ML20215B280

Interim Rept 1, Main Feedwater Pump Suction Pipe Rupture Incident

ML20215B280
Person / Time
Site:	Surry
Issue date:	12/29/1986
From:	VIRGINIA POWER (VIRGINIA ELECTRIC & POWER CO.)
To:
Shared Package
ML20215B065	List: ML20215B064 ML20215B135 ML20215B179 ML20215B280 ML20215B358 ML20215B402 ML20215B435 ML20215B540 ML20215B592 ML20215B624
References
FOIA-87-20 NUDOCS 8706170279
Download: ML20215B280 (24)

Text

e

.. INDEPENDENT TECHNICAL INVESTIGATION CODMITTEE 1

MAIN FEEDWATER PtMP SUCTION PIPE RUPTURE INCIDENT INTERIN REPORT #1 SURRY POWER STATION UNIT 2 ISSUED: DECDBER 29, 1996 I

i l

8706170279 870611

$0ELLNG87-20 PDR

ACKNOIA.EDGDtDITS The independent technical investigation committee was composed of R. W. .Calder, }

E. S. Grecheck, J. M. McAvoy, J. N. Maciejewski, J. W. Ogren. T. .B. Sowers, '

L..L. Spain, and W. A. Thomas, Jr. The comittee would like to thank the many people who helped with the investigation. In particular, we would like to j thank R. H. Blount, I. L. Breedlove, J. M. Mutt, J. E. Lewis, E. Homer, G. W. Heitus, M. B. Shelton, R. E. Tolbert, E. W. Throckmorton and E. Watts for  !

their technical assistance and T. F. Kelsey for his long hours and high quality i photographs. ,

f 1

o a

5 l

6 TA8LE OF CONTENTS SECTION PAGE f I. EXECUTIVE

SUMMARY

. . . . . . . . . . . . . . . . . . . . I

!!. INTRODUCTION . . . . . . . . . . . . . . . . . ... . . . 2 III. OVERVIEW OF EVENT, . . . . . . . . . . . . . . . . . . . 3-IV. SYSTEH ANALYSIS. . . . . . . . . . . . . . . . . . . . . 4 i A. Purpose B. System Conditions Prior to the Event C. Possible System Conditions Thru the Event D. Other Possible Conditions Which Could Have Contributed to tha Event Conclusions E.

V. PIPE FAILURE ANALYSIS. . . . . . . . . . . . . . . . . . 8  ;

A. Background B. Investigation and Conclusions C. Probable Failure Sequence VI. RECOMMENDATIONS. . . . . . . . . . . . . . . . . . . . . 11 VII. ASSOCIATED INVESTIGATIONS AND OBSERVA7 IONS . . . . . . . 13  ;

A. Purpose B. Maintenance of "A" MFP Suction Piping C. "A" MFP Suction Pressure Indication D. Work Ongoing in Turbine Building E. High Pressure Heater Drain Pump High Pressure Trip Circuit F. "A" MFP Discharge Check Valve G. "C" MSTV H. Actuation and Discharge of Halon and CO2 ystems S

I. Associated Reconenendations l' VIII. LIST OF FIGURES. . . . . . . . . . . . . . . . . . . . . 19 l IX. LIST-OF ATTACHMENTS. . . . . . . . . . . . . . . . . . . 20 l X. LIST OF REFERENCES . . . . . . . . . . . . . . . . . . . 21 1

i l

k e.

..)^

"- I. EXECUTIVE SISNUtY Based on preliminary system and pipe failure analysis, the Surry Unit 2 "A" feedwater pump suction pipe rupture appears to have occurred during a nonnal secondary side pressure transient following a reactor trip. l The pipe ruptured because its wall was thinned to as low as 0.048 inches by a bulk single liquid phase erosion / corrosion phenomenon.

At installation, the pipe had a nominal wall thickness of 0.500 inches.

The pipe was correctly specified per the applicable code and installed-per the installation specification. Ultrasonic testing has revealed similar but less severe thinning in similar sections of pipe on Surry Unit 1.

