ML20116J325

From kanterella
Jump to navigation Jump to search
Draft Rev 0 to GENE-523-136-1092, Engineering Assessment of Hatch 2 Ahc Indications
ML20116J325
Person / Time
Site: Hatch Southern Nuclear icon.png
Issue date: 10/31/1992
From: Dias K, Ranganath S
GENERAL ELECTRIC CO.
To:
Shared Package
ML20116J287 List:
References
GENE-523-136-10, GENE-523-136-1092, NUDOCS 9211160133
Download: ML20116J325 (26)
v • d • e


Text

,

t

.w..' -*

A GENE-523-136 1092 DRF.137-0010 -

Revision O.

' ENGINEERING ASSESSMENT OF THE-

' HATCH 2 AHC INDICATIONS r

1

.0ctober.992 Prepared-by:

A

'K.PLDias,l Engineer:

StructuralMechanics q

p _ m >L.L.

'/

~

Ap~ oved by:

/ )-

t/

(SiRanganath,PhD., Manager

-Structural Mechanics' 9

GE NuclearEnergy.

9211160133 921110-6 DR.-ADOCK 0S00 m.

wt

_ ;, n.

T ~i

GENE-525136-1092 IMPORTANT NOTICE REGARDING CONTENTS OF TIIIS REPORT -

Please Read Carefully ne only undertakings of the General Electric Co pany (GE) respecting information in this document are contained in tne purchase order between the Georgia Power Company and GE, and nothing contained in this document shall be construed as changing the purchase order. The use of this information by anyone other than Georgia Power Company, or for any purpose other than Jiat for which it is intended under such purchase order is not authorized; and with respct to any unauthorized use, GE makes no representation or warranty, express or impiled. and assumes no liability as to the completeness, accuracy, or usefulness of the informatior, contained in this document, or that its use may not infringe privately owned r:ghts.

l GENE-523-136-1092

SUMMARY

This report doeur-as the evaluation of structural margin with regard to circumferential eracking of the Hatch. access hole cover (AHC) weld to ledge region detected during UT examination. The results of two types of structural analyses are presented in this report to assess the crack indications.

Although finite element analysis was performed to assure structural integrity of the AHC, it is important to note that the access hole cover weld does not represent a pressure boundary weld and is therefore not strictly covered by the flaw acceptance criteria of the ASME Code. Furthermore, if some through thickness cracking does occur, accompanied by fer.kage, there is little impact on plant performance or safety. Finally, in the lin'.iting case where separation of the cover is postulated, the reselting core bypass flow would be readily det% > as a result of the SIL 462 criteria and shutdown implemented in an orderly manner, Even a N cover lodge.< h< elf in the recirculation line and blocks flow, this is still covered by the pumi seizure event which has been evaluated Thus, separation of the cover is bounded by existing safety analysis. Therefore, the AHC cracking does not represent either a pressure boundary integrity issue or a safety concern. The NRC has implici*.ly accepted this in Information Notice No.88-03 (Attachment A).

Two types of structural analyses were performed to evaluate the crack indications. The first considered a crack in the AHC at the weld interface with the ledge (Figure 1). The second considered a crack entirely in the ledge with the crack tip just flush with the top of the cover. The results of the two finite element analyses are described here.

1.

The analysis with the crack at the weld interface (Fig.1) showed that a ligament of 1/8 in, is 7

sufficient to maintain the required margins against limit load and hinge formation.. After accounting for potential crack growth, this translates into a ligament requirement cf 0.14 inch.

The available ligament of 0.26 inch is well in excess of this. The analysis was performed for normal operating differential pressure of 27 psi and a safety factor of 3. Therefora, if the remaining ligament (and crack) were to exist in the AHC weld interface, there vauld still be sufficient margin to assure structural integrity of the AHC.

1 l.

4 w

.q-GENE-523136-1092 2.

The second analysis with the erack entirely in the ledge and with the tip Oush with the top'-

surface of the plate also shows that failure is not predicted even when the crack tip is offset by just 0.125 inch from the weld interface. Although assuming a crack at the weld interface is more limiting, the UT results (Attachment B) indicate that the crack actually lies 0.125 inch or-more into the ledge. Radial crack growth for a ledge crack was judged to be negligle.

