ML20045B279

Forwards Correspondence on Plant Low Pressure Turbine

ML20045B279
Person / Time
Site:	Pilgrim
Issue date:	06/14/1993
From:	Eaton R Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
NUDOCS 9306170125
Download: ML20045B279 (44)

Text

,

(* *Fou r-

.A a

S UNITED STATES -

3 NUCLEAR REGULATORY COMMISSION

,8 WASHINGTON, D.C. 20555-0001 June 14, 1993 Docket File 50-293 MEMORANDUM FOR:

File FROM:

Ronald B. Eaton, Senior Project Manager Project Directorate I-3 Division of Reactor Projects - I/II Office of Nuclear Reactor Regulation

SUBJECT:

CORRESPONDENCE ON PILGRIM LOW. PRESSURE TURBINE Please put the attached subject correspondence in Docket file 50-293.

,o Alb Ronald B. Eaton, Senior Project Manager Project Directorate-I-3 Division of Reactor Projects I/II Office of Nuclear' Reactor Regulation

Attachment:

As stated

.i

~

l l

i l

i

15005o, j

i i

. t.

i Fus. s r~

' ), _

c., y,

g IN!O nI i:,a u.

'f/ p 9306170125 930614

'(

PDR ADDCK 05000293 P

PDR l

June 14, 1993 Docket File 50-293 MEMORANDUM FOR:

File J

FROM:

Ronald B. Eaton, Senior Project Manager Project Directorate I-3 Division of Reactor Projects - I/II Office of Nuclear Reactor Regulation TlBJECT:

CORRESPONDENCE ON PILGRIM LOW PRESSURE TURBINE Please put the attached subject correspondence in Docket file 50-293.

/s/

i Ronald B. Eaton, Senior Project Manager Project Directorate I-3 Division of Reactor Projects - I/II l

Office of Nuclear Reactor Regulation

Attachment:

l As stated

{

DISTRIBUTION w/ attached cover letter & its attachment:

Docket File DISTRIBUTION w/ attached cover letter only:

NRC & Local PDRs PDI-3 Reading Slittle

.REaton WButler B.D. Liaw l

OFFICE PDI,-3)LA PDI/3:PM PDfMb/s j kle REhn:dt WButlek j

NAME DATE (p/10/93 (7 / /93

[ /u//93 0FFICIAL RECORD COPY

'/

l

-DOCUMENT NAME: A:\\PIDOCKET.MEM

=

k f

f

.s 3

m ll GEPower Generation a-BOSTON EDISON COMPANY D'NJll7D' TURBINE S/N 170X341 reoocommerom Po 30 5043 L.P. "A" ROTOR #FD3518

$7,*3,'*l'l$*8'5"##3 INSPECTION RECOMMENDATIONS F

l I

May 11,1993

[

Mr. Ed Kraft, Jr.

V. P. Nuclear Operations & Station Director Boston Edison Company Pilgrhn Station RFD #1 - Rocky Hill Road Plymouth, MA 02360

Dear Ed:

In April 1993, a complete in-service rotorinspection was performed on the above rotor. The.

inspection included ultrasonic examination of the shrunk-on wheel bore areas, including the i

wheel axial keyways. Based on the wheel bore mspection, it is Gene al Electric's recommendat on not to return the rotor to service without either removing the 7th stage, generator end wheel, or pre-warming the L.P. rotors prior to startup.

Numerous diamamions have taken place with your station's engineering personnel since these recommandadons were delivered. GE's main concern is to optimire reliability and the prevention of a wheel burst The analysis of the wheel bore areas is based on criteria reviewed and accepted -

by the Nuclear Regulatory Commission (NRC), and is consistent with all other analysis done by

. GE on operating nuclear units.

t

)

I

-i

?

1 Mr. E. Kraft, Jr.

2 May 11,1993 GE Power Generation will continue to make personnel available to discuss this issue with Boston Edis'on before the unit is returned to service. If you have any other questions, please call.

Ve:y truly yours, William C. Roode Operations Mower Power Generation Services New England WCR/mm

/beco0506. doc cc: R. Fairbanks. BECO W. Clancy, BECO L. Schmelling, BECO C. Jungclas, BECO M. Potkin, BECO T. Burns, BECO W Rothert,BECO Wozniak, BECO 1

]

C6 12 93 CC:18 U. S t i. P. C PEGICal 1 4Ill 002 d MAY 11 '93 11:26 PILGRIM NRC RESIDENT P02

+.

I

>N1 3t50 Almoden Erpreneway-,

M L 9 1(5 ren Jose. CA 9511e Ibesil Plant Operosions 66 re,e Mmer Reed (42) ne-ezco May 5,1993 5* 2 tax. ooers w+euw DAR-93-012

.aem ruos SIR 93-047, Rev. O (215) 964-8886 ru nm m.s=

Mr. John Gerety Boston Edison Company 25 Braintree Hill Office Park Braintree, MA 02184

Subject:

Evaluation of the Pilgrim Unit 1 Iow Pressure Turbine Rotor 7th Stag Shrunk-on Disk Dear Mr. Gerety-This letter documents results of Structural Integrity Associates' (SI) evaluation of th stage shrunk-on disk for the Pilgrim Unit I low pressure turbine rotor. This evaluation General Electric Company (GE), and an independent evaluation by SI to assess the remaining life of the rotor, considering the parametri of several variables on the inspectioninterval/ time to fsilure for the 7th stage shrun These results indicate that an inspection interval in excess of 4 years can be j Resuhs of the evaluation performed by SI are summarized below, it should be noted these results were based on disk dimensions and a disk bore centrifugal stress of 42 1800 rpm, provided to SI by Boston Edison.

We can supplement this deterministic evaluation with a probabilistic fracture mechanics (PFM) analysis, to demonstrate fur benefits of using a statistical approach.

Intmduction

~

A recent ultrasonic in-service inspection of the low pressure turbine shrunk-on disks a Pilgrim Unit I was performed by GE. A summary of these inspection results, a[

comparison to prior inspections of the disks are contained in Table 1.

GE presented recommendations to Boston Edison based on these results as shown in recommendations call for a removal of shrunk-on disk "7GA" at the minimum, to assure an inspection intervallarger than 2 years, for a 75 *F startup temperature. The GE evalu concluded that the smallest critical crack depth was 0.3 inch in the keywry of shrurik-on 7GA.

