ML20092A439

Rev 0 to N71 Pipe Rupture Evaluation

ML20092A439
Person / Time
Site:	Perry
Issue date:	01/31/1992
From:	CENTERIOR ENERGY
To:
Shared Package
ML20092A430	List: ML20092A426 ML20092A439
References
EA-182, EA-182-R, EA-182-R00, NUDOCS 9202100113
Download: ML20092A439 (40)

Text

_

s:

G/C REPORT:

EA-182 REVISION O PERRY NUCLEAR-POWER PLANT (TA #92-0002) s' TITLE:

N71 PIPE RUPTURE EVALUATION JANUARY 31, 1992 4

9202100113 920203 PDR ADOCK 05000440 9

PDR g.

a

+

G/C REPORT EA-102 REVISION-0 PERRY NUCLEAR POWER PLANT (TA #92-0002)

TABLE OF CONTENTS SECTION PAGE NUMBER I.

PURPOSE 1

II.

BACKGROUND 1

III.

PROCEDURE 1

A.

INITIAL EVALUATION 1

B.

ADC7TIONAL ANALYSES 2

C.

PIPE SUPPORT STIFFNESS 4

D.

HYDRODYNAMIC LOADS EVALUATION 5

IV.

RESULTS AND CONCLUSIONS 5

APPENDIX I INITIAL C/C ANALYSIS 1/3/92 APPENDIX II SUPPORT EVALUATION APPENDIX III HYDRODYNAMIC LOADS EVALUATION

G/C REPORT:

EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

=I.

PURPOSE:

The purpose of this evaluation is to provide additional documentation that the modified 36-inch Fiberglass Reinforced Plastic (FRP) elbow and nearby anchor, both located in the above ground _ portion of the N71 Circulating Water Auxiliary Condenser inlet piping, are acceptable for continued service.

In order to accommodate this-requirement, more detailed evaluations of the piping along with a rigorous evajustion of the failed anchor were performed considering both conservative operating conditions and an imposed displacement criteria based on field measured movements.

In addition to the inlet piping, the N71 Circulating Water Auxiliary Condenser outlet piping was also evaluated under system operating conditions to document the piping stress levels and anchor' loads presently existing in the outlet piping in the region of the abovc ground FRP elbow.

II.

BACKGROUND:

Initial evaluations of the failed N71 FRP elbow centered on performing a ecnservative evaluation of the stresses at the critical location of the-modified elbow.

Because this evaluation was intended to give-assurance that the elbow stresses were acceptable, a conservative method was used to maximize stresses by artificially displacing the anchorage to provide a displacement envelope which could be monitored-in the field to provide assurance that the FRP elbow stresses were within acceptable values.

At this point in time, the emphasis was on the fiberglass piping, not the long-term adequacy of the anchor support (IN71-H0013).

In the interim until such a long term evaluation was

-performed, anchor IN71-H0013 would be monitored via a baseplate scratch pad to provide indication of possible overload.

II

I. PROCEDURE

As a consequence of the_ required repairs to the 36-inch diameter fiberglass elbow in tae inlet line to the auxiliary condenser and the nearby anchor 1N71-H0013, G/C performed several analyses to 1

provide _ additional assurance that the system is operating long-term within a safe envelope.

These evaluations consisted of several phases as described below:

A.

INITIAL EVALUATION:

The initial analyses were performed by G/C shortly after the rupture of the fiberglass elbow to provide CEI with a pipe movement criteria which could be monitored at the site.

This PAGE 1

G/C REPORT:

EA-182 REVISION O PERRY' NUCLEAR POWER PLANT (TA #92-0002) criteria was_ produced using a truncated model of the Auxiliary

' Condenser Circulating Water-(CW) inlet piping with conservative boundary conditions to evaluate maximum stresses in the fiberglass 3

elbow._ Three separate values of 30, 50, and 70 lbs/in for the soil modulus of subgrade reaction were vtilized in order to conservatively estimate the maximum stresses in the region of the FRP elbow.

Initial results' produced by this model indicated that concurrent displacements of 0.125 inch in both the vertical and horizontal directions produced a maximum combined stress of 1923 psi in the critical region of the failed elbow which resulted in a factor of safety of_2 when compared to the long term strength of the FRP piping of 3800 psi.

A more detailed description of_the original model and the results of this initial evaluation are provided in Appendix I of this report.

B.

ADDITIONAL ANALYSES:

Based on this initial criteria, a monitoring system was installed by CEI on_the inlet piping to record the movements of the system at the flange connection to the FRP piping.

In conjunction with this monitoring system, it was decided to perform additional analyses of the system to more accurately determine the adequacy of the piping and anchor in the vicinity of the FRP' elbow.

These additional analyses included an expanded piping model of the Auxiliary condenser circulating Water (CW) inlet line to include the piping to the condensers and also portions below ground in order to include additional effects which could be influencing the stresses in the FRP elbow.

The below ground inlet piping model was extended for approsimately 35 feet to the connection to the 144-inch CW line.- An expanded model was also utilized to evaluate the loads and stresses existing in the CW outlet piping, and was extended-for a similar distance underground.

In addition to the operating case, other load cases were analyzed including a flow-transient case in order to assure that no significant dynamic loads occurred in the piping during system 1

operation,.and a target criteria case to envelope predicted worst case movements of the-piping.

cl.

INLET MODEL:

(Figures 1 and 2)

EXPANDED CONFIGURATION:

The. original model of the inlet piping was expanded to include the additional piping going to the auxiliary condensers and additional underground piping to account for displacements PAGE 2

e

.G/C REPORT:

EA-182 REVISION O PERRY. NUCLEAR POWER PLANT (TA #92-0002) which_may be influenced by the buried FRp piping thermal and pressure effects.

Because of the importance of determining i

the nature of the. existing anchor loads, a detailed finite element-model of the support 1N71-H0013 was constructed to provide detailed spring constants for use in the p,iping analysis.

As described below, this model was also used to-evaluate the imposed piping forces and moments resulting from

(

the various loading conditions.

The loads on the anchor were evaluated utilizing a STARDYNE finite element model as described in Appendix II. _ Localized stresses and the weld at the pipe to anchor connection were evaluated separately utilizing the WERCO computer program.

+

SOIL PROPERTIES:

The initial piping analysis utilized several values of horizontal soil modulus of subgrade reaction.

Based on the i

results of this initial evaluation and a more detailed review of the analytical data, the expanded piping models were evaluated for two different pairs of values for the soil modulus of subgrade reaction in the horizontal and vertical directions.

A furtner review of the resulting analyses determined that the most critical support loads were caused by the stiffer soil properties.

Based on this evaluation, all expanded model piping load cases were run with the stiffer-soil properties.

OPERATING CONDITION:

The expanded inlet model was run for a design / operating condition with a 4T of 65'F (30-95*F) above ground, 40*F below ground ~, and a pressure of 60 psig.

Horizontal and vertical soilspringswereincludedcorrespondinggothevaluesofsoil vertical and 70 modulus of subgrade reaction of 135 lb/in

.lb/in3 lateral as previously determined.

2.

OUTLET MODEL:

(Figures 3 and 4)

OPERATING CONDITION:

l A separate model was utilized for the Circulating Water Auxiliary Condenser outlet piping because of differences in configuration both above and below grade.

The outlet piping was analyzed = atla 12*F higher temperature than the inlet r

piping and at a pressure of 39 psi.

As was done for the inlet piping, concurrent maxirum loadings resulting from a flow

'~

transient ~ analysis were includea in the analysis.

)

PAGE 3

..m...

_. _ -.. _ - - _ _ _ _ _ __ _.,m _.....,___ -- _._ _._ -_, -

G/C REPORT:

EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002) 3.

INLET / OUTLET MODELS:

FLOW TRANSIENT LOADS:

Because of the concern that'the initial rupture may have been caused by an unanalyzed loading due to flow transients in the wystem, a detailed evaluation of the syatem operating modes was made. -It was determined that the o..ly flow transients that were possible were those resulting from pump switching l

operations in the CW system.

The evaluation of these transients are described in detail in the following section.

As may be seen, the magnitudes of the forcing functions obtained from this analysis-are relatively small; however,-

they were included in the piping analysis for completeness.