Based on the results of the two finite element evaluations it is concluded that the structsral integrity of the weld will be maintained. Olhar reasons for justifying continued operation such as low crack growth due to hydrogen water chemistry, tiie presence of metal path due to crack

~

branching even with thrcugh thickness cracking and the implementation e ML 462 guidelines remain valid..

t c F 35 176I" i

4

.' i 9:R?:27 PLOT NO.

5 m r2 s Rrcs ETEP=1 4

(

IIER=1 Ex CauC)

S CLOBAL j

oge4 DMX =9.B197&7 SMN

=-C Z 2 9 OMNB:-<S3E2 i

EMY =56886 i

]

SMXE=C2153

n

_- 1 D.OT=1.711-XF

= 9. 2 44 f.

YF

-2.BC3 I

-C229

)

-21216 i

M sum t:B61B3

)

M 1522 auuD 12c2s H Z3Bc8 i

34861 i

i C5874 56886 I

'i I

I l

l l

l

)

AHC S7 PIOS AHALYSIS - 20% P.EMAIMINC LICAKCh7 Radial Stress Distribution for 20~. Remaining Ligament Case FIGURE 1

\\

I

9 GENE 523-136-1092 ANALYSIS OF CP.ACKING WITillN TIIE AllC WELD Crack Growth Ratti Crack growth rates under sustained load IGSCC conditions in Alloy 182 are dependent on the applied stress intensit'j factor and the water chemistry - conductivity and electrochemical corrosion potential (ECP). But available data (Reference 1) indicate that crack growth rate is not strongly dependent on the applied stress intensity factor. In fact, beyor,d a threshold K value of-approximately 15-20 ksiv'in, the crack growth reaches a platecu value of 2-5x10-5 n/ hour for i

conductivity levels up to 0.5 pS/cm. However, the crack growth rate drops rapidly at lower conductivity values.

Similar behavior has been observed for SCC growth rate for Alloy 600 (the AHC material and the SSP material). Figure 2 shows the weekly average reactor water conductivity for Hatch-2.

From the tigure, it is clear that the conductivity levels at Hatch-2 were maintained below 0.2 pS/cm for almost the entire fuel cycle 9. For 0.2 pS/cm and normal water ehemistry (NWC), the typical ECP is + 100 mV. For NWC conditions at a relatively low conductivity, it is then reasouable to assume an expected crack growth rate of 2x10-5 n/ hour, consistent with the observed crack growth i

plateau discussed above.

Hatch-2 has been operating on hydrogen water chemistry (HWC) at 0.6 ppb H. Since the 2

conductivity is Iow, the GF CR&D " PLEDGE" model can be used to predict crack growtn rates for HWC conditions. Figure 3 shows measured ECP data as a function of hydrogen injection (Reference 2) for a typical BWR, Based upon this data and a hydrogen injection rate of 0.6 ppm, the ECP is approximately -100 mV. For this ECP, the " PLEDGE" model predicts a significant decrease in crack growth rate as compared to NWC conditions at +100 mV ECP. The crack grawth rate is reduced by a factor of 16.3 to 0.122x10-5 n/ hour. The predicted _ ligament raiuction i

over the next operating cycle is calculated as follows, assuming 12,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of operation, (0.22x10-5)(12000) = 0.01464 in.

of 0.01464/0.625 x 100 = 211 where the AHC through wall thickness is 0.625 inch.

~.:.;

-t GENE-523-136-1092 Minimum Recuired Licament Size Finite element analysis has been performed to show the minimum required remaining ll; ament size around the AllC as a function of AHC pressure differential. From " Reactor Internal Pressure Difference Databook" (Reference 3), the following pressure differences across the AHC for Hatch 2-were obtained and adjusted by appropriate safety factors:

Safety dP (psi)

Factor Allowable dP (nsi)

Condition 26.46 3.0 79.4 Normal Op 47.7 1.5 71.6 Faulted The limiting dP of 79.4 psi is used to determine the allowable remaining ligament size for the AHC Figure 4 sliows the finite element predictions. For dP=79.4 psi, the remaining average -

ligament size is approximately 20%, or 0.125 inch.