The deep-seated Fracture Appearance Transition Temperature (FATT) for thi location was +12'F. Adding 35'F to this FATT value for conservatism, and u generic GE lower bound fracture toughness curve (Figure 1), the fracture toug{

estimated at this location is 105 ksiv' inch; from which the resulting critical cra i

inch is derived.

3

./?

i h 1983 _/

y,,,,,jo y,,,,.,,,g,,j,,,,,,,,5 E m uc.,cr [ N I993

g u. y on a

u. 5 u P. C PEGIOi 1 i !!i O'C I IHY 11 'ya 11:27 PILGRIM NRC RESIDENT P03 Page 2 Mr. J. Gerety-hiny 5,1993 DAR-93-012/ SIR-93 047, Rev. 0 Statistical fits of crack growth rates used by GE based on inspecdon d t 1

either upper bound literature data or by usinE the cr a a, which included NRC.

y e

The evaluation performed by Si consisted of three phases. Phase I consisted form stress analysis to estimate the distribution of hoop stress as function c osed-from the disk bore due to centrifugal and shrink-fit effects.

In Phase II, a fracture factor (Kj) at the crack tip as a function of crack size models.

Finally in Phase DI, a parametric study was performed to assess the eff several key variables on critical crack size and time to failure. A brief description evaluation phase is given below followed by a summary of results and conclusions 1

Phase I:

Stress Analysis i

A closed-form stress analysis of the 7th stage shrunk-on disk was performed to est distribution of hoop stress as a function of radial distance from the disk bore d centrifugal and shrink-fit effects.

3 Cent rifugal Stresses:

A hoop stress value of 42.4 ksi at the disk bore due to rotation at 1800 rpm was provided to $1 by Boston Edison, based on GE input. The radial distribution of hoop stress due to rotation was estimated from the well known classical schnion fo rotating cylinder (plane strain) as given below (Reference 1):

t o*,13-2v)p o g

(1.py)

(1)

_ (3-2v),,

(1 -v)8 2

7

where, u is poisson's ratio p is the material dernity t> 2s the angular speed in radianshee rg as the inside radius of the disk r, is the outside radius of the disk, and, r is the radial location at which hoop stress is calculated.

For inside and outsid= disk radif of 16.5 and 43.5 inches, measured by Boston Edison radial distribution of centrifugal stress at 1800 rpm was computed and scaled to yie i

bore stress value of 42 4. The resulting distribution due to centrifugal effects alo f

.cc,12 9; 07:19 U.

i :. P. - ;E M C#4 1 5 !f1 004

-~

MAy 11.'93 11:29 PILGRIM NRC RESIDENT P01 Page 3 May 5,1993 M r. J. G eret g DAR 93-012/ SIR-93-047, Rev. O in Figure 2. Centrifugal stresses at any other speco (rpm) can be obtained by scaling these tesults to the ratio of the square of the rpm values; f.e., by a factor of (rpm /1800)2, Shrink-nt Stresses: Stresses developed due to shrink-fat of one cylindrjcal member onto another arc a func. tion of the contact pressure (p) developed between the two ruembers.

The contact pressure can be computed for a given radial interference and geometry (Referencc 2) as given below:

8 8

r p,, E6 (c -b9(b -a )

(2) h 8 a 2b (c _,2)

where, E is the modulus of clusticity, 6 is the radial intettercnce, and, a, b, c are the inside, middic and outside rac111 of the shrink fit members.

A radial Interference of 0.0345 inch was used for the 7th stage shrunk on disk based on the following diameters provided by Boston Edison:

Disk inside diameter = 32.998 inches Shaft outside diameter = 33.067 inches Values used for radii a, b and c were 0.0,16.5 and 43.5 inches, respectively. A zero shaft l> ore radius was assumed for conservatism, yielding a contact pressure of 26.851 ksi.

' Die hoop stress distnbution is then given by the well-known Lame's solution:

2

'3[

(3) b o*. p 2c -bt E

t r

j

whert, b and c are the inside and outside radii of the outside disk, and, r is the radial location ut which hoop stress is calculated.

The resulting stress distribution due to shrink-fit is shown in Figure 2. The hoop stress computed at the bore of the disk was 35.876 ksi.

Combination of Centrifugal and Shrink-fit Stresses:

Whue the centrifugal stresses are independent of the magnitude of the shrink fit, the shrink-fit stresses reduce with increasing rpm, because the radial displacements at the disk ID are proportional to the square of the angular velocity, it was reported that the GE design UK

mer ceco

v. s n s..

pesicni 1 i its oos nsy 11 '93 lis 30 PILGRIM NRC RESIDENT P02 Page 4 Mr. J. Gerety_

May 5,1993 DAR 93-012/ SIR-93-047, Rev. 0 results in zero shrink-fit stresses at 120% of rated speed (2160 rpm). With this design condition and the inverse relationship of shrink-fit stress with the square of angular velocity, shrink-fit stresses can be calculated at any speed and then summed with the t~ ropriate centrifugal stresses at that speed to yield a resultant stress distribution, as & o e Figure 3, for speeds of 0,450,900,1350,1800 and 2160 rpm.

~

The combined disk bore stress at 1800 rpm was 53.362 ksi; i.e., a centrifugal stress of 42.4 ksi plus a shrink-fit stress of 10.962 ksi.

Stress Concentrating Errect of Keywiny:

A stress concentration factor (K ) of 2.2 was reported by Boston Edison for use with the GE i

keyway design. This K value has also been computed in analyses performed under EPRI i

project NP 3340 (Reference 3). ne decay in the peak K value of 2.2 with distance from 1

the keyway surface (from Reference 3),is pven below:

r,(x) - 1 +(r,-1) 0.25(- S--)* + 0.75(R*x)*

(4)

R +x

where, R is the keyway depth (0.625 meh), and, 2 la the radial distance from the keyway surface.

As illustrated in Figure 4, the effect of K diminishes rapidly, reaching a value of only 1.06 g

at 1 inch from the keyway surface.

Combined centrifugal and shrink-fit stresses with the K effect superimposed are shown in i

Figure 5.