Conservatively, maximum loadings were applied concurrently to the pipingLsystem at each change in direction.

The resulting loads and stresses were combined absolutely with the loads and stresses obtained in the deadweight and thermal analyses.

4.

INLET MODEL:

DISPLACEMENT TARGET CRITERIA:

The expanded inlet model was run with forced displacements at the flange location to better define the original criteria with respect not only to stresses in the FRP elbow but also the resulting loads on the anchor 1N71-H0013.

As a means of providing a criteria to monitor operation of the system, displacements were induced in the piping models corresponding to movements of 0.125 inches in the vertical direction and 0.135 inches in the horizontal direction.

These displacement values ~were supplied by CEI as conservative envelopes of monitored displacement data.

The-loads on the anchor were evaluated as described above utilizing a-STARDYNE finite element model.

Localized stresses at the pipe to anchor connection were evaluated separately utilizing the WERCO computer program.

Because the target criteria case is considered to only assure that pressure integrity is maintained in the piping-system, only primary loads due to deadweight and flow transients are evaluated.

The1 displacement loadings are considered to produce secondary

~ piping stresses which are self-limiting and are therefore not included under primary stress evaluations.

C.

PIPE SUPPORT STIFFNESS AND ANCHORAGE EVALUATION FOR PRYING EFFECTS:

PAGE 4

G/C REPORT EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA /92-0002)

See Appendix II of this report for a discussion of the methodology and analyses used to evaluate the pipe support anchors in the Circulating Water Auxiliary condenser inlet line (Mk No IN71-H0013) and outlet line (Mk No IN71-H0021).

D.

HYDRODYNAMIC LOADS EVALUATION:

see Appendix III of this report for a discussion of the hydraulic analyses performed to determine potential hydraulic loads (Impulse and transient) in the 36-inch diameter circulating Water Auxiliary Condenser piping.

IV.

RESULTS AND CONCLUSIONS:

A detailed analysis of both the inlet and outlet N71 Circulating Water Auxiliary Condenser piping has demonstrated ths adequacy of the piping and anchor under operating conditions, including all possible flow transients.

The I sximum stress occurring in the inlet piping TRP elbow region above ground Vas determined to be 2014 pai resulting in a factor of safety of 1.9 When compared to the long term strength of 3800 psi.

The maximum stress occurring in the outlet piping FRP elbow region above ground was determined to be 2232 psi resulting in a factor of safety of 1.7 when compared to the long term strength of 3800 psi.

As a means of providing a-criteria to monitor operation of the system plus to demonstrate added margin, displacements were artificially induced in the piping model corresponding to movements of 0.125 inches in the vertical (upward) direction and 0.135 inches in the horizontal (North) d$rection.

Under this envelope condition, it has been demonstrated that the maximum stress in the above ground FRP is 1948 psi resulting in a factor-of safety of-1.9 when compared with the long serm strength of 3800 psi for the FRp material.

The anchor evaluation for the target criteria case has shown that the anchor components cannot be shown to be adequate for the full amount of criteria displacement.

The maximum allowed displacement hat is acceptable based on standard design critoria (Reference 1, Appendix II, Part B) for the anchor has been estir.rted to be 0.081 i

inches in the horizontal (North) direction and 0.075 inches in the vertical (upward) direction when measured at the flange location.

The displacements at which the anchor components achieve functional 'imits have been estimated to be 0.115 inches in the horisontal direction and 0.106 inches in the vertical direction, again measured at the flange location.

PAGE 5

N71 CIRC UATER'TO AtlX' CONDENSER (OPERATING. II)

CIRC-1HP-YW s

Z x

L

~

l 4

r O

i m

mRE mE 147I-Hoo r 3 m

3 w*

\\(

i20 N

53 68 M63 2D 0.0963 SPECIFIED DISPLACEMENTS (INCil.DEG) y..

i

,O CIRC-IHP M71 CIRC UATER TO AUX CONDENSER ~(OPERATING. II) y:

R 2

>X (M7 f-HOOl3

.[L fdi Ei F115 128 s-S a2 s

1 '44 m.

19G m$

K[$

bidH bl#

Y 58 153 168 163 17 0 173 175 J\\

.A.\\.

.A\\

Q l

w l

h

FU PilMP TilHB COND-COOL UATER' PIPE (4. DESM /0PER)'

CIRC-OHP

'R IL z/\\x g

/

- 1 uns-noo2 s y

x=

28 03,4 s

13 dd G

50 6

?'

7

/

s.--

.,.,u._

,____.___...__________.__.______..,__,,_n,..

q:

4

.e.

i 3 h

i t; w=

- o S

le

' /

We C

b c

L, /

E

- W L

- Q w

K W

- CC O

CN

- K W

HC3 05 0

- w NZ FIGURE 4 PAGE 3D

~e-g erm

,n4 p

-r m

--,a-w e.

a

,--n.-

e,

,e--

a w-

PORT:

EA-182 REVISION O m POWER PLANT (TA #92-0002)

APPENDIX I-INITIAL G/C. ANALYSIS OF FIDERGLASS INLET PIPING 1/3/92 OBJECTIVE The' purpose of this analysis is to evaluate the potential stress levels in the reinforced fiberglass portions of the pNPP FW pump drive turbine cohdenser. cooling Vater piping for a variety of-environmental-conditions.

The_objectiva is to provide CEI with some pipe movement criteria-that can be used in monitoring the system to provide reasonable-assurance that the piping is operating within a safe envelope of deflection.

The results of this analysis are to be considered approximate due to the fact that the_ procedures used are typical of those used to analyze

'oteel pipe and.may not be entirely appropriate for reinforced fiberglass piping; however, the results are adequate for the intended ~

purpose of defining an acceptable pipe movement criteria. -Further, the i

results are influenced by the variability of the modulus of elasticity for fiberglass pipe depending upon the orientation of the reinforcing-

= fibers and the-possibility that properties of the existing pipe may-

=have~ changed over time.

ANALYSIS ~ BOUNDARIES 9

The analysis was. performed'with the use of the CAEPIPE program, a PC-based general ~ purpose pipe stress program.

The model consisted-of a truncated section of_the pipe running from a rigid hanger 1N71-H0014 at

' Elevation 634' to the buried fiberglass elbow at Elevation 608'-7".

The:model includes the new-fiberglass elbow at the transition-to the carbon steel pipe north of anchor HO13.

INPUT. DATA The following raterial characteristics and environmental conditions l

were incorporatsd into the analysis:

i:

1.-Material properties:.

Steel' pipe p_er A,,_ljlif, Grade KC 60 CL2 e

l E=27.9x106 psi a=6.07x10~ -in/in/_'F Fiberclass PAGE 6

~,

_m O/C REPORT EA-102 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002) 6 Elbow - Flexural modulus E=1.29xig psi Pipe. - Flexural modulus E=1.2x10 psi Elbow - Expansion coefficient a=12.0x10-6 in/in/*F Pipe

- Expansion coefficient a=11.1x10-6 in/in/*F

-(The tensils or hoop moduli are not considered in the CAEPIPC-program.)

2. Environmental conditions:

Temperature range ST=65'F above Elevation 615'-1" Temperature range ST=40*F below Elevation 615'-1" Pressure P=f8 psi Soil modulus of subgrade reaction X = 30-70 lb/in/in a (Soil springs corresponding to 30, 50, and 70 lb/in/in2 were used in order _to bracket the range of k.)

.3.

Imposed movements at Anchor HO13:

Movements of 1/8" northward (+x) and 1/8" down (-y) were selected as reasonable boundary limits of motion for the pipe at the anchor l nearest the flange connection between the fiberglass elbow.

.The northward movement corresponds to the diametric clearance in the holes fcr the anchor bolts in the HO13 baseplate, and a vertical movement of similar magnitude was arbitrarily chosen.

The directions of these movements-(northward and down) were chosen conservatively so that they would be additive to the effects of thermal expansion.

A run'with-the anchor moving 1/8" up instead of down was made in order to confirm that the assumption of downward movement was more conservative.-

ANALYSIS RESULTS

-The following results are considered approximate and tentative at this time, pending review and verification of the inputs and calculations.