To justify continued operation through the next operating cycle, the above allowable ligament size is adjusted to account for one cycle of crack growth as follows:

Mir.! mum Required Ligament = 0.125 + 0.01464 = 0.140 in or,

= 0.140/0.625 x 100 = 221 That is, the measured average remaining ligament (= total thickness minus average crack

~

depth) must be greater than 0.140 inch (or 22% of the total thickness).

Results Of UT insnection Due to limited access in the ledge cegions around the AHC, only 67% of the circumference could be inspected, as shown in Figure 5. UT data was obtained using both 45' shear and 60.*.

signals. These results are shown graphi_cally in Figures 6 and _7. Although the 60* signal is -

l 5

g

?

GENE-523-136-1092 generally more reliable, the average crack depth was conservatively computed using the more limiting signal at each 5' measurement interval. The boundhg crack depth profile around the circumference is shown in Figure 8.

^

Based upon the bounding crack depth profile (measured at 3* intervals), the average crack depth was computed to be 0.365" (or 58% of the thickness). It is assumed that the average crack depth in the ledge regions, where UT inspection was not possible, is also 0.365" Therefore, the measured remaining ligament size (assuming the crack lies within the AHC_ weld) is, i

Measured Average Remaining Ligament = 0.625-0.365 = 0.26 in or,

= 0.26 ~.625 x 100 = 42 %

Clearly the measured ligament size is well above the minimum required ligament size of 0.140 inch (or 22%).- The local peak crack det.th w.ts 79.2%, which corresponds to a minimum ligament size of 0.130 inch. This value differs from the minimum required ligament by only 0.01 inch.' This difference is small and occurs only over a 5' interval.

- c

M i

Hatch-2

~

Fuel Cycle 9 7

- 7 +3.

99.5

[- -

/

99.0

/

2 95.0 90.0

- -- 2. -

+1 80.0 b

f

.70.0

[

..j '

soo 00 R

50.0

---. ~

--~

-40.0 g

p-

]g.

_.30.0 l'

- 20.0 o.

/,

-1 3

10.0 4

o 5.0 -

. _2

- o

. 1.0 0.5

~

_I

._,,-3

' O.01 0.1 1

Weekly Average Reactor Water Conductiwty (pS/cm).

. FIGURE 2 '

}

w" 2

3-5 r'

,m'N v-

.f y

y

}-

W9>h, y

p pd g

y y

g' w

q

300 ;

l i

200 '-

100 L w

5 0

E 6

- 10 0 w

-200 BOTTou = 1

-300

-400 O

0.4-0.8 1.2 1.6 2

2.4 2,8 rdEDWATER H2, PPM FIGURE 3 Measured ECP at bottom electrode positions (Ref. 2).

5 4

AHC PRESSURE CAPABILITY VS. LICAMFNT cc -

l I,

J5s,

?

l e

'UC-i.ivENT =4E0;ctCn

'/

d.

200 -

i k

I h

250 -

o=

5 200 -

iv i

E 150 l o

A h

i 100 -

i u

4 50 -

l O

i i

i 0

20 40 60 30 IC REMAINING LICAMENT, PERCENT Pressure Capability as a Function of Remaining Ligament FIGURE 4

e No Scan Ledge Side

\\/

'~~'N 25

.340 s 'A/

s

\\

/-

Access Hole Cover i

/

[

'\\\\

'/

14 0 2 15 x'N,.

No Scan Ledge Side Total Circumference Scanned = 240 degrees (67%)

Avg. Remaining Ligament = 0.26" (42%)

i FIGURE 5

w Ad'-

9 4-.

3 L

4 4

a m

e i

I I

a r

t I

fHATCH 2-180* ACCESS COVER (SHROUD SIDE) THRU WALL LAYOUT N 5 hkAR DhTM m

I m

I f

f

{

.<\\.