Phase II:

Fructure Mechantes Analyses The primary objective of the tracture mechanics analysis was to generate curves of stress intensity factor (K ) versus crack size (a), which then permit determination of a critical crack i

size (a,) for a given fracture toughness (K ). The equation used by GE to compute Kj wa c

ic compared with other more appropriate solutions using SI's in-house fracture mechanics code PC-CRACK (Reference 4). The PC-CRACKW computer program was developed under the S1 OA program, which meets the NRC code of federal regulations 10-CFR 50, Append B, ANSI /ASME NOA-1-1986, and NUREG-4640 (for computer software). A more accurate determination of K results in a reduction in the inherent conservatism in using the GE 3

equation.

i

c,5 1; r creo u.s ti.c.

pes t or: 1 i irt 006

.-MAY 11 '93 11831 PILGRIM NRC RESIDENT P03 Page 5 May 5,1993 Mi, J. Genty, DAR-93-012/ SIR 93-047, Rev. O GE K Solution: The K solution used in GE calculations is given by the equation:

3 i

K ;- 2c/5/8" + a N) f whe re, a is a uniform stmss normal to the crack tace,

$/8-inch is the keyway radial depth, and, a is the radial extent of the cruck measure from the keyway surface.

De above GE equation can be expressed in the tollowmg well known form for an edge-crack m a uniform stress field (Figure 6a):

K, - 1.120/na,

@)

whe re, crt s the effective cd e-crack depth = $/8 + a, as given above.

i a

S De GE crack model (equation 5), embodies the following conservatisms:

- q (i)

Tbc edge-crack model represents and upper bound computation for K i

because this model (Figure 6a)is far more compilant that the actual geometry I

which is a radial crack emanating from the ID of a thick hollow cylinder or a hole in a plate (Figure 6b)

(ii)

The uniform stren field used in the GE equation is cver-conservative becauf e it does not incorporate the radial decay in hoop stress with increasing distance from the bore.

"Kj versus a" calculations for the GE model are shown in Figure 7 for crack depths of upto 1 inch.

SI's Ki Solutions:

To incorporate the reduced compliance associated with the relatively large size of the shrunk-on disk (OD of 87 inches) the crack model for a hole in an infinite plate (Figure 6b) was used in the S1 calculations (Model I in SI's cornputer program PC-CRACK, Reference 4).

The stress concentrating effect of the keyway (K ) was intcorporated using two t

l upproaches:

i

)

05129; 07:21

' i. s u. R >- PEGIOu 1 Ilu 00'

, I'lHY l1

'f3 11s31 PILGRIM HRC RESIDENT PO4 Page 6 Mr. J. Gerety_

May 5,1993 DAR 93-012/ SIR 93-047, Rev. 0 (1)

Model A:

As shown in Figure 8a, for a hole in a plate, the K, effect was implicitly accounted for by using the unconcentrated stresses (Fjgure 3) and an effective crack length of(5/8+a); the hole radius used was the inside radius of the disk (16.5 inches).

(11)

Model B:

As shown in Figure 8b, for a hole in a plate, the K, effect was explicitly accounted for by using the concentrated stresses (Figure 5) and a crack length measured from the surface of the keyway; the hole radius used was the inside radius of the disk plus the keyway depth (17.125 inches).

"Kg versus a" calculations for the Si crack models att shown in Figure 7 along with the GE model for comparison. The SI crack models which more accurately represent the actual crack geometry result in a significant reduction in Kj. For a fracture toughness of 105 ksM' inch the critical crack size predicted by the GE model is 0.34 inch compared with 0.5 inch for the St Model A and 0.56 inch for SI Model B.

Phase III:

Parametric Study (EfTect of Key Variables)

In this phase the effect of several key variables which govern the basis for determining an in-service inspection Interval were studied. Calculations were performed for the 7th stage disk 7GA, the critical disk controlling the in-service inspection interval, as identified by GE.

An in-service inspection interval (ta,,) can be computed based on the time.to-failure (trau) for the subject disk using the equation:

I, tpjSF (1) where SF is a safety factor, and tren is given by the equation:

t, - (4 %)

u (3)

(da/dt)

where, "cr is the critical crack size, nj is the initial crack size, and, da/dt is the crack growth rate.

Based on the detection of"non-measurable" indications during GE's inspections of disk 7GA and the characteristics of the GE inspection system, an initial crack size, ag, of 0.25 inch was assigned to disk 7GA by GE.

O

05 12 E o":21 U. S t i. P. : PEG I C# 1 16 lu 00:3 MAY 11 '93 11:32 PILGRIM HRC RESIDENT P05 Page 7 Mr. J. Gerety.

May 5,1993 DAR-93-012/ SIR-93 047, Rev. O For the assigned initial crack size of 0.25 inch, the time-to failure (equation 8) is then dependent on the critical crack size and the crack growth rate, which is primarily governed by stress corrosion cracking (SCC).

Based on the forego!ng discussion, the following variables were identified by SI as having a significant effect on the determination of traD and therefore twp as given in equations 8 and 7, above:

Effect of 35*F Adjustment in Deep Seated FATT Effect of Fracture TouE ness Variability

=

h Effect of Prewarmmg Effect of Crack Growin Rate Stress Intensity Factor Model The efleet of each of the above variables was studied, one at a time, and the results are summarized in Table 3.

(a)

Effect or35*F Adjustment in Deep-Seated FATT From Reference 5, the best estimate deep-seated FATT values for disk 7 were:

Disk 7TA 9*F Disk 7GA 12"F These values were further adjusted for safety by adding another 35'F to the result, yielding FATT values of 44*F for disk 7TA and 47aF for disk 7GA. For the base case shown in Table 3, a lower bound fracture toughness (K c) of 105 ksVinch was estimated from Figure i

1, conesponding to the FATT value of 47'F for disk 70A, yielding a critical crack size of 0.34 inch and a time-to-failure of 1.5 years. If the 35'F margin were not included, the lower bound Kic estimate increases to 128 ksVinch, with an iticrease in critical crack size to 0.82 inch and a time to-failure of 9 5 years.

