - ~ - - - ~ ~

7.

d a

l o

F F 1 0 g n

H1 0 i

k I

e 3 c a

v r d c a d l

o s. _

d s .

' d F 0 a o a F 0 o t o H1 l d

l

• F% u v  % .d F0 t s

0 a H9 ) _

1 o i r l s o 1 d k f

(

F C

% ao g 0 m I

L 0

0 l 2G a r

S g m 1  % E a v- F 0 C 9 F

T I i d

L F S )

h mC N t EL A i

m E S

(

M n

) a v s in d

ele o o

vc G ry 0 I u c g 1

c .

4 F 0 e 1 C 01 v 3 L( r

)

u s e r

% c60le F 1 c i.

u g

F> y F H( c v

O 4 0 8 2 8 4 0 2 2 1 1 Cax$ .'mf u oze<z@F s o

?$$5i v

j(1 ll l I

l

- l 0

(

O 4.0 FRACTURE AND FATIGUE LIFE EVALUATION 4.1 Block Top Crack Initiation Damage Model This section analyzes the initiation of block top cracks. A key part O~

"of this analysis is the observation that operation at 90% to 100% of rated load is in the f atigue initiation region of the Goodman diagrams for minimum

-trength cast iron as shown in Figures 3-13 and 3-14 However, the normal variability of material properties could result in no crack initiation under 90% t 100% 1 ad perating c nditi ns. The 110% load point is far into the O region where initiation is expected with minimum strength cast iron, and it is clearly more damaging relative to 100% load than 100% load is relative to 90%

- load. At load levels of 80% or less, the minimum acceptable Class 40 material probably will not initiate HFF cracking and will require greater than 100 startup cycles to initiate LCF cracking.

To predict the minimum amount of additional service before crack re CF and ex eked to occur between stud holes, it is necessary to consider bo O-HFF an the current level of damage. If a particular block has erated for a substantia 'eriod of time without initiating ligament cr ;s, this provides an estimate o e minimum LCF and HFF damage requir for ligament cracking in that particular ock. Since the F.a!- 'nc a rnatinn stresses for ini-

O tiating ligament cracks o for initir. N  % stud cracks in the presence of ligament cracks are alm' t the same ..upare Figures 3-13 and 3-14), at least the same cumulative dama w be r equired to initiate stud-to-stad l cracks af ter the ligament crack in iate and grow to a significant depth as

-vas experienced in producin he ligamen cracks.

lO i

l A conservativ ay to estimate the curre level of damaga is to divide 1 ne time at load nd the number of startups to d into categories, i.e.,

t,<70%, 70<Lc9 ., 90<L<100%, 100<L<105%, 105<Lc108%, an >108%. For LCF, the l

O oumber of artups that reached the particular load rsge tabulated. After , ,

an in etion revealing no ligament cracks, subsequent ope ti.on without l

l inspecticn can proceed 50 long as the number of additional start s in eac gtategory are less than the accumulated number at the time of the last inspe 4-1 O l 5

O l l

1 e' .

hite. The appearanc ofthemicrostructurii I- m" ro-poro ty and degenerate  ;

O e DG103 is au different than that th: ;r,ghe ;hd: :. fqn1 and nr.1n9

-- - N The presence of . aegenerate graphite microstructure has been shown to reduce the strength of cast iron significantly (4-4, 4-5]. Specific materials testing is required to quantify any degradation in fatigue or fracture pro-0 perties of the thick section block casting. A conservative projection of the cracking potential of other engines was ebtained by extrapolating the ex-perience of DG103 and assuming that other engine blocks are of equivalent

' ** * * 'i 'l -

O 7 I A olock wtth no existing stud-to-stud cracks and material prope s sufficient etter than those of DG103 should be able to c ete the LOOP /LOCA requireme without any cracks as deep as the .2-inch crack in O 0G103, while continuing to v normally. Engines wi etter material or more favorable operating parameters or ands w have less damage. Therefore, these calculations indicate that peri spection for radial cracks between the s'tud holes, in combination th site-spec analysis of operating his-O' tory, material properti , and operating stress, ld assure that block ne's ability to cracks will not w to a size which will impair the e provide th wer levels required during a LOOP /LOCA.