I t

)

\\ \\f f.

m f

/

[o I_

\\

}

r m

i I

T i

H u

i

\\'

1 StW Watt

_y

-f

.n.-

t

/

m

{

v 3 ia i siTTTTTTT

7 T1 O

~ 'TTg :

0 5 5 5 3 2 3 3_*$_4_s E E 3 5 5 3 3 4 3 8 8 E S 4 8 9 8 2 j

NCHEs m CActn6ERENCE

_: m. -

i

'.l t

t FIGURE 6 l

i

!!i4 1

t e

.I fi[

?

eit!

4.

4 n

m v:

gl

,s i1' i

e m

e 4

s

_ IA

+

e T

A D

e L

R 0

i i :

6

. h

(

s Tb L

\\

'-z YA

\\

e L

/.

L k

u LA W

i U

R H

r r.

T E

r 7

)

I D

r E

R S

y na_

U D

r x U

r G

O I

r F

R H

S

=

(

R

=

E V

OC 1

SS I\\

E

,\\

C I

C

, p

  • f,,
ll A
  • 0 8

1 2

\\

H i

C T

A l

1

\\

i, n-0 O

0 0

o o

o o

o 0

w O

0 t

4

?

6 1

4

,,RU WALT Lll l

I l

,l l

i l\\IIl,fI(

lll

- - - ~ - - - - -

- - - ~ ~ - ~ -.

---n-

--n--.~

4 I

[

I 4

i I

s.

i 4

,7 F

  • ~

j sp k

g es

~# '.

t s

w W

i l

w__

i

-sr-

.g j

}'

n-

~

l l.s i

)a at P

CC 1

i I

_s N

g x-f c

fl w.

.3 g

3..', p I

l

-r c

r

- w~

1, D

  • .i yh W

N l

5 I

= f m.

- se:

W

-g oc:

~ ne p

- n' O

~

l

[

,,- J

~

n g

U N

J

~

it wA

  • as f

t, eem assumes op 4

jv"~

5 x

k i

s-

~

Ih ]

.e

=

4 7W h*

I i

i i

i i

~

M..

m-m.

b i

1 GENE 5231M 1092 ANALCIS OF CRACKING IN Tile AllC LEDGE The previous analysis conservatively assumes that the cracking is within the AHC to-ledge weld. Ilowever, a careful review of the UT plots confirms that the crack is entirely in the ledge. In most cases, the tip of the crack is at least 1/8 inch or more from the vertical weld interface.

Therefore, even if the tip of the ciack extends to the _ap surface of the cover, there is still a remaining ligament since the crack is in the ledge.

Based on the above findings, a second finite element analysis was performed for the special case of when the crack tip is just at level of the top of the cover, but 1/8 inch ;nto the ledge. The analysis was performed for a normal operating pressure equal to 27 psi. The structural integrity of the remaining 1/8 inch horizc :hl ligament is m iluated based upon the axial stress d:stribution results (Figure 9) of this special finite elemem analysis. The maximum stress v.17.6 ksi.

Assuming a value of 3Sm=70 ksi for limit load failure, there is still a factor of 4 margin remaining.

For the limited faulted condition (approximately 48 ksi pressure), the peak stress at the ligus.it/ hinge is.cated up to 17.6 x (48/27) = 31 ksi. Herefore, for fat 'd conditions there is approximately a factor of 2 margin remaining.

For this analysis, where a ledge crack was considered, radial crack growth across the horizontal ligament was t-glecteo Mr the following two reasons:

1. At flatch 2 hydrogen water chemistry (llWC) will be maintained with high availability.
2. Crack growth in the radial direction is not likely unless significant branching exists.

Therefore, the results of the FEM analysis (ledge cr

  • ing) are sufficient to justify continued operation for one trore fuel cycle.