(b)

Effect of Fracture Toughness Variability If the menn value of Kje of 145 ksvinch were used, from Figure 1, instead of the lower bound value of 105 ksVinch at the same FATT of 47'F the critical crack depth and time.

to-failure beco:De 1.22 inches and 16.2 years (see Table 3). Derefore an inspection interval of 4 years with a safety factor of 4.05 would be justifiable, based solely on mean venus lower bound fracture toughness data.

INC

05 12 cc 0~; 22

u. s it. P. C PEG 10ri 1 i 111 CO

. MAY 11 '93 11:32 PILGRIM NRC RESIDENT P06 Page 8 Mr. J. Gerell May 5,1993 DAR 93-012/ SIR-93-047, Rev. 0 (c)

Efrect of Prewarming The GE inspection intervals (Tabic 2) are based on startup teroperatures of 75'F,80*F, 90*F and 100'F to address brittle failure of the disks during cold startup. Prewarming of the rotor prior to cold startup results in significantly higher toughness and critical crack sizes as shown in Table 3. At 80*F the lower bound K estimate is 110 ksVinch and the cr ie crack size increases to 0.46 inch, with a time-to failure of 3.5 years. At 90'F the lower bound Kjc estimate is 115 ksVinch and the critical crack size increases to 0.54 inch, with a time-to-failure of 4.8 years.

(d)

Crack Growth Rate The primary crack growth mechanism in low pressure rotor shrunk-on disks is stress corrosion cracking (SCC). Based on information provided in Reference 5, a crack growth rate of 0.06 inch / year was used for the. base case analysis (Table 3), for a reported st state operating temperature of 172*F at the 7th stage. This crack growth rate represents two standard deviations above the mean, based on GE inspection results, which included non measurable indications. For comparison, two alternate sources of crack growth data were used.-

(i)

Laboratory crack growth data representing a compilation of GE, Westinghouse and CEGB data (Figure 9)

(ii)

The SCC crack growth equation accepted by the NRC for use in the analysis of cracking in bores and keyways of LP turbines in nuclear plants (References 6 and 7), as given below:

In(da/dt) - -4.980 - (7302/I) + 0.0278o, (9)

where, T is the operating temperature of the disk in *R (*F+460),

o is the yield strength in ksi, and, y

da/dt is the growth rate in inches / hour.

An upper bound crack growth rate of 0.02 inches / year at 172*F was estimated froin the laboratory data (Figure 9) yielding a time to-failure of 4.5 years as shown in Table 3.

For a reported disk yield strength of approximately 120 ksi, the crack growth rate predi using the NRC equation at 172*F was 0.0164 inch / year, resulting in a time to failure of 5.5 years as shown in Table 3.

4

0". 12 9; 07:2; U. 5 H. P. '.' PEGIC# 3 1 1 1:1 010 MAY 11 '93 11: 33 PILGRIM HRC RESIDENT P07 Page 9 May 5,1993 Mr. J. Geretg DAR-93-012/ SIR-93-047, Rev. 0 (c)

Stress lutensity Factnr Model As discussed in the fracture mechanics analyses (Phase II) alternate stress intensity factor solutions to the conservative GE model were used by S1, based on the SI computer program PC-CRACK (Reference 4). "K versus a" results from these alternate models are plotted in Figure 7.

The " crack at a hole in a plate" model used by SI is a more accurate representation of the keyway crack geometry, incorporating both the added constraint imposed by the large disk size and the decreasing stress gradient with increasing radial distance from the bore.

Both alterflate models retain the conservatism of assuming an infinitely long edge crack (i.e., the totallength of the keyway). Further increases in critical crack size would be obtained with a more realistic assumption of finite flaw aspect ratio.

SI Model A:

For the crack model shown in Figure Sa (without the K effect), the critical i

crack size predicted for e lower bound toughness of 105 ksVinch was 0.54 inch, with a time-to-failure of 4.8 years; i.e., a critical crack size inr:rease by a factor of 1.6 over the GE K-solution.

SI Model B:

For the crack model shown in Figure 8b, incorporating the keyway stress concentrating effect, K, the critical crack size predicted for a lower bound toughness of 105 i

ksVinch was 0.56 inch, with a time-to-failure of f.2 years.

Summary &

Conclusions:

An independent stress and fracture mechanics analysis of the 7th stage shrunk-on disk 7GA was performed by S1, followed by a parametric study of the influence of several key variables on the recommended reinspection interval for this disk. Specifically the effects of the following variables were studied:

PJfect of 35*F Adjustment in Deep Scated FATT Effect of Fracture Toughness Variability

• Effect of Prewarming Effect of Crack Growth Rate Alternate Stress Intensity Factor Models This study revealed that substantial improvements in inspection interval could be achieved by use of one or rocre alternate parameters described above. In particular, more accurate modelling of the strets intensity factor combined with the use of an upper bound crack growth rate of 0.02 inch / year, based on literature data, is recommended. This approach yicids a time-to-failure of 14.5 years or an improvement in inspection interval by a factor of ten over the base case.

C6 12 93 CC:20 U. 5 t i. P. C PEGICt l 1 1 IN 011

. MAY 11 '93 11:33 PILGRIM NRC RESIDENT P08 1

Page 10 May 5,1993 Mr. J. Gerety DAR 93-012/ SIR-93-047, Rev. O Also noteworthy is the fact that um of more representative material properties (PATT, K c),

i instead of conservative lower bound values, result in a significant improvement in inspection interval. For example, use of mean fracture toughness alone results in factor of eleven improvement in inspection interval.

Although deterministic in nature, every effort was made in the foregoing analyses to utilize conservative assumptions on distributed variables ('.e., two standard deviations), along with the conservatism of assuming an infinitely long crack at the keyway.

In conclusion, the u a of a more accurate crack modelin combination with use ofliterature upper bound crack growth data, appears to be more than adequate to justify an inspection interval of 4 years, with a safety factor of 3.6.

Very t

yours, Reviewed and approve.d by; Darry A. Rosano, P.E.

P. C. Riccardella, Ph.D., P.E.

Senior Engineer President

[p

/

/

ygo/c L w av

05 12 93 O~ 24 8 i. 5 l l. P. '.- PEGICtl I itil 012 MRY 11 '93 11:34 PILGRIM HRC RESIDENT P09 Page 11 May 5,1993 Mr. J. Gerety DAR-93-012/ SIR-93-047, Rev. O References 1.