This bulk single. liquid phase erosion / corrosion phenomenon appears to be  !

confined to situations with high purity (low oxygen) water in combination with pipe configurations (elbow, Tee or orifice) that result in high fluid flow rates and severe turbulence. Pipe mater system temperaturearealsothoughttobecontributingfactors.{}and The tenn erosion / corrosion is slightly misleading and the ' phenomenon is perhaps better described as flow assisted corrosion.to distinguish it' from pure erosion or cavitation image.

Low oxygen content of the feedwater is believed to retard the build-up of magnetite (a black oxide) that would otherwise protect the pipe. . In the areas of high turbulence, the magnetite that is forred on the inner.

surface of the pipe is removed by dissolution. More stringent water chemistry controls were implemented following the steam generator replacement outages to protect steam generators and turb:.es from corrosion. However, these chemistry controls did not nece!.. s tate or result in a reduction of feedwater oxygen concentrations.

The failurr location was in a 90 degree elbow following an 18 inch T-joint off of a 24 inch header. An independent failure analysis laborato y has been contracted to perform additional testing of the  ;

f failed pipe. Ultrasonic inspection of locations with other pipe geou tries shows less severe wall thinning. There is a program underway with the help of a consultant to' identify for North Anna and Surry inspection locations for potential bulk single phase erosion / corrosion.

Inspected locations that do not meet acceptable minimum wall thickness will be replaced. The following replacement options are under .

evaluation and the selection will be documented in a Technical Report.

o direct replacement with increased inspection l o thicker wall pipe i o change pipe material i o change piping geometry .

i

11. INTRODUCTION An independent technical investigation comittee was appointed by Dr. J. T. Rhodes, Senior Vice President - Power Operations to investigate the December 9, 1986 main feedwater pump suction pipe (condensate) rupture on Surry Nuclear Power Station Unit 2. The comittee was tasked to determine the cause of the pipe rupture and examine potential contributing factors. The committee worked in parallel with other incident response groups as shown in Figure 1.

This interim report provides an oven few of the event, an evaluation of potential contributing factors (eg., design, . installation, testing, operation and maintenanc of associated systems) and the results- from preliminary nondestructive examinatien of the pipe. This interim report also provides the preliminary conchts. ions and recomendations of the comittee.

Additional reports will be issued by the comittee with data from destructive examination of the ruptured pipe.

l l

2

f e -

III. 0VERVIEW 0F THE EVENT At approximately 2:20 p.m. on December 9,19%' Surry Unit 2 tripped due to low-low level in the "C" steam generator. This trip was initiated by an inadvertent closure of the "C" main steam trip valve (MSTV). The unit had been at 1005 power since 1:00 p.m. on December 8,1986. The reactor protection systems and other safety related systems . functioned as designed with two minor exceptions and the norinal plant transient resulting from this type of trip ensued.

Approximately one-half minute after the trip, a carbon . steel elbow in  !

the 18" suction pipe to the "A" main feed pump failed catastrophically. l The failure was in an eroded non-weld region in the inner radius of the '

90 degree elbow approximately one foot downstream of a -tee from the 24 i inch header. The free - end of the seiered suction line- moved ~

approximately)

(See Figure 4 6.5 feet before becoming wedged between adjacent piping.

The water / steam flashing from both sections of the severed pipe erigulfed equipment and personnel in the area. Several workers in the area were seriously. burned by the discharge. The Halon and CO fire suppression systems inadvertantly actuated due to electrical shor,ts caused by water in their control cabinets. -Discharge of CO 2 froze repeaters for plant communications systems. Additionally, a door key-card reader shorted out causing security doors to malfunction and the fire suppression sprinkler system activated due to the heat released by the break.

The primary system was virtually unaffected by the pipe break and the secondary systems were secured within approximately one minute. Two reactor coolant pumps (RCPs) were runging with the reactor coolant system (RCS) temperature maintained at 520 F by relieving steam ' to the atmosphere through the "C" steam generator power operated relief valve (PORV). A second RCp was secured and a cool down commenced using atmospheric vents and auxiliary feedwater. At 7:30 a.m. the next day, the plant was in cold shutdown on the residual heat removal system. No detectable radioactivity was released because of this event.