,4.3 O } Block Material Properti:.s in stud hole Th comparison of stresses under full load opera regions of the ock top with the ultimate strength fatigue resistance of Class 40 gray cast on [4-63 shows that fa e cracking of the ligament O region can occur in mater with minimu pecified properties. On the other hand, if the ultimate strength atigue . resistance of the block are above average for the Class 40, gue er initiation may not accer without a large number of cold st s and extended ope ion at or near full power. The

'O cumulative damage dex approach provides a meth to quantify the ef fects of alternative ine usage.

Clearly the block top area is not so conservatively igned that acking will never occur, nor is it so highly stressed that ligame cracks lO L CL_

4-5 01 [

( 6

.O -

N w ill a s occur.' under these circumstances, the cumulative time / load <

-Q~

1 evel/ number f cold starts at which cracks develop and the rate at which t p ogress is str ly dependent upon the materials properties of the cific e isting. .

'aray cast iron is ticularly sensitive to material properties degra-

.O l dation due to small amounts tramp elements, like 1 d. The ultimate ten-sile strength of thick section tings has be shown to be reduced by as nuch as 80% of its normal value by pres ce of greater than 0.01% lead O_

ll4-4, 4-s]. These tramp elements reduc strength and ductility by modifying she normal structure of the graphi flak to produce degenerate graphite j structures with interconnected manstatten cicular) structure. '

The presence of v extensive degenerate g hite microstructures can O be identified by co entional metallographic examinat n. In-situ polishing l of block top sur ces. light etching and taking of cellul e acetate (plastic) replicas fo microscope examination provides a non-dest etive method to detect erely degenerate casting stuctures. Small pieces o lock material O so be removed for more detailed meta 11ography and for antitative yan semical analyW t: d;te,; Uiu presence or unassiraois i.rau, ui .. n q l

4.4 Cm n11ery Cracks O

An inspection of the emergency diesel generators at Shoreham revealed crack indications in the cam galleries of all three SNPS engines. These indications were of varyir.g lengths, the longest Deing '4 1/2 inches long and 0.375 inch deep in DG103. A typical cross-section of the can gallery is l

)

O displayed in Figure 4-2, indicating the crack region. jTDI installed strain 1

I gag on an experimental engine (DSR-46) at the cations of the crack-like defects recorded the dynartic strains in a run e ine. The strain gage data were re d by TDI to obtain the me and alternat stresses [4-7].

These stresses are roduced here in able 4-2. For the pre nt ar31y ,

stresses obtained at the e No. location at 100% load were used.

h A fracture mecha s ana s was performed to evalu the fatigue 9 4 l

4-6 l

9 '

5

O-(- Section 4 References ,

O 4-1 A. Yuen, et al., " Correlations Between Fracture Surface Appearance and Fracture Mechanics Parameters for Stage 11 Fatigue Crack Propagation in

~

Tt-6 AC-4V, Metallurgical Transactions 5, p. 1833, August 1984 2 -Stone. and Webster Letter of 12/15/83 to LILCO,

Subject:

Two-Year O Operating Cycle Emergency Diesel Generators SNPS.

4-3 C. F. Walton and T. J. Opar, Iron Castings Handbook, Iron Castings Society, Inc., 1981.

4-4 C. E. 8ates and J. F. Wallace, " Trace Elements in Gray Iron," American O Foundrymen's Socienty, Report of Research Project.

4-5 C. E. Bates, "Effect and Neutralization of Trace Elements in Gray an'd Ductile Iron," Ph.D Thesis, Case Western Reserve University (1968).