The_ highest stresses in the fiberglass piping occur at the top of the vertical section of the 36.8" diameter run where it connects to the tapered transition section on the bottom of the new elbow.

This is due to.the-telatively high bending moment at the critical section combined with'the minimum section modulus where the thin wall (0.400") ends.

.The results of the analysis, which include the combined _ effects of weight, thermal expansion, and imposed movement at the IN71-H0013 anchor are as follows (based on soil k=50 lb/in/in2):

1. Movements at points'of interest PAGE 7 a

i G/C REPORT:

EA-182 REVISION O PERRY NUCLEAR. POWER-PLANT (TA #92-0002)

Point Vertical movement North-south movement Anchor HO13.

.125" (imposed)

.125" north (imposed)

Flanges

.121" down

.163" north Critical section

.206 down

.099" north

'(Top of thin wall section) 2.-Axial force and bending moment at the critical section Axial force Fy=10,948 lb in compression Bending moment M =19,816 ft-lb (237,792 in-lb) g 3.

Combined longitudinal stress at the critical section Longitudinal stress S=1911 psi The effects of varying the soil modulus of subgrade reaction by more than 100% are slight, resulting in a total variation in combined stresses of only 2% at the critical section.

The critical section compressive force, bending moment and combined stress for various values of X are as follows:

Soil modulus ]s Axial foreg Eendino agment C_qm11ILedi stres_q 2

30 lb/in/in*

10,045-lb 18,854 ft-lb 1883 psi 50 lb/in/in a 10,948 lb 19,816 ft-lb 1911 psi 70 lbfin/in a

11,504 lb 20,221 ft-lb 1923 psi In order to. confirm that a downward movement of 1/8" at Anchor HO13 was more critical'than an assumed upward movement of 1/8", the analysis model was modified tos impose an upward movement instead-of a downward movement.

A comparison of the results for the critical section is as follows: (soil k=50 lb/in/in*, HO13 movement unchanged at 1/8" north) :

Hv1 c; HO13 Axial f.orce Bendino moment Combined EtreSR 1/6" down 10,948 lb 19,816 ft-lb_

1911 psi 1/8" up 2,175 lb 7,794 ft-lb 1561-psi The maximum single component of the combined stresses is the longitudinal pressure stress of 1291_ psi in tension, due to the internal prersvre.of 58 psi.

CONCLUSIONS PAGE 8

G/C REPORT:

EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

For postulated movements of 11/8" vertically and 1/8" northward at the IN71-H0013 anchor, the stresses at the most critical section of the fiberglass piping do not exceed 1923 psi Which results in a factor of safety of 2 when compared to-a 3800 psi long term strength for the FRP piping.-

These movements are considered to be the maximum credible movements at the location of this anchor, considering liberal installation tolerances at the anchor.

Since the largest single component of the critical pipe stress is the longitudinal pressure stress of 1291 psi, the maximum combined effect of thermal expansion and postulated anchor movement is only 632 psi.

Therefore any movement of less than 1/8" in any direction at the anchor or adjacent flange is considered to be trivial with respect to stresses in the fiberglass piping. Wide variability in the soil modulus of subgrade reaction has been shown to have little effect on forces, moments, and stresses in the piping.

l j..

PAGE 9

O/C REPORT:

EA-102 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002) d APPENDIX II i

PART A

\\

SUPPORT BTIFFNESS AND EVALUATION F

This section documents the methodology and results of'aaalyces used to determine the support stiffnear at pipe supports It171-h0013 and 1N71-H0031, and the suppe-t c/aluation of loads obtai, from the piping analysis.

Tho support stiffness in utilized in t.

,' ping analysis to i

obtain the support reactions and the resultant pipe stresses.

MOD.T:

The two pipo supports are very similar in design; hcwever, there are minor differences in the actual as-t'.11t locations of thq Drillco Maxibolt anchoragos and in the fabrication details of the vertical j

member which attaches the 36-inch Circulating Water pipe to the support baseplate.

These differences are not significant and as a result only one finite element model is used to determine the stiffness and evaluate the support.

The support is modelled with the STARDYNE computer program.

A sketch of the model is shown in figure 11-1 along with the material properties utilized in the analysis.

STIFFNESS:

The support stiffness is calculated by imposing loads at the conterline of the 36-inch pipe in each of the six degrees of freedom.

The resulting stiffnesses used for the piping analysis are as follows:

TABLE 11-1

.RECTIO!1 STIFFNESS (kip /in)

X-TRANS LATIO!!

197.4 (149, target crit)

Y-TPANSLATIOtl 1140.0 2-TRANS*_.ATI ON 338.9 DIRECTION STIFFNESS (in-kip / radian)

X-ROTATION 305044.

Y-ROTATIO!I 363459.

Z-ROTATIC.

151404.

SUPPORT LOADS:

The piping analysis was initially oerformed for uoveral values of PAGE 10

i O

8 O/C REPORT EA-182 REVISIO! 0 PERRY NUCLEAR POWER PLANT (TA #92-0002) hori: ental soil modulus of subgrade reaction.

Dased on the results of this initial evaluation and a more detailed review of the analytical data, the expanded piping models were evaluated for two different pairs i

of values for the soil modulus of subgrade reaction in the horizontal i

and vertical directions.

A further review of the resulting analyses i

determined that the most critical support loads are caused by the stiffer soil properties.

Based on this evaluation, all piping load cases were run with the stiffer soil properties.

I The operating, fluid transient, deadweight, and target criteria loads are presented in table II-2.

The loaa combinations used for the STARDYNE ccmputer analysis are presented in table 11-3.

1 i

i 1-l l

L

~

PAGE 11

~

i s

TABLE !! 2

$UMMAtt OF SA$lt LOADS ON $UNPott ME h38. N001) & ND021

.................................................s..........................

$UPPot1 ME h0$. N0013 8 N0021 LDAD CA$E SUPPott LOAD $ IN PIPihG C00tblhaft silllM (lbs, f t)

FX Ft F2 Mr Mt M2 (lbs)

(Lte)

( Lte)

(f t lts) (ft lbs) (ft lba)

DEAD WT 995 17376 19 222 49 3301 Ot$W/0P (H0013) 2118 3591

• 277 4561

!W8 1 73 DE$v 0P (N0021) 3t4 8277

+ 3M 5616

+3521 1280 ftAh5 (>0013) 1705 1245 1

135 238 477 ftAks (H0021) 1003 413 14 266 94 fLANGt MVMT 2 19928 383&6 66 1309 445 8230

(.135"W,.125"))

................................a.............................................

TABLE II 3 SLMMARf 0F COMB 1NED tMDS CJ $UPPORf Mt WOS. N0013 & N0021

$UPPOR1 Mr Mos. N0013 & N0021 LOAD CA$t LOADS lei PIPino C00E0 st$f tN AT CENf tt CF 36" DI AM PlPt itMAtt$

fx FT F2 MX MY M2 (ktps)

(k(ps)

(kipe) (in tip) (in kip) (in ktp)

..............,.........................+................

FX, FY, M2 Most Postfivt Come N0013: OPtt 3.823 2.166 0.276 56.352 33.120 3.648 F2, Mu, NY..". MAXlMl!! MAGNlflOE

+/* FLUID ftAkt N0021: OPtt 1.387 7.864 0.378

+66.816 45.444 14.232 FM. FY, M2 ". W>$f PollflVE COMB

• /. FLUl0 1tAkt F2, ME, NY ". MAXIMlZE MAGN!ftCE OtAD +

20.923 21.010 0.047 13.044 4.M2 59.148 (1

1.00 L1*(FLG MVMi 2)

DEAD +

17.934 15.252 0.037

'Q.688 3.951 44.334 K2 0.85 (2*(FLG MvMT 2)

OfA0 +

14.945 9.494 0.027 8.332 3.150 29.520 (3

0.70 (3*(FLG MVMT 2)

DEAD +

12.952 5.656 0.021 6.761 2.616 19.644 to.

0.60 K4*(FLG MVMT 2) i DEAD +

10.959 1.817 0.014 5.190 2.082 9.768 t$.

0.50 l

K5*(FLG MVMT 2)

OtAD +

8.966 2.022 0.007 3.619 1.548 0.108 K6 0.40 K6*(FLG MWT 2)

.........................................u...............................................a..........ca........