Operator response to the event was appropriate and the overall handling of the emergency was professional. A more detailed sequence of events is provided in Attachment A. " Summary of Significant Operating Event."

A sketch of the secondary system is provided in Figure 2 and a general arrangement of the turbine building is provided in Figure 3.

i 6

e . . . .

1 l

IV. SYSTEM ANALYSIS A. Purpose 1

The purpose of this section of the report is to provide an analysis of the pertinent portions of the Unit. 2 Feedwater and Condensate Systems innediately prior to and during the "A" feed pump suction pipe rupture event. Possible system conditions which could have potentially contributed ta the event will be- explored and j conclusions presented. '

8. System Conditions Prior to the Event i'

Operating Parameters -

Unit 2 was l stable at 100% power with slight fluctuations in "C" stearn and feed flow indications. Feedwater pumg discharge header conditions were as follows: 1040 psig, 374 F, 24,884 gpm. The "A" feedwater pump suction piping conditions were calculated as follows (asguming equal distribution): 367psig,12.442gpm(17.6 f t/sec), 374 F-(62,F subcooled). At these conditions, the. bulk fluid is in a single liquid. phase. For references on the above operating conditions, refer to Attachment B.

Both feedwater pumps, 2-FW-P-1A & 16, were running, high pressure feedwater heater drain pump, 2-SD-P-18, was running; condensate pumps, 2-CN-P-1A and 1B were running. System valve lineups were for 4 nonnal operation as follows: the nonnal condensate and feedwater flow paths were open to the steam generators; the feedwater pump discharge recirculation valves FCV-FW-250A & B were closed. Refer to Figure 2 Feedwater Pump Suction Pipe Wall Thickness: '

An evaluation of the class 301, 18" feedwater pump suction piping revealed that it was originally installed in accordance with the Surry Piping Specification. NUS-20, which specifies A-106 Gr.B extra strong pipe (.0. 500" nominal wall) with A-234-WPB extra strong fittings (0.500" r.ominal wall). The nominal wall thickness of the originally installed piping 'is well above the piping design code (USASB31.1.0,1967 ed.) minimum wall thickness of 0.360 inches for 18" pjpe at feedwater pump suction-line design conditions (610 psig at 380 F). l Teedwater pump suction pipe wall measurements subsequent to the l event revealed wall thicknesses in the general failure area (i.e.,

the area of corrosion / erosion) of 0.300 inches to 0.140 inches with local spots down to about 0.048 inches. These eroded wall thicknesses are well below the piping design code minimum wall thickness of 0.360 inches The wall thickness at which the 18" feedwater pump suction piping would " burst" was estimated by using the piping material's ultimate strength (60,000 psi) in the code minimum wall equation in place of the material's allowable stress (15,000 psi). This estimated pipe wall thickness for "burstirP was calculated to be 0.090 inches for an internal pipe pressure of 600 psig. A pressure of 600 psig is expected in the feedwater pump suction piping during a nomal secondary plant shutdown, i

Refer to Eection C below for possible system conditions through the event. Refer to Attachment C for details of these wall thickness calculations.

C. Possible System Conditions Thru the Event i Nomal Transient:

The following trip transient would be expected upon closure of the main feedwater regulating valves. FCV-2478. -2488, and -2498 due to a reactor trip and low Tave, the feedwater pump discharge pressure l would increase to near shut-off head, about 1,610 psig (1450 psig 1 based on the simulator). The feedwater pump recirculation valves, FCV-FW-250A & B, would open and the discharge pressure would slightly decrease.

Since the nomal flow path is closed, the feedwater pump suction pressure would increase to about 600 psig because the condensate pumps and the high pressure heater drain pump would also reach shutoff head. The condensate pump shutoff head is about 610 psig.