4-6 L. E. Tucker and D. R. 01berts, " Fatigue Properties of Grey Cast Iron",

SAE Paper No. 690471, SAE Transactions, Vol. 78, 1969.

g 4-7 Dean T. Ripple, "R4-L6 Cam Gallery Strain Gage Test," Transamerica Delaval Inc., Engine and Compressor Division, R and D Report, FR-01-1983, Rev. June 21, 1983.

!f ,' s f

C

'O i

g

. 4-8

)

3 '

i .

'P c

!O l

• I

,O -

A O

1.1" 1

A O

0.375*

Y

,m 4.5" >

C. '

Figure 4-2. Cam gaklery crack shape.

O

O

, I l

0 i

O -

1 1

i FaAA-84-5-4 0 .- -

y

g.

.l O- -5.o conclusions Ase RECCIMENDATIoMS

- Review of operating experience with TOI R-4 and RV-4 cylinder blocks indicates that precautions are necejssary to avoid the potential consequences

.of block top cracking between stud holes of adjacent cylinders. The results O of strain gage testing, combined with two-dimensional analytical models of the bloc't top and liner and cumulative damage estimates, provide the following conclusions and recommendations:

O. 1. Initiation of cracks in the ligament between stud hole and liner counterbore is predicted to occur after accumulating operating hours at high load and/or engine starts to high lead. These cracks are benign

' because the cracked section is fully contained between the liner and the region of the block top outside the

,0 stud hole circle. Field experience is consistent with

both the prediction of ligament cracking and the lack of innediate consequences.

]

1

^

a 2. The presence of ligament cracks between stud hole and liner counterbore increases the stress and the proba-

.f O- bility of cracking between the stud holes of adjacent cylinders such that stud-to-stud cracks are predicted to initiate after additional operating hours at high 4

load and/or engine starts to high load. The deepest 1

measured crack in this region (51/2-inch depth) did not degrade engine operation or result in stud loosen-ing.

3

3. The apparent rate of propagation of cracks between stud holes in the DG103 block at SNPS, when compared with the LOOP /1.0CA requirements, indicates that blocks with ligament cracks (e.g., DG101 and DG102) are predicted to withstand a LOOP /LOCA event with suffi-O cient margin provided that: (1) inspection shows no stud-to-stud cracks detectable between heads, and (2) the specific block material of DG103 is shown to be sufficiently less resistant to fatigue than typical gray cast iron, Class 40.

1. 4.- The block tops . of engines that have operated at or above rated load should be inspected for ligament cracks. Engines such as those at Catawba and Grand Gulf that are found to be without ligament cracks can be operated without additional inspection for combina-tions of load, time, and number of starts that produce O

5-1 9

db l - - . _ - . .. .

P 1

less ex'pected damage than the cumulative damage prior The allowable engine usage D to the latest inspection.

without repeated inspection can be determined from cumulative damage analysis.

~

5. The blocks of engines th'at have been operated at or .

- above rated load without subsequent inspection of the block top should conservatively be assumed to have 3 cracked . ligaments for the purpose of defining inspec- l I

tion intervals.

6. For blocks with known or assumed ligament cracks, the j absence of detectable cracks between stud holes of g adjacent cylinders should be established by eddy  !

current inspe'ction before returning the engine to '

emergency standby service after any period of opera-tion other than no load. If crack indications are

• found, removal of the adjacent heads and detailed inspection and evaluation of the block top are neces-sary. In addition, it.is necessary to ensure that the l microstructure of the block top does not indicate inferior mechanical properties.

7. Engines that operate at lower maximum pressure and temperature than the SNPS engines (e.g., San Onofre) may have increased margins against block cracking that 3 could allow relaxation of block top inspection re -

quirements. Modifications to other parameters such as increased liner-to-block radial clearance will reduce stresses, and site specific analyses of such modifica-tion could also permit relaxation of inspection re-quirements.

?