1 PAGE llA 1

l l

l l

l

!7til MCCJt.tn Of (US"/C8,J

• N:CC % t 9:t:CCA/ AA /C t C,50 i

.DENftTY

• ES4 */c'4 %s C*:sf4*/C/ENo# Cf ikf4AML, SN,%'!/;t)

• 6. Set 0%4y?:

CCAJCA!Ti

~

/,%"p P/f'f -y C:4 cdi Ti CsN A CEC S,y f $7Af,M1N

• 3 0 0 0,o st' MCDULV5 Cf f4 AC7*/C/Tf

3/22 %

~

ibt$50u AAT C s 0,/ 7

't 'a

/G Y $CH $O P/f[ ^c

• Q ~ /A/0/C47[Z /a4/O A EA H 3 1

/

/ft5

{*4

% <*q,

N X

N N

e

/f3

,,,/

/"/7/X /[ ]

k d'

s6 V

s p,

[

d,q

/y

.: xt N/-l N

s x

m 15

,r vi to; 0

/6#

@/,'

N pl I

(8 N

/E; ffi

$,3

)

"D g%

N

,r'

/

e t, x

j/

JI

.I i

'xsy y

a X

,/

FIGURL: 11-1 PAGE llB

I G/C REPonT EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

APPENDIX II PART D ZIPE SUPPORT ANCHORA!1E_ EVALUATION FOR PRyMG RFFECTS Pit a support base plates attached to reinf orced concrete structures using concrote expansion anchors are ovaluated in accordance with the pNPP anchorago design guide (Ref. 1).

The particular type of anchors used to anchor Pipe Support MX Nos,1N71-H0013 and 1N71-H0021 are DRILLCO Maxibolts.

These anchors require significant proload for propor installation which tends to complicate the determination of the factor of safety provided in the design with regard to pullout.

A straight forward analytical proceduro has been developed by C/C for determining the anchor's " engineering tension load" Which properly accounts for prying effects.

The procedura provides a pesctical solution to the problem of high anchor tensile forces due to preload effects, which in some cases, may approach or exceed allowable tension loads befcre any external loads are applied.

An outline of the procedur( in given in Tablo II-1.

The STARDYNE computer program (Ref. 2) is used in this task to perform the finita element analysis described in Step 1.

This program is used throughout the nuclear industry to perform base plate analyses.

The rigid plate analysis required in Step 2 is analogous to a reinforced cenerate beam analysis by worXing stress design acthods in which a plane section remains plane.

Also, this conventional cracked section analy, in method conoiders only the compression resistance of the concrete and the tensicn resistance of the anchors.

Prying effects are not considered in the rigid plate analysis.

Tho and result (step 3) of the procedure outlined in Table II-1 is the engineering tension force in the anchor.

This force does not include proload tension which is consistent with engineering design practice for all types of connections.

It to also consistent with regulatory requirements and does not account for prying effects.

I Resultant shaar forces on each anchor can be determined using i

conventional statics methods. they can be obtained from the finite element analysis results if proper restraints are included in the model.

In this task, anchor shear forces are calculated manually for each of the four DRILLCO anchors on each base plato.

The interaction ratio for shear and tension loads on the anchors is calculated using the method outlined in the PNPP anchor design guide which is Attachment No.

3 to Reference 1.

Allovable tension and chear loads are obtained by dividing the SSE design allowables in Table 2 by l

l PAGE 12 l

l

)

,~

O/C REPORT:

EA-102 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002) 1.4 which provides the appropriate f actor of safety for the loading conditions ovaluated.

IADLE II-1 1

Procedure for obtaining the Engineering Tension Load in Preloaded Anchors

1. Perform a finito olement enalysis of the base plate using conventional industry methois.

The model accounts for the compression only contact surfaco between the underside of the base plate and the concrete surface and the tension only restatance provided by the anchors.

Toutdation flexibility and plate flexibility effects are also a: counted for in this type of analysis.

In addition to the design load cases, include a load case in which no external loads are applied so that effects of praload only can be obtainod.

Raoults of the finite element analysis are used to obtain values for the following variables.

Input Variables:

To = anchor " lift off" load from design specificationa Eb = anchor stiffness in tonsion from design specificationn FEA output variables do = anchor displacement due to preload only di = anchor displacement due to proload plus external loado for load case aia calculated Varinbles:

1 Tbo anchor tension after elastic comproosion of foundation due a

to preload and without external loads Tdo = To + K *do d

Tba = anchor tonalen due to preload and external loads Tba = To + K adi b

K, = ABS (T /d ) - K= equivalent foundation stif fness in the anchor vicinity Ke en o o b

2. Perform a rigid plato analysis to obtain anchor tennion forces, Tbre for each of the design load cases.

Effects of foundation flexibility are included in this analysis.

I

3. Using results of the finite olement analysis and the rigid plate analysin, the following calculations are made to determine the anchor's engineering tension force, T be' PAGE 13

l O/C REPORT:

EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA /92-0002)

a. Calculato prying forco, Tbp bp = T a - (Tbo + Ter/(1+Kco/K ))

T b

b

b. Calculato ongineering tension force, T eb Tbo = Tbr + Tbp 5Tba

REFERENCES:

1. G/C, Inc., Structural Dept. A0414 Dosign Calculations, Perry Nuclear Power Plant, " Design Guide for Anchors", Calo. ID / 1 39.1, WO/ 04~

5250-716, Rev.

O, 4/2/91, with Attachment Hos.

1, 2 and 3.

2. GMC, "STARDYNE User Information Manuala, Cenoral Microelectronico Corp., San Diego, CA, 190".

Softwarot STARDYNE, Version 3.5(R), HAY 1, 1989 Hardware: G/C Personal Computer, IBM compatible (386) l PAGE 14

G/C REPORT:

EA-102 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

APPENDIX III HYDRODYNAMIC _ LOADS EVAkD TION:

This report documents the results of hydraulic analyses performed to determine potential loads on the three foot diameter Circulating Water Auxiliary Condenser piping for the Perry Nuclear power Plant.

The Circulating Water system was modeled using the HYTN41 computer program.

HYTN-41 is a general purpose thermal / hydraulic system analysis computer program.

It can be used to simulate both transient and/or steady state non-isothermal fluid flow in a network.

The effect of pump start /stop, valve optn/ closure and other system control operations can be modeled.

R r the transients addressed below a transient simulation of the Cit ulating Water system was performed.

The system model includes the pumps N71-C001A,B and C.

The pumps were modeled using homologous curves which provide flow, head, speed and torque characteristics in all four regions of possible pump operation (normal, reverse speed dissipation, dissipation zone and turbine zone).

The remainder of the system was hydraulically modeled including the pump discharge valves (F020A,B, and C) flow coefficient vs stroke, the 12 ft.

underground piping to the main condenser and associated piping, the discharge piping and the losses associated with the cooling tower.

In addition to the condenser loop, a simplified model of the cooling tower blowdown was included.

In addition to the primary loop, a detailed model of the Auxiliary condenser, inlet and outlet piping, and motor operated inlet valves (F150A and B) was integrated into the overall model. The HYTN41 simulation is used to determine unbalanced piping loads on the Auxiliary Condenser inlet and outlet piping for three different scenarios.

The method used to-develop axial piping unbalanced forcing functions is attached.

The System operating Instruction (SOI-N71) for the Circulating Water System was reviewed for any system operating transients which might cause water hammer.

No operational transients were identified that would create a water hammer.

Initial till of the Circulating Water system is very carefully controlled to avoid starting a pump or opening a valve with flow into a voided pipe.

The only operational transients which.will have any inertial affects, i.e.

surge involve starting or stopping a pump or opening or closing a valve with the system full.

Tre three scenarios evaluated include; PAGE 15

9 0/C REPORT:

EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

1. Normal system operation with three pumps running.

one pump trips and its associated discharge valve closes.

2. System operation with two of the three pumps running.

The remaining pump starts and its associated discharge valve opens.

3. Normal system operation (three pumps operating).

Auxiliary Condensor Valve F150A closes.

Unbalanced piping forces on the Auy111ary Condenser inlet and outlet piping were calculated for each of the above scenarios.