F. D*Isa," Mechanics of Metals",1975, Addison-Wesley Publishing Company, 2.

J. E. Shigley,

• Mechanical Engineering De sign",1972, 2nd Edition, McGraw-Ftdl Book Company.

3.

EPRI Report, NP-3340, Stress and Fracture Analysis of Shrunk-on Steam Turbine Disks,1984.

4.

Structural Integrity Associates, "pc-CRACK", Verston 2.1, 1991.

5.

Presentations by General Electric Co to Boston Edison, 4/29/93,4/28/93,12/7/92 and 3/8/92, on axial keyway cracking of Pilgrim Unit 1 LP rotor shrunk-on disks.

W. G. Clark, B. B. Seth, and D. M. Shaffer, Procedures for Estimating the 6.

Probability of Steam Turbine Disc Rupture from Stress Corrosion Cracking",

presented at Joint ASME/IEEE Power Generation Conference, October 1981.

[7.

~

F. F. Lyle, Jr., "Imw-Pressure Rotor Disc Cracking and Remaining Life Analysis",

Proceedings of the Fossil Steam Turbine Disc Crackmg Workshop, Apnl 1991.

5 l

t.

C6 12 9' CC:24 U. S ti. P. : FEGICr1 1 t ill 013

,-MAY 11 '93_11: 34 PILGRIM NRC RESIDENT P10 GI I

EE EE IE EE o-e s i a j o v.

ea ax

a I

o O O

o o O 6 o t

t i

e se t

a m-QWi c

M O.

ci c> 37 h[

IE m

a t C L.'

eN-=

6 e o

-"I

-o i

83 l

c' Q

CO l E$

$5 h EE KE EE EE E

=-

M e'Al, a 4 E

Er l

sR t,

e a ;;

2m (f)

,, e s s e e e e ci e W

1

.o I we Em

~

hXLQTC,o#y-ze ml CO L EE EE gc Oc a

ee 33 33 2S O

's Ng

, 5 o.

i o ~g -

c y

r-

_m_

e g

W

-- O zu O'$)

2.c.

1 l

=,

o'E fa i

3 O

% C/)

7.$

i e

~4 2 6

e i

e "O E C l Q39 1

)

We o

5.

-O'

>o E D5 EE w

n

~

G' O

t9 co =.g -

1

?>

E e,

E c" =

o C

o_ CC

?j'@

5 ei a'

4 a m.-

O.

C

!2 aF2 i

8es $ 7 o

5 ir UUM l a& Fr $

5 W

l RR t#

ii JH2 9@

i J; t a s o O

e l

v o

t.p h h YN C

i T@?

f ~[. }

$ =

iw M*

5*

g5 m@

1 l P 3s t

g

'h' r -

" 0) i 3b-Jr "E" t a

<c 4

s m

O&

en a E, m

Y

~- SIR 93-047 Rev. 0

~

12 g

s$k Sy

. D' =. f,($ o -

~

{

E= g s fg5zO mwgmz T"

~

~

)

e M

M M)

M

)

)

c b)

N NN N

3vWI.

(

e n i

(

(

(

9r

(

5 5

5 5 5

/e 2

2 2 7

4S 0

0 0 0

~

~

~

e s r

s aa t

De M

M)

M)

M M M @)

)

n

.Y

)

) I

)

b

).

N

(

u n

~

p0

(

(

(

( I (

t i

o I (

1 s

0 4

4 2 2 2 2

~

~

n 0

2 2

1 1

1 I

1 1

~

I 1 0

0 0 0 0 O 0 0

8 i

t

^

~

1 a

5 e

)

5 7 c M

~

c14D/ vee W (n 8

N ib)

(

r I

2 5

)2 i

S 2

3F r d X:

e 0

n

~

2 s

tar n 0 # D a' eg j

a e

ye.

P 7

.Y M

a>

)

(

I p

b

).

N w~

s l

1 t

a s8 un

(

y l

~

l y:i n

oo I

s 8i(

0 e

r M.

I R

1 0,

/w.a.

a e

n 0

w#S o

oe i

t ns ce 1

E.

p i

8 b) m r:

n e s

t

/

e 0 sWI N

o i.

n i

nd 1,

c a1 1

o 0'n yio r

(

i s

5 v

i 1 a r

oom G

t el il ol9o p:. Y r

Ga b

).

t i iR s 3. k n.

N o

1 ndi c a1 i

oo KP n

i n l

6ij 0 n 8 I

b I

(j

,n Al

~

I

[

W)

P 0

h) l

~

Mm 01 u n.

N o

a L

( i nd

/

l 1

c a1 I

35.

l (

i o n 3

Go i

t n i

f x

t yn A

M W) ao

)

)

)*

wi

)* 0" 0

t ya E

5 K K 5 l

E t

1 c

L M T M(

1 2

{

(

(

( (

e o K

0 2

1GG G L

G G s

1 H R R n

l R R I

u e

2e ow f

9 G

A A

A 2a T

G T

a G

O 7

7 ei ge Pw a

y ja E o I-0 p

l

{l1\\

il

Summary of Axia -Xeyway Recommencations [

l Pilgrim #-1: 17DX341

=

S LPA Rotor: Serial -#.: FD5518 t

e"

~

'n y8 ProbaaiListic Insaection

-r r"

interval Resu ts W

i

'i la Whcol(s) Removed

'I" E

"' E "E

"P Temporatura Ternparature Temporature Temporaturo 0

None 0.0 1.2 3.3 5.0 y

(

y c

~c l

t 7TA 0.5 1.7 3.7 52 5 ~,

I

.g 7GA 2.1

'3 f

yj 5.6

\\

N-1 w w m -.i-m l...