The pressure would be limited to about 600 psig because of protective pressure switches, CN-PS-217A & B, which would trip the high pressure heater drain pump at 600 psig.

For calculations giving estimated system pressures during a nomal system transient, refer to Attachment B.

D. Other Possible System Conditions Which Could Have contributed to the Event Feedwater Pump 2-FW-P-1A Trip:

To date, there is not sufficient information to conclude precisely when 2-FW-P-1A tripped. If the feedwater pump tripped prior to the i pipe burst, it could have contributed to the event by allowing the pump's discharge check valve, 2-FW-127, to slam closed. This could have resulted in a hydraulic shock to the system. However, continued flow during the centrifugal feedwater pump coast down would have slowed the discharge check valve closure if the pump did trip prior to the pipe burst.

Closure Time of the Main Feedwater Regulating Valves.

FCV-2478, -2488, & -2498:

Extreme rapid closure of these valves could cause a hydraulic si.eck to the system. There is no evidence of this occurring.

Recirculation Valves, FCV-FW-250A & 8:

To date, there is insufficient information to conclude specifically how these valves operated during the event. If these valves did not open when required, the pressure transient during the event would be slightly greater because the feedwater pumps would be dead-headed.

Feedwater Pump Discharge Check Valve, 2-FW-127:

Subsequent to the event, this check valve was found damaged. If feedwater pump 2-FW-P-1A tripped prior to the rupture while feedwater pump 2-FW-P-1B was running, backflow through 2-FW-P-1A was possible. This backflow could pressurize the suction piping of l

2-FW-P-1A up to the discharge pressure of 2-FW-P-18. The as-found j condition of check valve 2-FW-127 indicates that the valve would 1 open and close, but would not seal fully closed.

{

High Pressure Heater Drain Pump Trip Pressure Switches, CN-PS-217A & B:

1 If these switches failed to trip the high pressure heater drain pump l at 600 psig as designed, the feedwater pump suction piping pressure could have exceeded 600 psig. Since the cable connecting the pressure switches to the breakers appears to have been severed during the event, it can not be esttblished that this instrumentation was operational during the event. The two pressure switches mre tested and found to be within tolerance and the breaker trip tested satisfactorily from- the turbine building i junction box. Thus, it is likely the circuit was functional at the j time of the event.  !

E. Conclusions

1. Calculations based on system pressures produced by the normal .

feedwater and condensate system transient associated with a '

main steam trip valve closure predict gross failure for sections of piping with the measured pipe wall thicknesses of the failed section.

2. Individually, or in combination, the following conditions could have led to increased pressure or hydraulic shocks in j the feedwater and condensate system greater than the normal transient conditions. However, these potential conditions are judged not to be contributors to the event.

)

o trip of feedwater pump 2-FW-P-1A and quick closure of discharge check valve 2-FW-127, o quick closure of feedwater regulating valves, FCV-2478,

-2488, & -2498 1

- .. .. . -. . .l

i o failure of recirculation valves FCV-FW-250A & B to' open

' properly, o . backflow through feedwater pump discharge check valve, 2-FW-127, and o failure of the high pressure heater drain pump trip circuitry.

i f

l l

J 7

r h PIPE FAILURE ANALYSIS A. Background A field metallurgical investigation was made of the Surry Unit 2 "A" main feed pump suction line failure. The failure occurred at an 18 inch diameter schedule XS WPB elbow as shown in Figures 4, 5, and 6. '

The elbow material is carbon steel, ASTM A-234 Grade . B, which is equivalent -to ASTM A-106, Grade B seamless pipe. The nominal wall thickness was 0.500 inches at installation and the elbow had seen service since Unit startup in 1973.

The field metallurgical investigation consisted of the following:

o Visual inspection of the system failure location.

o Removal of the fractured elbow from the suction line, i o Visual 5X magnification evaluation and photography of fracture surfaces, and elbow surface conditions.

o Ultrasonic wall thickness measurements, on a 2 inch grid pattern, of the failed elbow.

o Metallurgical replicas taken on the elbow at wal surface locations, o Mechanical measurements of elbow thickness.