The results of the analysis are presented in Table III-1 r

F b

i PAGE 16 l

l G/C REPORT EA-182 REVISION 0 PERRY NUCLEAR POWER PLANT (TA #92-0002)

TABLE III-1 MAXIMUM PIPING LOADS Pipe Segment Scenario 1 Scenario 2 Scenario 3 Maximum Load Maximum Load Maximum Load Pos.

(Neg.)

Pos.

(Neg.)

Pos.

(Neg.)

Lbf.

Lbf.

Lbf.

i i

Inlet 1 188.

(327.)

282.

(379.)

188.

(175.)

Inlet 2 73.

(138.)

103.

(141.)

76.

(72.)

Inlet 3 82.

(120.)

194.

(175.)

89.

(86.)

Inlet 4 61.

(135.)

163.

(140.)

65.

(66.)

Inlet 5 200.

(508.)

585.

(455.)

206.

(230.)

Inlet 6 59.

(170.)

178.

(162.)

58.

(64.)

i Inlet 7 18.

(44.)

45.

(41.)

14.

(26.)

Inlet 8 7.

(24.)

19.

(18.)

22.

(7.)

t i

outlet 1 9.

(24.)

14.

(17.)

22.

(5.)

Outlet 2 16, (43.)

4 4. -

(39.)

12.

(25.)

Outlet 3 46.

(117.)

75.

(89.)

59.

(25.)

Outlet 4 115.

(244.)

177.

(190.)

75.

(70.)

outlet 5 50.

(105.)

70.

(88.)

37.

(37.)

outlet 6 50.

(131.)

80.

(110.)

37.

(37.)

i Outlet 7 69.

(181.)

111.

(147.)

46.

(51.)

Outlet 8 680. (1760.)

952. (1342 )

73.

(146.)

Outlet 9 721. (2142.)

1350. (1776.)

838.

(791.)

Note the piping segments are numbered consecutively from the 12 foot diameter line to the Auxiliary Condenser and

• rom the Auxiliary Condenser to the 12 foot return line with the loads occurring at each change in direction.

f The analysis is documented in Calculation 2.6.14 Rev.

O.

t l

l l

l PAGE 17 l

t

0/C REPORT EA-182 REVISION O PERRY NUCLEAR POWER PLANT (TA #92-0002)

IHB!LALL9Apit Although the steady state, impulse loads due to momentum changes at the elbows are normally considered insignificant, calculation N71-08 was prepared to determine the magnitude of these loads.

These loads are due to the fact that as the fluid changes ' direction at the elbows, a force is imparted to the fluid so-as to maintain the conservation of momentum.

There is also a reactionary force imparted to the piping that is equal and oppocite in direction to the force on the fluid.

These impulse loads are over and above the pressure loads, which impart a force of P x A.

Calculation N71-08 documents that these forces are indeed small and are comparable to a 1.7 psig pressure force.

Although small, this I

additional load has been included in the piping stress analysis by utilizing a conservative operating pressure of (58+2) = 60 psig for the inlet piping and (37+2) = 39 psig for the outlet piping.

i e

1 b

l

+

PAGE 18

h ATTACllMINT 2 ENCL 45URE 2 SUNNARY OF SIGNIFICART ISSUES RECARDING AUX 11.IARY CIRCU!.ATING VATER PIPE RUPTURE I

L l

l l

I l

~

PY-CEl/Olt-0.188 h Page 1 of 13

SUMMARY

OF SIGNIFICANT ISSUES REGARDING CIRCULATING VATER PIPE RUITUltE I.

CIRCULATING VATER (N71) PIPE RJPTURE On 12/22/91 at app: utmately 0200 hours0.00231 days <br />0.0556 hours <br />3.306878e-4 weeks <br />7.61e-5 months <br /> th) 36 inch anxiliary circulating vater supply line catastrophically failed. The fallute vas located in a fiberglass e4bov in the pipe just prior to the point where the pipe transitions from fiberglass to carbon steel. The pipe was located in the yatd en where the pipe exits the around prior to entering the heatet bay building. As a result of the failure, approximately 2.87 million gallons of water spilled into the yard area, the plant underdrain system, i

and into other plant attas. The plant was shutdovn until repairs could be completed =

Probable Causes Of Failure The plant staff immediately contacted a fiberglass piping consultant to evaluate the failute. This individual had been used in other issues surrounding fiberglass piping on this site and others.

Shard mapping of the failed fiberglass pieces was completed.

Following mapping and photo documentation, the pieces vere collected for rewnstruction to determine the f ailure nicchanism. Adjacent piping, supporta, and adjacent structures vere thoroughly inspected as part of the toot cause effort.

Failed bolts from the piping suppcrts were analyzed at Centerior's analysis laboraiotles.

Several elements of cause vere identified following the above efforts.

These causes were examined for their respective contribution to the failure. They are as follovst A.

ADDED STRESS ON FIBERGIASS PIPE AS A RESilLT OF AN IMPl.0 PERI.Y SPACED 0-RING PLANGE SEAL.

An 0-ring supplies the sealing mechanism between the flanged portion of the fiberglass piping and the flange of the carbon steel piping.

The 0-ring sits in a 316 stainless steel 0-ring tetniner. The 0-ring and retainer ete manually positioned in the center of the flanged connection and the flanges are bolted together so that the 0-ring meets the flange faces to provide a pressure seal. When installej correctly, the 0-ring contacts the one flange face and there is an approximate.0625 inch gap between the 0-ring tetainct and the other flange face. This allows some amount of telative motion of the flange faces.

Inspection of the installed 0-ring shoved that the flange faces vere pulled together to the point that they were both contacting the 0-ring retainet. Thus, any steel flange motion vould be more dilectly transmitted to the fiberglass piping than originally intended by 0-ring flange seal design.

Initial judgments vete that this situation may have induced appreciable additional stress into the fiberglass piping such that this item vas one of the primary causes of the eventual piping

failure, llovever, follow-up piping analytical votk has shown the i.

PY-CEl/010-0388 L Enclosute 2 Page 2 of D relative insensitivity of the fibeiglass piping to imposed displacernents of such magnitudes thus, the overall contribution of this item is judged to be less important than originally consideted.

B.

STRESS CONCF#11 TAT 10N IN TitP. El.lLOV DUE TO AN APPROXIMATE 8 INCil NON DESIGN AXI AL GR00VP. IN Tile El.POV TO PIPE BONDING ZONE The vettical section of the failed elbow is joined to the adjacent pipe by a butt type joint. A butt joint is joined together by a sleeve. Typically, the bonding atens ate prepared and a primer coat is applied and cured.

Following cute of the primer, the sleeve joint is applied to the prepared surfaces and cuted.

Upon cure, the sleeve joint attains the approximate physical propet ties of the pipe and provides circumferential suppeti and axial coupling of the pipe and elbow. The maximum thickness of the sleeve is at the butt joint.

Ptom this maximum thickness, the sl+ eve is tapered for a distance of approximately 20 inches on eithet side of the butt joint on the parent pieces.

A typical elbow is composite material approximately 1/2 inches thick. The interior.090 inches of the elbow is a resin rich mixture which provides good cottosion tesistance but little strength. The next approximately 0.400 inches of thickness is composed of voven roving and mat glass fibers in a polyester resin

matrix, This ptovides the attength characteristics of the elbow.

The exterior.010 inches is also composed of a corrosion tesistant but relatively veak material.

Two axial grooves vete observed on the exterior of the failed elbow. These axial grooves appeared to be made by a high speed grinder during original construction of the slbow or during initial installation.

The axial glooves vere in the Vest by Northwest quadrant of the vertical section of the elbow approximately at ground level. 1he axial grooves were approximately 60% through the vall thickness of the elbow. The existence of an axial groove in the elbow tesults in both a net cross-sectional area reduction in the load bearing portion of the elbow and the creation of a stress concentrat'on 3oint at the notch. The end. result of the groove is a reduction in the hoop strength of the elbov. This reduced strength / stress tiser point was located near ground level, vnere soll backfill around the piping causes a relatively highly loaded area vithin the fiberglass piping.