- 1

~ x s n.m

u &

=

c.

v N

m t

I, gly j

Fr ? y c F'

I 1

~>

e) g 29/93 Meetirig

[F

\\

l

__m-m.__m.-_

1 52 U

N E

Tabic 3 x

i g

Summary of Results Pilgrim Unit 1 Shrunken Disk 7CA Keyway Cmcklog Evaluation (n}

c

$b a

daldt tra

'~ "

Case Analyzed TugE,,

FNIT K

Crack as

• F ksM,nch Model inches incla inch /yr years es 53 Base Case 75 12+35 105 GE 0.25 0.34 ROG 1.5 U

Parametric Study (Effect of Key Variables):

'.06 9.5 (a) FATT without 75 12 128 GE 0.25 0.82 0

35'F margin g

g (b) Mean Kic 12+35 145 l

l 1.22 l

16.1 (c) Prewarming o

m' 80*F 80 i 12+ 35 110 l

l a46 l

3.5 Z.:

• J 90* F 90 12+35 115 l

l R54 l

4.8 (d) da/dt @ 172*F lab data 75 12+35 105 l

l 0.34 a020 4.5 y '!l NRC egn.

75 12+35 105 l

0.34 0.0164 5.5

-5 (A -

(c) SI crack models Model A 75 12+35 105 Si Model A 0.54 G06 4.8 Model B 75 12+35 105 Sl Model B 0.56 a06 5.2 (d)+(c) 75 12+35 105 St Model A a54 a020 14.5 1

75 12+35 1 05 St Model B a56

- 0.020 115

'J.

m m

~

er E

-c m

s 6

e

'e Power Generation M

3 Critical Stress Intensity vs Excess Tem oerature ij s e

o 300 3

Valid Kie Data

["l All Sources l

250

$38

/

3 20e l

150 E

/+4 5?1

/100

/

w

/

si

/

} & /

i 5

l 0'

)

I

-400

-300

-200

-100 0

100 200 300 400 Excess Temperature ( F) = Test Temp. - FATT R

Fiigure 1.

Fracture toughness (Kg) as function of exc-ss temperature (T-FA~lT) from Reference 4.

4

!I E!

HOOP STRESS VS SHRINK-FIT & RPM w

t 45

[

s

<h x

2 40 ta i

@g O

~

'.)

f 35 p3 c)x 30 il 3

+

gM m

o n

25 g

a m

20

\\

g:n a

'b+

hh O

I 15 -

D T

10 -

cn e

5 I

1 i

~

f 0

5 10 15 20 25 30 RADIAL DISTANCE FROM BORE SURFACE (IN.)

0 1800 RPM ONLY

+

SHRINK-FIT ONLY 2

0)

Figure 2.

Hoop stress as a function of radial distance from the disk bore, due to tie ir.dhidual contributions of: (i) rotation at 1800 rpm and (ii) 0.21% shrink fit.

l

2D

..g p$

C HOOP STRESS VS (SHRINK-Fir + HPM) 70

.w N

~

9 60

(

I b 3 o

N_

l')

' \\

M 'it 50 V

{"

x c

3 '3 y

40 z Ej v

n hg n

a y

m E

~

O Cn VI H c o ".

1 m '-

-i m g

o 10 -

f

-g (n -

O I

I 3

I I

I 0

5 10 15 20 25 30

}

RADIAL DISTANCE FROM BORE SURFACE (IN.)

D RPM

+

450 RPM o

900 RPM a

1350 RPM 4

1800 RPM v

2160 RPM Figure 3.

Hoop stress as a function of radial distance from the disk bore, doe to Oe combined effects of rotation and shrink fitat speeds of 0,450,900,1350, 1800 and 2160 rpm (without the K, effect).

n

-C '

W Kt DIE-AWAY FROM KEYWAY SURFACE 8

KEYWAY DEPTil - 0.625 INCH

~

k d

2.3 U

U a

?

2.2 2

w o

I 2.1 u;3 N_

s 2

m>

p3 m

1.9 N

59i 1.8 7 l';.

s t.7 x

O G

q 1.6 xy M

1.5

@Cin z.:

1. 4

--i y 1.3 1.2 -

xE g

1.1 -

1 -

l 3

I I

I I-0.9 O

0.2 0.4 0.6 0.8 1

1.2 1.4 1.6 1.8 2

RADIAL. DISTANCE FROM DORE SURFACE (IN.)

b Fgure 4.

Decay of peak keyway stmu concentration factor, K, of 2.2 with increasing

{

radial distance from the keyw-surface.

/

. 'D

-c e,

HOOP STRESS VS (SHRINK-FIT + RPM)

WITil Kt = 2.2 140,

U yg e

o 130 c

120 "

$8 110 l

][

E 1 00 x

90 d

g3 80

-I z !U' h

v m

en 7C

?

O (n

x E

~

y 50

(

@{

10 -

v.:,;;4:n+w

-n

-a 00 -

I I

I I

I I

0 0

5 10 15 20 25 30

)

RADIAL DISTANCE FROM BORE SURFACE (IN.)

0 RPM

.+

450 RPM o

900 RPM a

1350 RPM X

1800 RPM 3

V 2160 RPM Figure 5.

Hoop stress as a function of radial distance from the disk bore, due to the combined effects of rotation and shrink fit at speeds of 0,450,900,1350,

(

1800 and 2160 rprn, including a peak keyway stren concentration factor, K, of 2.2.

I

ge, g; 9-

- Cel:28 U. s i t. P. C FEGitti 1 6 114 022 MAY 11 '93 11:38 PILGRIM HRC RESIDENT P19

^-

4 4

=

~

o 1P e

o u

o n

~

t

/

~

(a) 7 l

a

~

o - c, + c,x + c, x *. c, x e h

S (b)

\\s

.~

X e

m_

a

~

?

- ~b I

Figure 6.

Crack models used to perform the 'K versus a* calculations: (a) GE model and (b) crack at a hole in a plate.

)

SIR-93-047, Rev. 0 21

3 DISK #7: K VERSUS A w

o COMPARISON OF PC-CRACK VS G.E. SOLNS 140 U

O

~

g 3

0 I

En o

130 E

G$

2

[

T"

?

/

?*

S

}

W R

1 20 g

(~

/

2B d

j Wa 110 O

d.

$c

?

M 4

k E

/ //

$E i

/

m u

a0

'E

/

80 3

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

CRACK DEPTH'FROM KEYWAY SURFACE (IN.)

O G.E. SOLUTION

+

PC-CRACX: NO Kt o

PC-CRACK: Kt=2.2 8

E Figure 7.

Stress intemity factor, Kg, wrsus crack size, a, for the fonowing crack g

models: (i) edge-crack in a semi-infinite plate with unifonn stress (GE), (ii)

O crack at a bok: in a plate without keyway K, effect (Si model A), and,(iii) crack at a hole in a plate with keyway K, of 2.2 (SI nxxiel B).