B. Investigation Visual Evaluation:

The visual evaluation of the elbow inside surface revealed a thin wall and dimpled surface appearance (Figure 7). This condition has been noted previously, in bulk single phase systems, 'only in the Westinghouse steam generator J-nozzles. The J-non e surface condition was detennined to be the result of a bulk single phase system erosion / corrosion mechanism. This is a mechanism of electrochemical corrosion in rapidly flowing aqueous solutions.

Wall -loss by erosion / corrosion occurs by gouging-out-patterns on metal surfaces under the simultgous action of a flowing medium and an electrochemical dissolution.

Both the Surry J-nozzles and Surry feedwater pump suction line demonstrated a similar design geometry consisting of a header or large diameter pipe with a right angle discharge pipe extension and a 90 or 180 degree elbow or turn. This configuration is shown in Figure 8 for both components. The turbulent flow created by this geometry, and a possible vapor phase separation locally in the elbow and tee extension, coupled with the low oxygen content (average 4 ppb) feedwater, is concluded to have resulted the erosion / corrosion which thinned the elbow pipe wall. g The 1

1

K '

erosion / corrosion mechanism in the J-nozzles is known to result initially in a very low metal'1oss rate followed by a progressively higher nonlinear loss rate.

Ultrasonic and Mechanical Wall Thickness Meesurements:

Ultrasonic wall thickness measurements and direct mechanical measurements taken on the elbow indicate a gradually sloping wall thickness loss over much of the surface. At several. locations, usually near welds 3 localized cavity has bean created in the elbow surface. The remaining wall thickness at these localized areas has teen measured as low as 0.048 inches while adjacent locations may be 0.090 inches to 0.140 inches in thickness.

Metallurgical Structure Replicas:

Field metallurgical replicas taken on the surface of the failed elbow indicated a carbon steel pearlite / ferrite microstructure typical of ASTM A-106 Grade B chemistry pipe. No distortion of the structure or signs of strain where observed in any replica.

Fracture Surface Examination:

Visual examination of the fracture surface at SX magnification revealed an apparent ductile tearing m de, plane stress, slant fracture over most of the surface. Tears which may represent highly localized tensile overload areas are evident at two thin wall cavity locations shown in Figures 9, 10, and 11. Numerous defects of small size were observed along the fracture surface. Generally these defects represent laps, laminations, and inclusions less than 0.5 inches in length. At one thin wall location, near an apparent overload tear, an inclusiors or lamination was observed which may represent the final slant frecture (unstable ductile tear) initiation location. The slant fracture surface condition represents most of the failure surface.

C. Probable Failure Sequence Based upon the observations made in Section B above, a probable sequence of stress and failure events leading to the final pipe l rupture are as follows. )

i o The system is operating normally. System pressure is {

approximately g67 psig, and system temperature is i approximately 374 F. Wall thickness of the elbow is as low as i 0.048 inches in highly localized areas and 0.100 inches over a i someu /t more general area. Because of the high system temp cature, the toughness of the elbow material is on the i upper ductility shelf. Local membrane stresses are near yield I at low wall thickness locations.

o The system undergoes an upward pressure transient which results in a localized tensile overload failure (cavity

.g.

m

T,-

9 1

blowout) in a thin wall erosion / corrosion cavity. The. tensile overload tear arrests, and does not develop into the unstable I crack tearing mode. Water flashing into steam from the l localized tear is heard by station personnel, who also verify {

that the sound heard was not.the lifting of a safety valve.

].

o The steam flash continued for a few seconds as pressure continued to increase 'in the elbow. In the area of the small tensile rupture, or at a second thin wall area, the pipe wall experienced an increasing stress intensity approaching the 1 upper shelf tearing resistance of the carbon steel elbow material. An unstable ductile tear developed, possibly at a defect site. The pipe ruptured, ejecting a fragment from the elbow upward from the fracture surface.