Inspection of the failed elbov revealed that the groove vae an initiation site for the initial rupture in the elbow. The characteristics of the tear indicate that it initiated at the groove and traveled axially en the elbov from that rone. The failure vas relatively clean at the initiation point.

As the failure continued through the elbow, it changed from a " clean" fracture to shredded jagged tears.

It is probable that this secondary tearing occurred due to the hydrodynamic force of the water after the pipe initially ruptured. The axial groove within the butt joint bonding area is concluded to be a flav vith primary causative influence leading to the eventual piping failure, with the subsequent elbov damage most probably emanating from the initial flawed atea.

PY-CEl/0!r-0388 b Attochment 2 rnelosuse 2 page 3 of 13 C.

ADVERSE PIPE StJpPORT INFLtlENCE The fitst support (IN71-110013) on the N71 stcel piping to the south of the fiberglass-steel interface was intended to function as an anchor point.

As such, it should have rigidly held the 36. inch diameter steel piping so that minimal loads / displacements vete induced into the fiberglass elbov from the steel piping side of the interface.

Subsequent to the 12/22/91 event, inspections indicated significant damage to support IN71-il0013. All fout (4) existing anchois (3/4" diameter HILT 1 drop-in anchots holding the suppott to a concrete slab) vere fractured. The broken pieces were removed and fotvarded to Centerior's testing laboratory for fallute analysis.

Refer to

-Attachment 2, Enclosure 1, Addendum A for a_ completed copy of the analysis / test report. Observations / conclusions from this analysis ate summarized as follows:

1.

There are some observed fatigue cracks in the bolts located avay from the fractured surfaces. There is also se'ne evidence t

of corrosion influence on this cracking.

It is not possible to definitively deteamine if the fractured surfaces ate fatigue driven.

Hovever, out judgment is that they are not caused by fatigueduetothepresenceofsignificantboltiMstic deformation prior to fracture (fatigue failure vould typically be more of a brittle type failure with little distortion while overload failure vould typically have relatively latge distortion).

Refer to Att:chment 2, Enclosure 1. Addendum S for photographs of the failed bolts.

It therefore follows that the fiberglass piping failed finst, with subsequent anchor bolt I

failure due to extreme overload.

2.

Evidence exists that indicates that the nuts on the anchor bolts were " loose" prior to the piping failure.

This looseness permitted ths 1/2" thick support baseplate, and thus the entire support, to be relatively free to displace minor amounts (essentially within bolt hole clearances).

Thread damage on the bolts indicates a long term " hammering" action caused by lateral (horizontal) movements of the baseplate.

The amount of probable lateral displacement within the piping permitted by the " loose" support, in absolute terms, is not large (estimated at approximately 1/8").

Initial judgments vero that.this looseness may have had substantial adverse influence (with regard to stress) on the fiberglass piping vith a primary role in leading to the eventual piping failure.

Subsequent piping-analytical votk, however, has shown a piping displacement of this magnitude to be of relative little importance.

Therefote. although the presence of a loose anchor support during system operation is a probable contributor to the fiberglass piping failure, it is not considered to be a primary causal factor of the same, w

PY-CEl/01E.0380 L Attachment ?

Enclosute ?

Page 4 of 13 Based on the above diseumston. " loose" anthoi suppoit #1N71-H0013 is concluded to be a ptebable contributot to the piping fallutet however, it is not an initiator of the fallute and not a centributor of primary causal influence.

D.

NON UNIFORM 1.0ADING Or Tile EXTERIOR OF Tile FIllERGl. ASS P1P1NG This failuie mode vas considered, but was deteimined not to be a contributoi to the fallute.

If the bicak in the pife vote underground then exteinal loading isom fill vould be a consideration. The fallute was above ground vheio these vas no such loading.

E.

DEGRADATION OF Tile MATP.RI AL STRENGTil 0F Tile PIPE DUE TO AGl!

This factor would act as a recondaty contributor to other primaty factors when examining a failure.

Many factors contribute to the degradation of the material stiength of the fiberglass pipe over time. Two key factots include stress loading of the pipe under vetted conditions and ulttaviolet (UV) tadiation, it should be q

noted that this pipe was located on the Notth face of the building and was insulated. Thetefoie UV radiation would bn at a minimum for an exposed pipe and not a factor for an insulated pipe.

This vss considered not to be a palmary causal factot in the piping tallute as compared to other issues as discussed herein.

F.

DEGRADATION OF MATERI Al. STRENGTil DUE TO IIEAT TRACING Severe localized heating of the fiberglass could enuse degradation of the strength of the fibeiglass piping.

Ileat tracing is radially wrapped around the elbov to ensute it does not fleere.

Although the elbov exhibited some discoloration at the heat trace lines, the elbov did not tupture along those lines.

No pattein of s upture that Jould point to heat trace as the toot cause existed.

Conclusions As stated above, the probable causes vere evaluated individually and in combination.

It is generally agreed that no individual causal factot precipitated the failure.

The most probable and primary cause of the failure was strength weakening of the elbov vall caused by the presence of non-design axial grooves. Under operationn) loading, these grooves also acted as stiese risers and appeated to be the initial site fot elbov failure.

Some induced stressing of the fiberglass elbov vas also probably caused by " loose" anchor suppor t IN71-Il0013.

similarly, incorrect 0-ring installation may have caused some additional attessing of the elbow. The latter tvo factors, however, are not considered as primary causal factors of the piping failuie.

_Short Term Corrective _ Actions To prevent any tecurrence of a similar type f ailuie, the following steps vete taken.

pY-CEl/01E-0380 L Enclosute 2 Page 5 of 13 1.

The Auxiliary Condenser inlet line had additional fiberglass material added to the elbow to strengthen it and increase its pressure capacity.

Calculations vere petformed to determine the additional material necessasy to eliminate any possible mate:ial degradation concerns. This action was completed prior to plant startup.

2.

Careful attention to the co: rect assembly of the 0-ring to ensure the optimal gap between the 0 ring retainer and the flange faces was met. The discharge line vas also inspected and evaluated.

These actions were completed pitor to plant startup.

3.

Design Change Package (DCP) 91-0288 significantly upgraded the strength and resis>ance to loosening (under operational system loading,:,) for supports IN71-H0013 and IN71-H0021 (" anchors" on both inlet and discharge N71 piping.) This DCp was implemented prior to plant startup.

Long Term Corrective Actions In addition to the above steps, the Auxiliary Condenser discharge piping fiberglass elbov vill be evaluated to determine if seinforcement of this elbov vith additional fiberglass is necessary. This evaluation vill be performed in light of the contribution of age and strength degradation to the root cause of the failed suction elbov. the lover operating pressure of the discharge piping and the previously stated pipe support modifications. This evaluation vill completed by the end of RF03.

An evaluation to determine the need for heat tracing on the fiberglass elbovs vill also be completed by the end of RF03.

The 0-ring retainer and flange face spacing for the discharge elbov vas inspected during the plant shutdovn and did not h the.0625 inch gap required for optimum spacing. This line vill be reworked to cortect the spacing error by the end of RF03. An evaluation was performed to justify interim operation based on the lover operating pressure of the line, the support modification performed, and the results of inspections performed prior to plant startup.

All major portions of the fiberglass piping vill be visually inspected for flav indications during RF03.

II.

EQUIPMENT PROBLEMS AND ANOMALIES Various equipment problems were experienced folloving the December 22, 1991 circulating vater pipe rupture and subsequent plant shutdovn. A summary of the significant problems encountered and the associated corrective actions is provided belov, A.

Electrical Equipment

1. -

Bus L11 Failure to Transfer Upon plant shutdown, i.e., turbine trip, the plant auxiliary loads are transferred to plant startup power sources.

This is I.

accomplished automatically by:

(1) opening 13.8kV breaker l

l

PY-CC1/01E-0300 L Onclosute 2 Page 6 of 13 L1102 and closing breaker L1006 an<l (2) opening 13.0kV breaker L1202 and closing breaker L1009.

Both of these breaker automatic transfer schemes ate driven by the same relay logic.

The L1202 to L1009 transfer properly occutted, and the L1102 and L1006 transfer failed. Upon inspection of 13.8kV breaker L1006, maintenance found that its closing springs were dischanged. All spring charging switches, fuses, etc., vere found to be in proper position.