N U. S t i P -

ou 31 ys 11:

P LGR r t4RC RESIDENT P21 4

4 e

s Stressos without Kt

\\ S

%.w r

R = 16.5 1

TJa

+

~

Si Model A a

Strosses with K '= 2.2 t

%,5 R,g - 16.5 + S/g

~

f, 8

lm Si Model B Figure 8.

Schematica of the crack models used by S1 Illustrating the meitsureme crack size versus hole size and stress distributions; Model A: Crack at a hole in a plate without keyway K effect, and, Model B: Crack at a hole i a plate with keyway K of 2.2 (SI model B).

g n

2 SIR 93 047, Rev. 0 23 INC

05 12 93 07:2G U. S t L P. ': FEGIO1 1 i !!!

025

..MAY 11 '93 11:39 PILGRIM NRC RESIDENT P22 4

4

.9 b

%s.,

c bao m

2 c

g

.].

t c m

EE E

E$

8

=

\\

O

$555 l[

e.4 ic o

CD a. CD O.

O h

$b h

o9D4 9

h$

?5

=

u o

a53

=

n c

sa Eyn o

Q D

00 u

v 2@R I

oI gg c2 s

ESE gw "o

(

14 55 oo o

m age e

s e

n.In o

.9 j

aoo "

i 8

E

=[-

u 6uef> o o j

o g

.8 Ps.

Ei I

sw m

=

J

.8 0 *ur o

O c4 J

8 h

h h~

k e

(Jyul) eles ynmoso

{

e SLR 93-047, Rev. 0 24

_h

_l is e w :

v r

2.

-yh$g.

. !d

~

x i,,

n g %_

!=

'd

.54g. eg, ilm - w w/.. ~..

y e

4

.s ec

%; b,;s

?lb' e5 $

4, Ifhp h

k f i i k f $ $ h T a a w m m m m'g g.

3 n

( WWh,htff.w -

1 s u.

. ' J ?.'

,.z;gf 2

l'"..~'",..,; & w..*

...M i (- *+94 _IE".

'E

+ y# '

~-,

,..r,f.$i.Y,

'? )

'T

• i,_.

fW

.4

- LEU.Vy 4 8 4 9 7 s

h *1~b

^

e f

W O.~: 0

=mpllllv

' asan. gg w:2

\\

t 1

~,

. ' '?.:' 7,,s. %YV>0~s:A4h f y u.

?~'.

G

S J Shfh.#[yuweqp&gNM-h_m.

~ :;~

u g - @hMj;N e-

- [s - e..w.,

-A[.'.w

.; f, g

/ //e /

s

= =.

nn n,,

e

.t;. x ;g g,

p, ff * ~

~

- e n

.c

,./

s>. g, N 'n ?s n., k l N [ f m. n..

g,}

\\

D:

af 3 /

. j '~

If 5 5Q 9f "a

I a n

+y7 y.s wlg lj

- lL '

mms Ma

%z % %s. r ";. ;,f.:. k,y, T q

.a

w g g 4 g%, y 39 L. E z.

, m., A

+

y up s,

I

. )s $: M. _i _..I k M m k,f-

+

4

. y.C, e '., y, c4- #

,,N, d m

c:xy n_

" h-.*: -

5,.yt.

f

• _Q (:,:JCR.r~M

.,f e #MN ' F ~.~ h

%g->t

?

$ f?' ~

hk$q% ';f

./3

' 4;.

y A

,[*

4&

_ y. & v.,.q

>f f

.. E P ; ar;L, E:;-

. w..'m

.+

f$.

yz,.%.1

.y M

~ ns

~ ~.- Q.

.5p n y r.

• \\ \\:;

f_

t '.,* $,, u

'Y"gm.' O.' t fh f

.Q-g,

- - ~..-c:

w f,

Th 5.!

~

h N$si,6$e87.fsWN.?

~p# l:

wo.m

.. a

~

.. n

f. f. 4

.,4b;.

14.

' Q. y y

}.

); *:

3

\\

C

),'. i 7~ '

a

w...

g $ g km ME s

s,

t t

c.

PP f

h Power Generation f

l i

Test Concept a

.i e.

t, ;f; Y

1 e

s' d

.ct

,. ' I

,I l

2 s

i

. I t

,f

\\

0 ;

g 3

N

'i.,

t s

\\

s' Dual Transducers V,'yf \\

? [h@!N c 6

s s From Web Surface i

S, xpu.u.

., y:r,,.

n..'i!4;rj.:..g s,y.

. n y t.. -;. >

i;, k'", ', ! I.

[,'.

s k

c..,

e z 's -

' {xfg i..,l

.[

..y.

S( <

i,

{

t{ > -

r y < ;.

~,

• e> p t '.

y.

I.' e #

T j,,,, '.

d s,,.. -,

+

g e

s, y: 3.u s,,

. J.*%,,.},[ j '

, 7

?>t 9

k

3

. (S -

1*

Power Generation Measurement of Radial Depth Longitudinal Pitch-Catch Wave v/

Shear Wave Transducer Transducers 1

X 1; O

/

A

/

i Crack - '

Keyway Crack Determination of Flaw Size as important as' Flaw Detection Alone

.-s,

.,,-..,.,_.e..

.mm,-.

.. om..

r

e

• I Power Generallon Ultrasonic Testing of Keyway 0

Detection 9,

N ',

sk

.\\

O

~

[

l s.

1

'\\

/ k

.... -y l-..;.

,...g.. 3,.

-. - ei em up - aus

= eBD em

• 4 4

m--

w-e..

,,-,.r w==

n-,in,m.,

w y

e-w

=,, - -..

.-e c

>-%w

-+.

,es-

,.4, i

,m--e e

=

a r...ic e

~i=.%

i ws

= = --. -

nu-wam' eda m

n a-y (h

Power Generation L

Ultrasonic Testing of Keyway sizing

1 1

1..

7l, l

C 3

\\

11 A

II

\\

l J

'{

\\

,/

L F L L

J i

a

\\l n

rz-

._.;_..,.-,y

..w.s,_

v,

,,,,s

.. -., +.

_i,..