For the vertica: ilgament, the previous AHC cracking analysis is bounding and has shown that the remaining vertical ligament of 0,26" is sufficient to maintain required margin for continued operation... -

i

's

'I n

i G

Nm tw

@ n H

1 VN M

@MM m@@M3h T

e W

^ TTN f Ob@TSHMDb3 Ve DC U H @M @ th efN@nnneN*ered WM *F D>33b@NT CD eMb f MTMO @e@N NOM as.HNosn

.S H 1HooNnbare i

HG a*Z 4,4 v iE 31 i e.4

  • S H 3 I I l I i i bn @e I

ti ll 1111 3 ll till 1118mme l

> N th w ll il verg w

%3814 >?l HEE b lllll{ },

i maen 7gyg c e r-em i

I,luluil m.im V wppysyggyggI

, I llilllil H P ir8 H

rawnnesumi ef:a lt

p.

o p ;p qur si+

i MLi'i lll lll lifiili&dllj [&

,l.;

1 i l a

e a

L W;](MHiQf

  • wr dpudw o

l l

s FaitEnj i i

1 H j,'...,,

H h7 I*!j ;j.l e

,, e

m a.ml l [

DififiOMI:Eli@8 i

1 If h7$71TI'"

l I

I I,'I!hri lTil l

l l

l o:cs j

Idl _EE%ws ^ ^

l n.

I I

I e

U U4 N

3 4

2 t

FIGURE 9 4

i

---_ ---___ _.=-.---.- ----,.

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

- - - - _ - - _ _ _ _ _ _ _ _ _ _ _ _ = _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _

j GENE-5231361092 CONCLUSIONS AND RECOMMENDATIONS dased upon the preceding analysis and discussion. the following conclusions and secommendations are offered.

1.

Based on the stresses in the vicinity (1/16 in.) of the crack tip, limit load failure (due to hinge formation)is not predicted even with the crack tip close to the vertical weld interface. Even if the highest stretsed region (1/8" horltontal ligament) were to forra a hinge, tearing will not occur because of the inherent ductility of the material. In any case the cover would be restrained by the ledge, 2.

Even if through thickness tracking occurs over 50 perce.nt of the weld circumference, analysis shows that separation of the covtr will not occur.

3.

llatch Unit 2 is operating under hydrogen water chemistry. The injection level is sufficient to-lower the ECp in the region below the cover to -100mV. This w*ll reduce IGSCC growth _

j rates significantly.

4.

Because of the extensive crack branching in the weld metal under IGSCC, even if through thickness cracking is visually confirmed, there appears to be sufficient metal holding the cover in place. No gross deformation or incipient failure condition has been observed even in AllCs (in other plants) with much more severe cracking. On the other hand, at llatch 2 there are no visual indications, confirming that there is no through thickness errnking.

5.

The separation of an AllC would result in the introduction of a new flow path in parallel with the reactor core. The hydraulle resistance of the r' circulation system would decrease and this-would lead to an increase in the total recirculation (jet pump loop) flow. However, because a portion of the jet pump loop flow would bypass the core (through the access hole), the actual U

core flow would decrease and lead tu a reduction in core power consistent with the normal power flow map Since the access hole is approximately 20 inches in diamete, the impact on recirculation flow and core power would be significant enough that it would be readily'.

p i

GENE 5231361092 e

apparent to the operator, who would be expected to take corrective action. Attachment B l

describes the consequences of a loose AHC getting separated and approaching the recirculation pump, it concludes that there are no safety consequences for such a postulated event.

t 6.

The. recommendations of SIL 402 will be fully implemented. Even if failure of the cover (or signi: leant opening) occurs, it will be readily detected by the change in the core bypass tiow and the plant can be brought to normal shutdown.

7.

A safety evaluation of the circumferential cracking showed that, even if potential separation of l

the AllC from the SSP were postulated, such a failure could be readily detected based on the

+

observed changes in the core bypass flow. The safety consequences of circumferential cracking %ve been fully reviewed with the NRC and the staff concurs that the circumferential cracking does not represent a safety issue. No radial cracking has been observed at flatch Unit 2 Thus there are no pressure boundary integrity concerns. Therefore, it was concluded that the circumferential cracking did not represent a safety concern.

8.