VI. REC 0f04ENDATIONS The following recomendations are made based on this investigation:

1. Continued Evaluation of the Failure:

It is recommended that the failed elbow be given to an outside failure analysis laboratory for development and implementation of a failure analysis plan. The plan would include but not be  !

limited to:

o tensile tests (of elbow material) o bend tests o hardness tests o micro-hardness tests o operating temperature Charpy V-notch tests o chemical tests o optical metallography o scanning electron microscopy of fracture surfaces ad defect sites o preparat..an of a final report i

2. EPRI Analysis of Ductile Tearing Resistance:

It is recommended that an adequate sample of the elbow materir.1 be provided to EPRI for establishment of the operatfng temperature ductile tearing resistance of this material. This can be accomplished by the generation of J-resistance curves.

3. Establishment of a Single Phase Flow System Augment Inspection Program:

a. It is recommended that piping systems and design-specifications at North Anna and Surry be reviewed to establish a list of inspection locations candidates among the bulk single phase flow systems based upon:

1. Pipe material '

2. Operating temperature

3. As-fabricated design detail

4. Fluid velocity

5. Industry Data

b. A consultant, knowledgeable in this bulk single liquid phase erosion / corrosion phenomenon, should be utilized to identify potential areas.

c. Piping systems, or individual details, which are identified as potential problem locations should be placed in an augmented inspection program similar to that established for secondary system two phase flow regions.

4 Replacement of Piping:

) a

, j l

a. piping damaged by the event must be replaced. I

b. Any piping found by ultrasonic inspection' to be below a conservative minimum wall thickness must be replaced.  !

c. A Technical Report should be prepared to establish the methodology for calculation of an acceptable minimum wall thickness and to evaluate design options for pipe replacement. 3 ~

\

l l

VII. ASSOCIATED INVESTIGATIONS AND OBSERVATIONS A. Purpose The purpose of this section of the report is to record investigations and observations associated with the comittee's

-evaluation of the event, and to provide resultant recomendations and conclusions.

6. Investigation o't Maintenance History on "A" Feed Pug suction Piping The WPTS database was searched for evidence of previous related problems with the feed pump suction piping. A relatively' large number -(15-20) of work orders were identified in the database as referring to the "A". feed pump suction. Upon closer investigation, however, these were determined to apply to other piping, such as the feed pump recirculation line, or piping related to the feedwater heaters. No work orders related to previous leakage on the piping involved in the event were located.

A history of leakage on the feedwater pump recirculation lines was noted. This had previously been identified by Station personnel, i

and a Type 2 conceptual design study (NP-5395, Main Feedwater Pump l l: Recirculation Line Replacement) is currently in progress.

The excessive number of work orders identified in the WPTS database as "A" feed pump suction" appears to have resulted from the use of the general designation of mark number 02-FW-PP-NA as the suction line. Thus, unrelated work was identified with the suction piping.

C. Investigation of the High "A" Feed Pump Suction Pressure 1 Indication in the Controi Room '

On the shift following the event, the "A" feed pump suction pressure indicator PI-CN-250A, was observed in the control room to be reading offscale high (1000 psig). An investigation was conducted to detennine the cause for this indication.

On Wednesday, December 11, instrument technicians removed PI-CN-250A from the instrument rack and carried it to the instrument shop for examination. The transmitter is a Fisher-Porter Model 6698009002, I and is located on an instrument rack imediately adjacent to the "A" feed pump discharge. The transmitter was last calibrated on May 7, 1986.

Upon removing the cover from the transmitter in the shop, water was observed to flow out. A water line was noted on the terminal block in the unit, suggesting a water depth of approximately 2 inches.

Corrosion of the positive terminals was also observed. The negative terminals were observed to be Llack, but free of corrosion. Minor corrosion was also observed on portions of the printed circuit board. Photographs were taken to document this condition. No mechanical damage or corrosion to any non-electrical component was noted.  ;

s s

On December 12, a transmitter calibration was attempted. When pressure applied to the transmitter was varied from 0 to 1000 psig, the transmitter output remained at approximately 8 volts, rather than varying from 1 to 5 volts, as expected. Various troubleshooting techniques were used with no success, until- the circuit board and terminal screws were thoroughly cleaned. 4 Following this cleaning, the transmitter, with all- its original l internal components, performed properly and was calibrated 1 satisfactorily.