D ct ,

J J

5-2 l

) '

[I l

k Appendix:

COMP 0 MENT DESIGN REVIEW TASK DESCRIPTION O CYLINDER BLOCX Classification A Part No. 03-315A Completion Date 3/20/84 PRIMARY FUNCTION: The cylinder block comprises the framework of the liquid cooled engine and provides passage and support for the cylinder ifner. The O block must provide cooling water passages, provide bores to support the cam shaft assembly, and react the dynamic loads from the cylinder firing pressure and valve assemblies. For the RV engines, the cylinder block is intercon-nected with an engine crankcase which supports the camshaft and associated

- bearings. Although these are separate parts, their generic function is simi-lar to the cylinder block of the R-48 engines and will therefore be evaluated O as a unit. The liner itself forms the walls of the combustion chamber con-taining the high temperature gas pressure and must provide a guide for the piston motion while reacting skirt side forces without excessive wear or scuffing.

FUNCTIOEAL ATTRIBUTES:

O 1. The cam gallery bearing supports must be designed to maintain concentricity during service and have sufficient structural strength to react the cam / valve train loads without fatigue cracking.

2. The support of the cylinder liner must maintain tight seals, react pressure and stud loads without unacceptable distortion and maintain O sufficient load distribution to preclude excessive cracking in the liner counterbore (landing) due to combined thermal, ga> pressure and preloaded stud induced states of stress. The cylinder head stud thread configuration is important in determining stress concentra-tions and stress distributions.

3 3. The cylinder liner itself must be sufficiently hardened to resist unacceptable wear associated with piston ring action and maintain adequate contact with the block counterbore to prevent high cycle contact stress and fretting. In addition, the compression of the head to the cylinder liner must be sufficient to avoid. axial fret-ting of the liner within the counterbore but not so great as to cause failures of the cylinder block liner landing.

3 .

4. The cooling water distribution within the block must be sufficient to preclude overheating of the block and liner and must maintain proper flow conditions to minimize or avoid cavitation or corrosion damage to the liner.

O A-1 0

3D

O .

SPECIFIED STANDARDS: None

.'O

. EVALUATION:

1. Review information concerning previous cracking and distortion of the cylinder block and liners of the R48 and RV engines..

O- 2. Review liquid penetrant inspections of cylinder block in the head I e stud and liner counterbore regions of the SNPS OSR-48 engines. l

3. Evaluate the steady state and alternating stresses in the liner landing / head stud region and compare these to yield and endurance ,

limits for appropriate materials. This examination riust consider O variations in head stud thread geometries and preload torques.

4. Evaluate the state of stress in the liner in the landing / axial seal

region due to gas pressures, thermal growth and head clamping forces

- . and compare to normal fatigue properties for liner material.

O 5. Evaluate critical flaw size and rate of crack growth considering combined head stud loads and thermal stresses for cracks located between head stud holes and cylinder block counterbore diameter.

l

6. Evaluate critical flaw size and rate of crack growth for. cracks eminating from the corner of the cylinder block landing and counter-bore diameter.

ff

7. Evaluate the loading produced on the bearing supports in the cam gear gallery and verify the structural adequacy of the design.

8. Review the insper. ion of the sampled SNPS cylinder lines following 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> at 100" load for evidence of unacceptable scuffing, cor-

O rosion, cracking, or scoring.

9. Review information provided on TER DR-220.

! REVIEW TOI ANALYSES:

1. Review any TDI analyses which consider stresses created in the liner i counterbore area and any design changes which relate to geometry or .

material .

l IN m MATION RE00! RED: J

o Manufacturer's drawings of R48 and RV cylinder blocks and liners, 1.

including material specifications and historical design changes.

l 2. Gas pressures and temperatures for R48 and RV engine designs.

!g . 3. Cylinder head stud drawings and torque specifications.

A-2 i

]

y

)

i 1

l

[ s

)

1

4. Cylinder head stud drawings showing design changes.

. -5. 1.iquid penetrant inspection of cylinder block counterbore (landing) on SNPS engines.

6s Cam shaft loads due to rocker arms, pushrods and valve springs.

3

~ .

J.'

J J

)

A-3

)

32.