The 0 degree cover is free from indications. Southern Nuclear Company is evaluating further all UT data to assess crack validity in the 180 degree cover. The we'.i represents a complicated configuration which could contain fabrication flaws and/or dissimilar metal ~

reflectors.

9.

GE provided reactor operating guidelines for detecting core bypass flow in Sill No. ' 62, 4

l Supplement 2, Revision 1 dated December 19, 1990. GE also rec ~nmended that the creviced Alloy 600 AHC plates in the affected plants should be inspected ultrasonically during the next outage and that the examination should be repeated at intervals of no longer than 3 years.

Southern Nuclear Company will evaluate repair and inspection. options to address the cracking concerns during the next Unit 2 oJtage.

~ In view of the above reaso,s, GE believes that continued operation for one more cycle is -

justitied.

.=..

-=.

w,

.O GENE-5231361092 REFERENCES 1.

S. Ranganath, *llatch Access l{ ole Cover Assesment", Rev. O, GE NE-523113-0892, August 1992.

2.

"In Core Response to flydrogen Water Chemistry at J.A. FitzPatrick Plant", Prepared by GENE for Empire State Electric Energy Research Corporation, EP89 32, January 1992.

3.

GE Doc. 3831{ A428, Rev 4, " Reactor Imernal Pressure Difference Databook." I

F t

GENE-5231361092 r

4 h

t ATTACIIMENT A

.n..

-.a

UNITED STATES NUCLEAR REGULATORY COMMISSION OFFICE OF HUCLEAR REACTOR REGULATION WASHINGTON, D.C.

20555 February 2,1988 NRC INFORMATION NOTICE NO. 88-03:

CRACKS IN SHROUD SUPDORT ACCESS HOLE COVER WELDS 4

' Addressees:

All holders of operating licenses or construction permits for boiling water reactors (BWRs).

Purpose:

This notice alerts addressees to the potential for cracks in the welds of the covers of th shroud support access holes within the reactor vessel.

The cracks could result in weld failure with resulting formation of loose parts and core by-pass flow. The event described highlights the importance of inspecting the access cover welds.

It is expected that recipients will review the information for applicability to their facilities and considor actions, as appropriate, to preclude similar pro'lems from occurring at their facilities. However, sug-gestions contained ic this information notice do not constitute NRC requirements; therefore, no specific action or writtan response is required.

Description of Circumstances,:

Jet pump BWRs are designed with access holes in the shroud support plate whien is located at the bottom of the annulus between the core shroud and the reactor vessel wall. Each reactor vessel has two such holes which are located 180 degrees apart. These holes are used for access during construction and are subsequently closed by welding a plate over the hole. The covers and shroud support ledge are Inconel Alloy 600 material. The connecting weld material is also Inconel 600 (Alloy 182 or 82).

The high residual stresses resulting from welding, along with a possible crevic9 geometry of the weld, when combined with less than ideal water ouality, present aconditionconducivetointergranularstresscorrosioncracking(ISGCC). This has been recognized by General Electric and, as a result, they have developed a remately operated ultrasonic testing capability for detecting cracks in the cover plate welds. The first use of this custom ultrasonic testing fixture was at Peach Bottom Unit 3.

On January 21, 1988, intemittent short cracks were found in the weld heat-affected zone around the entire circimference of the covers at Peach Bottom Unit 3.

It is estimated that cracking exists over 50% to 60% of the circum-ference with cusps as deep as 70% through the wall.

It is believed that cover plate welds have not been inspected previously on any other BWR.

It is possible that the cracking is generic and may, therefore, affect all BWRs with jet pumps.

8B02010084 I

IN 88 03 February 2, 1986 Page 2 of 2 t'

Discussion:

General Electric has identified three concerns if failure of the access hole cover plates is postulated due to weld cracking:

1.

Loose parts - In the event of complete failure of the access cover weld during normal reactor operation, the slightly higher bottom heari area pressure would lift the cover out of its recess.

It would most likely fall to one side, but there is a potential for it to be swept irto the re-circulation pump suction line causing severe pump damage.

2.