Maintenance determined that a subcomponent of the bicaker mechanism had broken.

Several additional problems occurred as a direct result of the failure of Bus til to transfer and vete tesolved when power vas restored to the bus. They areuns follovst a.

Control Rod Drive (CRD) Pump B tripped due to the momentary de-energisation of the " loss of oil pressute" relay.

b.

Switch $112 (345kV, Main Transformet disconnect switch) would not open due to loss of povet to the motor which operates the switch.

c.

Various containment isolations occurred due to the loss of Reacter Ptotection System (RPS) Bus-"B" and other lov voltage busest o

Reactor Vater Clean Up o

Reactor Vater Sample Lit.c o

Backup flydrogen Purge o

Balance of Plant d.

A Control Room Emergency Ventilation recirculation mode initiation occurred as a result of losing 120 VAC Panel K-1-H.

e.

Pover was lost to floor and equipteent drain sump pumps.

Short Term Corrective Actions

-The breaker subcomponent which failed (a control device relay) vas replaced and breaker L1006 vas successfully retested.

L,,ong Term Corrective Acti_or3 No long term corrective measures were required.

2.

Motor reed Pump (HTP) Breaker Failure to Close The MFP breaker logic was set in AUTO-START tesponse mode at the' time of the event. Vith the two Reactor Feed Pumps turbines tripped, the HFP vill feed water into the teactor vessel continuously or until a vessel Level 8 is teached.

After the Level 8 signal clears, the operator can reset the Level 8 trip signal and the NFP vill again auto start. This

PY-Crl/010-03P8 1.

o Pnge 7 of 13 trip /teset actton occusied 15 times ovet a tvo hnut pettod. On the sixteenth 'tip teset, the MrP did not automatically statt.

Short Tetm Cottective Actions An engineeting ieviev of the MrP motor's bteaket contiol logie did not tevtal any anomalies which explain the bienket's fallute to close on the sixteenth close actuation demand. This teview included examination of the bienker's anti-pump conttol logic.

In addition, the bt eaker vas t emoved f rom the cubicle and cycled satisfactotily using the bienket testing equipment.

The breaker vas disassembled and contacts vete inspected.

The breaker was seassembled and operated several times in the test position in the svit;hgeat. No problems vete found.

Doting the initial post event stattop of the MFP, the motot was monitoned Ior any unusual noise or vibration.

No abnotmalltles vere obsetved, long Term Corrective Actions A fallute of the MFP breaker occurred on Januaty 29, 1992.

A failure evaluation is cuttently in progtess and vill review any relation between the recent f allut e and the one which occut red on December 22, 1991.

3.

Startup Transformet Deluge Initiatiot3 This Fire Protection System featute functioned pet design when the rate of rise sensots detected a tapid temperature rise when the comparatively hot N71 vater (approximately 80 - 85 degrees Fahrenhelt) hit the much cooler transiosmet. The amount of water and location of water contact did not pose a problem as evidenced by the continuous operation at Stattup Transformet 100-PY-B.

Short Term Cottective Actions No cottectIve actions vere equi 41.

4.

Equipment Problems sesulting fro 4:.la t e t intrusion Electrical and communicati f. manholes Nos. 1, 2, 3, 4 and a.

7 vere flooded during thcA ent.

Security manholes Nos.

60, 66, and 67 vere also f %oded.

In n.anhole No. 2 a small amount of water was.bcerved leaking from conduits. This tesulted il vater in a Division Ill Unit 2 Motor Conttol Center (HCC). The MCC, from the partially completed Unit 2 plant, vas r.ot energized.

The only othet electrical equipment damage resulting from manhole l

flooding was isolated to manhole No. 3.

Vater from this manhole ran back into the south-east corner of the l

_ _. _ _ _ =

PY-CE!/01E-0388 L rnelosute 2 Page 8 of 13 Emergency Setvice Vater Pump flouse (ESVPil) to electrien1 1

l junction box JB1-2114. The vater then passed through a series of conduits into the HCC EP1A12, temperatute detector IP45-N08BA, and transmittet IP45 N090A. Most of the vater tan to the floort however,-a small amount of vater flowed into MCC compartment C causing a 120 VAC control transformer to short. Two additional instruments in the ESVPH, 1P45-N0100A and 1P4$-N0220A vete found to i

have vater in them which appeared to be unselated to the flooding event. No plausible pathvay for watet entry into i

these instruments vas detetmined.

b.

Ground alarms were expetlenced on operation of Service Vater valves OP41-r0420 and OP41-PO430 which vere also suspected to have resulted from flooding. There vas no l

vater found in the valve pit for these valves and the valves closed when required.

Short Term Corrective Actions The conduits entering junction box JB1-2114 in the ESVPil vere sealed to minimize watet entry.

Additionally a hole plug was removed from the bottom of JB1-2114 to allov vater to drain to the floor rather than following downstream conduits. The af fected instruments in the ESVril vere s epaired or repinced as necessary.

The motor operator for Service Vater valve OP41-F0430 vas found to be grounded and was subsequently replaced. Valve OP41-F0420 vas inspected and satisfactorily tested with no problem identified.

Long Term Corrective Actions i

Additional affected equipment vill be inspected as__necessary.

B.

Hechaatcal Equipment 1.

Scram Discharge Volume Failure to Drain l

The scram discharge volume (SDV) failed to drain following the l

manual scram insertion due to a failed stem coupling on the outboard drain valve 1011-F0181. The coupling joins the actuator to the valve stem.

A notification was made to the NRC at.2225 hours0.0258 days <br />0.618 hours <br />0.00368 weeks <br />8.466125e-4 months <br /> on December 22, 1991 to teport the SDV drain failure pursuant to the requirements of IE Bulletin 80-14. The failure vas similar to failures reported in General Electtic (GE) Huclear Services Inf ormation Let tar _(SIL)-422. The i

i consequences of the failed scram discharge drain valve stem connector was not signifleant. All control rods vere fully inserted with the scram signal.

1 l~

l l

l-

PY-CE1/01E-0388 L

+-

Page 9 of 13 Short Term Corrective Actions A replacement coupling was installed using the guidance l

provided in GE S11. 422.

Detailed inattuctions vote included in l

the associated Voik Order to ensute ptoper assembly during the reinstallation process.

. l i

Long Tetm Cortective Action As an additional enhancement to improve the teliability of this f

component. Engineering Design Change Request (EDCR) 91-0289 was initiated to evaluate potential design improvements to the stem coupling arrangement.

2.

Instrument Air Pressure Not Maintained During Event j

It was originally believed that a problem existed in the Instrument Air System due to an inability to maintain system

+

pressure above 86;psig with a scram inserted and the Safety Itelief Valves being cycled.

A detailed evaluation of the i

sequence of events, system pressure and overall system tesponse i

vas petformed. The analysis concluded that the system had functioned as designed during the event and the Unit 1 i

Instrument Air Compressor was able to supply all required air for important equipment-manipulations. The analysis revealed interrelations associated with operating modes of the compressors which vete not immediately understood.

j Short Term Corrective Actions The analysis of suspected Instrument Air System problems resolved previous concerns regarding overall system performance. No additional actions are tequired.

F C.

Structural The only significant structural' damage resulting from this event was i

confined to the pipe support. discussed previously and soll displacement in the area where the ruptured piping exited the ground.

Some of the soll and stone used around the yatd.' area structures was also displaced as a tesuit-of the flooding.

Two security perimeter detection zones in the vicinity of the pipe rupture were affected due to the vashout of'aggtegate.under the associated security fencing. Appropriate compensatory measures were i

taken. Additional arear affected include a concrete walkvay which was partially damaged and minor housekeeping problems-from displaced silt:and debris.

Short Term Corrective Actions The soll adjacent to the damaged N71 piping and support was replaced per direction of Engineering department personnel.

Aggregate which vanhed avay under-the perimeter security tence was also replaced.