4

,,._,-c

3

4.,.

.. g h

Power Generation Nuclear Wheel Inspection Stress Corrosion Crack 1

V 4

N Water Cutting Stress Corrosion

/

Crack pyJ;b

=Th..g u

e.............,

Keyway.

/.

Axial Key 1

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

-_m-,

.m

>=

4.

b, $ ?

p t "

11 Jj s

W Power Generation

!i:.

r. t Longitudinal View of Keyway 7

I Steam Wheel a

i Flow l

-[

11 L Stress Corrosion Crack Gap /d :n Circumferential 3 Between - ...(l., 'j.. d'l, ( Relief Groove q p [ gl' w Wheels o 4 ~ = = = r r _r_._v_ _m _-- _ _e_..fr,.: : : : : : : : l l Vapor H L________^K_EY________j Shaft - - i :- l l i (

Power Generation Keyway Geometry for Shrunk-on LP Wheels ('I) Steam Flow i> ~ ~ Shrunk-on Wheel Section A-A -A ......g

---

,..=..=

Shaft ______j V Locking ( . Axial Circumferential Ring-Key. Relief Groove I t f 9 . ~..

l'" 5-i g Power Generation b l ? 9 Branched Crack Fatigue Crack t i e 4 4-1 I m . m m -..

a ~

q. y 1

$) ~ rower cenerauon Assumptions When No Measured Size Average of Available Data: Under Hub 0.08 inch 1 Under Web - 0.25 Inch + Cause for Non-Measurable Indications Has Been: Pitting Scratches / Gouges Stress Corrosion Cracks in Keyway Up to 0.55 Inches Deep

• ,,-..-m

_..... - -...ma*---,.,

1. s,s

,----.w 4 e ~. ~v-,.,-, w--, -e.- m m m

p Recommendations for Boston Edison Company Pilgrim #1.170X341 e i: ins e ion Se Rotor Recommendation I ye s b LPA 3/80 5.1 4.5 Years 1 0/81 6.3 4.5 Years 2/87 8.8 3.1 Years Limitting Wheel STA LPB 3/80 5.1 4.5 Years 2/84 7.9 4.5 Years'

• i l

6/91 11.0 3.9 Years Limitina Wheel: 4TB l Current Service Years = 14.5 12/4/92

Hypothetical Optimum Dovetail Litrasonic Examination x Vane Ultrasonic Transducer. k \\.\\ l \\\\ L NJ t 1 \\ i w NJ 1 Wheel \\ N I

Dovetail Ultrasonic Examination Nr 3 b, %/. ^ c <w/ AX t O N ix uttrasonic. ~ 1,._ Transducer Ultrasonic Transducer { Wheel e S e J

~~ %ggf7 - Gii' bo d Boston Edison Company Pilgrim #1: 170X341 LPA Rotor: Serial #: D5518 Insp.Date ~ also insp. Date: 10/81 lnsp. Date: 2/87 Insp. Date: 4/93 5.1 Years 6.3 Years Service 8.8 Years Service 12.0. Years Service. Keyway SerVICO Hub Web Hub Web Hub Web Location Hub (In.) U n-) Un.) Un.) Un.) Ca.) Uni 4GA LE 0.13 (M) 0.14 (M) 0.30 (M) 0.2G (M) 0.53 (M) 0.42 (M) TE 0.00 (NM) 0.25 (NM) 0.22 (M) 0.32 (M) 0.51 (M) 0.44 (M) 4TA LE 0.28 (M) 0.30 (M) 0.33 (M) 0.31 (M) 0.35 (M) 0.43 (M) TE 0.23 (M) 0.31 (M) 0.24 (M) 0.42 (M) 0.30 (M) 0.49 (M) 0.50 (M) SGA LE 0.26 (M) 0.38 (M) TE 0.26 (M) 0.37 (M) STA ML 0.31 (M) 0.41 (M) MT 0.29 (M) 0.36 (M) I Page 1 of 2 LE: Leading Edge 4/28/93 Meeting TE:' Trailing Edge NM: Non-Measurable Indicat. ion

w. u,aai,,

u. i n a i,,,, + i ~, s., - u,- - - - - u -- ~ - - ~ ~ ~

Axial-Keyway Indications L Pi grim #1: ' 70X341 L3A Rotor: Seria #: FD5518 also insp.: 10/81 Insp. Date: 2/87 Insp. Date: 4/93 5.1 6.3 Years 8.8 Years Service 12.0 Years Service Keyway Hub Hub Web Hub Web Hub Web Stage Location (In.) (In.) (In.) (In.) (In.) (In.) (In.) 6GA MI N N N 0.25 (NM) 0.25 (NM) LE o o o 0.08 (NM) GTA M1 1 I I 0.08 (NM) 0.21 (M) 0.25 (NM) TE n n n 0.25 (NM) d d d 7GA-Mi i i i 0.25 (NM) RG (10') c c c 0.14- (M) a a a 7TA RG1 (KW) t t t 0.14 (M) RG2 (KW) i i l' O.12 (M) l RG (5 ) o o o 0.12 (M) l RG (10') n n n 0.12 (M) RG (25') s s -s 0.12 (M)- Page 2 of 2 RG (KW): Relief Groove in-line w/ Keyway 4/28/93 Meeb.ng RG (Anoin): nolof Grrova nt 9narmaa Annia - ~ -.

.r. 1 C'orrelatiori of Web Crack Depths With ho Measured Radial i: Shrunk-on Wheels -- 3 Actual Depth Deteription (Inches) Lab. Cracked L-2 (122N296VA2) - 0 Degree Keyway 0.296 90 Degree Keyway 0.669 180 Degree Keyway 0.046 270 Degree Keyway 0.216 Lab. Cracked L 2 Wheel (3991V1) B Zone Bore 0.029 9 -L O ,3' G Zone Keyway 0.324-H Zone Keyway 0.073 -R Ouad Cities #2 STA Keyway 0.320 SGA Keyway 0.380 Lab 41 Wheel AK2 Keyway - 0.160 AK3 Keyway 0.210 Average Actual Depth = 0.248" Four Comore #4 L2A 0.410 l L-2 B 0.550 = Mohave #1 ~L-1 0.250 y Kashima L-1 0.000 0 9 /A /00 .}}