Core flow bypass nonnal operation) - Loss of one or both cover plates would allow some r(ecirculation system flow to bypass the core jet pump discharge through the open access hole to the recirculation pump suction.

This flow transient would be readily detectable and would require reactor shutuown.

3.

Core flow bypass (Loss of Coolant Accident) - If the access hole cover plate welds were to fail as a direct consequence of a recirculation suction line break, the bypass path would prevent the emergency core cooling system from reflooding the core to the 2/3 level.

The core spray system would be capable of maintaining adequate core cooling provided there has been no degradation in the core spray piping.

No specific action or written response is required by this information notice.

If you have any questions about this matter, please contact the technical contact listed below or the Regional Administrator of the appropriate regional offict

~ d'eff b '

harles E. Rossi, Director Division of Operational Events Assessment Office of Nuclear Reactor Regulation Technical

Contact:

Warren Hazelton, NRR (301)492-0911

Attachment:

List of kecently Issued NRC Information Notices l.

t

-7 m

>w y

w

--q

o s

i GENE-523-1361092 i

j l

1 i

)

ATrACllMENT_Il I

f I.

l'

-22 l

- vw GENE-523-136-1092 ATTACliMENT B Preliminary Assessment of a 20.44 inch Loose Access Hole Cover Approaching a liatch 2 Reactor Recirculation Pump.

This is an assessment of what would happen if a marAole cover is carried by flow toward a Hatch 2 Reactor Recirculation Pump, and answer the question if the pressure boundary parts of the pump would be unacceptably degraded.

The postulate is that a 20.44 in. OD x 0.66 in thick plate (manhole cover) has entered the pump suction proper in one piece. Minimum (eye) diameter of the pump suction (per telecon with IU, the manufacturer of the pump) is also 20.5 in. Several scenarios are possible:

1.

He plate approaches the impeller nut (see sketch) at right angle to the centerline of the pump shaft, such that the plate becomes like a butterfly valve that is closing the suction of the pump. He core flow would become uneven similar to single loop operation. He operator would notice drop in core power, in this case the plate would be pushed against the impeller nut, which in turn would wear out a hole in the plate. However, this event would not go unnoticed, because the pump would be operating in supercavitiation mode with large vibration. He pump motor vibration switches would also alert the operator of the anomaly, if unnoticed for extended time the pump internal parts (impclier, wear rings, journal, hydrostatic bearings, seals, and throttle bushings) would sustain damage, in the extreme the pump would seize, and cause overload of the motor Re overload would cause motor to trip. His would terminate the event. The pump casing pressure boundary parts (volute, cover, and main studs) would survive this event, however, a visual inspection would be warranted. I l

f4 1

.v I

GENE-523 136-1092 2.

The plate approaches the impeller net diagonally, with the edge of the plate impacting and rebounding against the impeller nut, This mode of interaction between the plate and the impeller is more difficult to detect, because it would be intermittent, with the plate bouncing against the rotating shaft. On the other hand, this mode would not be as stable as the first case. The plate would eventually be rotated to choke the eye of the impeller.

This ensuing events and consequences described above would follow.

3.

The plate approaches the pump impeller and fractures to several pieces. There has been two occasions where large metallic loose parts were ingested by the pumps of the same size and design without damaging the walls of the pump casing or impeller. The loose parts (hammer, and 5" manhole cover) were ejected by the pump.

In a documented GE analysis it has been shown that in a hypothetical case that the pump impeller fractured in four equal pieces and struck the pump walls as a missile the pump wall was not perforated.

in summary: 1) various plant monitors would alert the operator to taku rtion,2) if unnoticed, the motor everload would trip automatically, and stop further damage,3) the puerp casing wall thickness can withstand large missiles without breach of pressure boundary as experienced in twc actual cases. It should be noted when the ace:. 4 hole cover is dislodged, it will be readily detected by the operator and plant shutdown initiated Thus it is very uniikely that the plate would reach the pump since the pumps trip on scram.,

Retrieved from "https://kanterella.com/w/index.php?title=ML20116J325&oldid=4286102"