Based on reports of satisfactory transmitter operation prior to the ,

event, the water found in the transmitter is believed to have entered through a conduit .during the event. The water and resulting

)

galvanic action on the terminals caused the transmitter failure.

Consequently, the high offscale reading in the Control Room was most likely a result of the event, and not an indication of actual feed I pump high suction pressure prior to or during the event.

D. Work Ongoing in the Turbine Building Basement at the Time of the Event Among the potential contributing factors investigated for the pipe I rupture were the activities of the personnel working in the innediate vicinity of the feed pumps at the time of the event.

Other than personnel passing through the area , no Virginia Power employees were working in the area. Crews from Daniel Construction Company and Insulation Specialties, Inc., were assigned to various tasks in the innediate vicinity.

Based on an interview with D. E. Schappel, E&C Construction Superintendent, the following Daniels' activities were in progress:

o a scaffold crew was erecting scaffold on the south side of the bearing cooling pumps, o an instrument crew was running tubing for Design Change 85-07, On-l.ine Chemistry Monitoring, and i o a team of pipe fitters and welders were prefabbing and hanging pipe for the new service air compressors (DesignChange 86-03).

This work was running pipe or tubing only. No tie-ins to any system '

were being made.

Based on an interview with R. E. Mudd, Insulation Specialties had two workers involved with installation of metal insulation covers on the "A" feed pump discharge piping. l The nature of the work in the area would not be expected to contribute to the cause of the event.

~ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _

Investigation of High Pressure Heater Drain Pump E.

High Pressure Trip Gfrcuit As part of Design Change 84-62, a high pressure trip was added to both HP heater drain pump control circuits, to protect against possible overpressurization of the feed pump suction piping. Two pressure switches CN-PS-217A & B were installed on the feed suction header. Actuation of both switches, on a pressure of greater than 600 psig, will cause a trip of the HP heater drain pumps. The switches are Static-0-Ring part number 9NN-Y5-M4-C2A-TT.

On December 10, 1986, the pressure switch setpoints were checked by Station instrument personnel. Both actuated within tolerance.

A direct test of breaker trip could not be performed from the pressure switches since the cable connecting the switches to the breaker cubicles in the normal switchgear room was damaged during the event. - However, on December 14, 1986, a test of the breaker trip circuit was conducted from the junction box in the turbine building. The circuit tested satisfactorily. l The results of these investigations indicated the high pressure trip circuit was probably functional at the time of the event. Since the runr.ing HP heater drain pump did not auto trip during the event, it may be concluded that the feed pump suction pressure was not sustained above 600 psig. .

i F. Observed Conditions of Feedwater Pump Discharge ,

Check Valve Z-FW 127 '

On December 12, 1986, the following conditions of check valve 2-FW-127 were observed:

1. One of the two disk pivot pins was missing.

2. One of two valve seat lock plates was displaced such that it l was not functional.

l

3. The disk was cocked in its closed position such that it did not seal properly.

4. The seat ring assembly was cocked out of place leaving a path j for backflow thru the valve assembly.

l l S. A gouge, about 3 square inches in area, was in the valve body.

G. "C" Main Steam Trip Valve TV-MS-201C L

l The plant transient was initiated when the "C" main steam trip valve l inadvertently closed. The minimum instrument air pressure was approximately 78 psig. Initial investi J pressure, which keeps this valve open,did gations indicate to not decrease thata the air point 1 1

1

_ _L

4 that would have normally resulted in valve closure. (See Reference 4 for results of additional inspection and testing of the "C" MSTV.)

H. Investigation of Actuation and Discharge of Halon and CD Systees Ion Actuation:

wpdu/

ie system was r 9:1 f: n to evaluate the reason for system

( actuation. In addition, interviews were held with Operations and Loss Prevention personnel.