&-e e

-i.wnw--

-.mev--.m-+--..e,-

,.w,,-y.w--m-cw,,,

y i%,e,-.,wwww--

-,,c

.,.m,4ye-,.ec,

,6s.-,.,m.,,,f.y.-ysme,.,:,_w,.-,

e-9.$-*-,+w+.--vm-r-y-er-.

m__

PY-CEI/01E-03BB L Page 10 of 13 Areas where housekeeping van degrada) as a result of the pipe rupture vere cleaned up prior to plant startup.

j I

Long Term Corrective Actions The remaining cortective measuies involve cosmetic repairs to the yard area. and repair of the damaged sidevalk. These activities vill be prioritized commensurate with ongoing plant activities.

III. RADIOLOGICAL CONSEQUENCES As a result of the December 22, 1991, event, slightly contaminated water and sludge vere deposited in the basement levels of the Intermediate Building, Radvaste Building, Control Complex, and Unit 2 Auxiliary Building. The contamination was sptead when floor drains in the buildings backed up during the event.

Power to building sump pumps was

. temporarily lost during the event, which contributed to the water level 1

in the buildings.

A portion of cor.tambated vater which ente $ed the Unit 2 Auxiliary l

Building was inadvettently discharged to the s!.te storm drain system t

through an unmonitored pathway. This resulted from a temporary hose l

connection which connected the Unit 2 Auxiliary Building sump to the Turbine Pover Complex sump and ultimately to the environment. The l

radiological consequences of the event are minimal as indicated by the table belov, which compares the conservative exposure-estimates to'the i

limits contained.in the Perry Technical Specifications.

Event Tech Spec

% of Toch I

(mrem) limit (mrem)

Spec limit Total Body Dose 0.000017 3.0 0.000.W %

i i

i Organ Dose

-0.000031 10.0 0.00031%-

Short Term Corrective Actions l

Building areas which became contaminated as a result of the pipe rupture event vere surveyed and subsequently cleaned up prior to plant startup on January 3, 1992.

c The temporary hose connection from the Unit 2 Auxiliary Building sump has been removed, eliminating any potential pathway to the environment.

The site storm drain system vas cleaned during the week of January 21, 1992.

1:

2.

pY-CEI/01E-0308 L EncloSute 2 i

Page 11 of 13 IV.

SAFETY ANALYSIS i

None of the equipment problems or anomalies described impacted equipment required to safely shutdown the plants therefore, this analysis vill focus mainly on the flooding aspects.

The water discharged by the 36" diameter N71 line brc K located north of the Heater Bay at approximately 620' elevation, generally flooded the yard area in the immediate vicinity of the break. Approximately one to two feet of water could have existed for a short duration at the vest boundary of the flooded area.

A.

Normal Design Flow Path Normally, most of the water from the break would be dissipated by surface run-off towards lov lying areas away from the plant.

(For this break, most of the vater vould run-off in the north and north-vest direction and some in the north-east direction).

Some of the water vould seep through the class B/C fill (at a very slow rate, as class B/C fill is nearly impervious to vater) around the i

building and reach the Underdrain system. The Underdrain system consists of a l'-0" thick porous concrete mat under the building foundations and a 12" diameter porous pipe routed around the perimeter of the plant. The porous pipe carries the collected water to nine (9) individual pumps located in manholes spaced around the

[

nuclear island. The water collected in the manholes vould be pumped to the gravity discharge piping (36" to 48" diameter steel pipe, at El. 588' [high pointI to El. 579' [lov point]).

In the unlikely event of the failure of all nine (9) pumps, the water level in the i

manholes vould rise to El. 588' and be drained to the ESVPH via the gravity discharge piping. The underdrain system is designed for a postulated break in the circulating vater system (12'-0" diameter fiberglass pipe) and is sized to handle the flow from such a break.

The break in the 36" diameter pipe which occurred above grade was determined to be bounded by the break postulated for the design j

basis of the Underdrain system.

B.

Estimate of Actual Flov Path A valk-down conducted on December 22, 1991, revealed that the cover for manhole No. 20, immediately to the vest of the N7) pipe break, had been left open. This provided a' direct and a much more rapid path for some of the flood water to the Underdrain system. This, along with the water that seeped through the ground to the Underdrain system, is considered to be the main flow path to the

.Underdrain system. The pumping capacity of the Underdrain pumps was exceeded for some time (this explains the high water level alarm 1

received in the Control Room after the break; the alarm is set at El. 568.5').

The pumped discharge portion of the Underdrain system was probably L

subjected to a more rapid flov from the break (due to the open l-manhole) than anticipsted by design.

Ilovever, this did not create a safety concern since the pumped discharge system is not the primary q w are--a n

~

-v,

,.-,v,-,,

,ne - pr, %

,a v.

-, - +.,

,-w.:r-n, m-,,,-

w

,-w-,,w-.,-n, r

r.,--,--

C '

pY-CE1/ ole-0388 h l

Enclosute 2 l

Page 12 of 13 i

system for keeping the water Icvel belov Fl. 590'.

The Gravity Discharge system, designed to perfotm this function, has been shown to be adequate to handle a break in the N71 system which envelopes the current break (discussed above).

Furthet, the ground vater i

level was lovered to El. 508.5' soon after the break as confirmed by a valkdovn on December 24, 1991, and pierometer vater level readings taken on December 26, 1991. This confitms that the Undetdrain system performed its function as designed.

Additionally, due to open manhole No. 20, thete is a possibility that the capacity of the gravity discharge portion of the Underdrain system was temporarily exceeded. This vould result in the cater level rising above El. 590' in the manhole.

However, this water vould be discharged to the lake via the Gravity Discharge system before it could illi the porous concrete and the Class A fill to El.

590f. Thus, the water level could not have exceeded El. 590' l

(design basis of the Underdrain system).

The path of ingress of water into the plant buildings has been determined to be as follows:

l 1.

Belov E1. 590', water most probably entered the safety-related

{

buildings through holes / tears in vaterstops/ vater proofing membranes at the building rattle spaces and pierometer tubes.

The amount of in leakage vas also somewhat aggiavated for this occurrence by the temporary loss of power to sump pumps within the buildings.

2.

Above El. 590', all the water came into the plant when the electrical manholes i111ed and vater rai bsck through underground duct banks into the plant, into the Service Vater pump house and into the ESV pump house. The amount of vater intrusion above El. 590' vas insignificant and as such had no safety consequences. The cables in the electrical manholes were specified by design to operate for forty years submerged in Vater.

The only safety-related equipment affected was in

{

the ESV pump house where vater entered into the building at the south-east zone Junction Box JB1-2114. Vater then passed through a series of conduits and boxes and ended up in Motor Control Center (HCC) EF1A12 causing the failure of a space heater transformer. Although this had no safety consequences, it is of concern due to the fact that water flowed into safety-related switchgear. The inlet point for this vater han been sealed to prevent any future occurtence.

The extent of in-leakage to plant structures can be attributed to a very rapid entry of-flood vater into the open manhole, causing the Underdrain system.to fill up rapidly.

It should also be noted'that, for the most part, the floor drains vere able to dissipate the water adequately. Thus the items designed to keep the buildings free of water performed in an acceptable manner.

The actual flood path for this break was not the path anticipated by design, largely due to the open manhole hovever, the systems designed to handle flooding performed adequately as demonstrated by the fact that no essential r,,,

--n----,,-,--,-.--,--,,.,,.,_,,,,_.,,,.,...-,,n,..

_,,,.,,... +,,,,

c,,

---r n.-

pY-cr1/010-0388 h Enclosute 2 Page 13 of 13 safety-related equipment was lost as a result of the floneling.

Therefore, this event is not considered to be safety significant.

,Short Team Corrective Actions Flooding damage from this event was primarily attributed to manhole No. 20 being left open. Administrative procedure pap-0204, "flousekeeping/ Cleanliness control," van revised to requit e manhole covers to be in place except when required to perfotm maintenance and inspections. The piocedure change was made effective on January 3, 1992.

The conduits for the electrical junction box in the ESVpil have been scaled, as described previously in Section 11.A.4 of this enclosure.

~

It was also suspected that some pierometer tube caps may not have been in place at the time the event occurred, allowing vater to enter buildings through the pierometer tubes.

A valkdovn was performed prior to plant startup to ensure pierometer tube caps vere in place. Additionally, applicable procedures vere revieved to ensure that sufficient conttols existed to maintain these caps in place after removal for periodic inspections.

Lon,g_ Term Corrective Actions No additional long term measures are tequired.

I

_