The Halon cylinders for Unit 2 are located approximately 50 feet i from the line break on Unit 2. The Control Panels for both units -i n

(1-FPH-CPI) w,4u4_ r as. .are

~ e 4mounted

,.. . n. on the ? e e' ' . th Unit 1 t#N 5 *6 i y ^ J N M~i'WK KM taaju~;f%. L.h.a .

$hortly after the feedwater pump suction line break, zones o bothL ithe systems were inadvertently actuated because water entered the j panel through an open condulet and shorted e*:t tne Units 1 and 2 push button circuits in the panel. Halon was then discharged to the .

Emergency Switchgear Room (ESR) on both Units. Fire dampers 4, 8, t 15, 21 anc 22 were inspected and found to be in the proper closed position. T uu M w A - d r ,J c _ ! }' N eeiv e a. 3% Ib e tca e sa.t k o b 4/m LJA f I,he affected control panel modules are:

1. 2N-30 Module Unit 1 Push Button l Unit 2 Pressure Switch ,

2, 2N-30 Module Unit 2 Push Button l Unit 1 Pressure Switch

3. T!-30 Module Unit 2 Time Limit CO 2 Actuation:

Units 1 and 2 Cable Tray Room CO systems inadvertently actuated shortly after the feedwater pump s,uction line break due to the entry of water into the control panels. The CO2 panels are located in the Unit 2 side of the 9 line wall on the 45 foot elevation as shown in Figure 3. The conduf ts leave the top of the panels and run to the edge of a cable tray directly above the cabinets. The sprinkler system and/or line break discharged water.into the open conduits.

The water then entered both panels and caused the circuits for both units to actuate the system. The Engineering report in Attachment D details the CO actuation. The bulk storage tank, located between Unit 1 and Unit 3 transfonners, was completely emptied. Unit 1 Cable Tray Room contained an excessive amount of CO after the incident due to shorting of the timer in the control pakel. l"fD'e'cu d..

dampers (Campers 27A, 27, 28 and 29) were inspected and found in the proper closed position.

l l

l

i Effects of Halon and CO 2 Actuation:

The actuation of both the Halon and COp impacted plant operations in several ways. Imediate and apparent effects were:

1. Undesirable in leakage of CO, in the Main Control Room (MCR)

& Main Control Room Annex (MCRA).

2. Undesirable in leakage of Halon into the MCR and MCRA.

3. Droplets of water behind Unit 2 vertical Control Board.

4. Discharge of CO froze plant comunications repeaters in the cable spreading r,oom, however hand-held devices functioned.

5. During the event personnel ran into areas protected 'by CO

. fire suppression for protection from the steam and had to b$

evacuated.

6. The discharge of Halon and CO limited access to the switchgear room and the cable sprea, ding room.

Evaluation and

Conclusion:

Both Halon and C0 systems discharged as a result of water entering intothecontrolp$nelsforeachsystem. The systems functioned as designed, discharge was excessive because the timer was shortedexcept thatdisc out. The CO,harge ofHalon and CO affected plant operations; operations personnel had to cope with 2both the accident and the resulting effects of the discharge of the these systems.

I. Associated Recomendations Maintenance Work Order Designation:

o It is recommended that a better designation of the specific component or piping systems on WPTS be established.

Feedwater Pump Discharge Check Valves:

o Inspect and repair as necessary the damaged feedwater pump discharge check valves in both units, o The feedwater pump discharge check valves should be periodically inspected at both Surry and North Anna Power Stations. A functional test of the check valve during plant startup may not be adequate to predict future operability. ]

i o An engineering evaluation should be performed to investigate the adequacy of the feedwater pump discharge check valve design.

o Review the equipment list for locations of similar check valves.

"C" Main Steam Trip Valve TV-MS-201C:

o The "C" main steam trip valve and the supporting air system should be inspected, repaired as necessary, and tested.

l l

1 Halon and C02 ystems:

S

}

o Several actions should be taken before returning the Halon and  !

CO system to full service. These are listed in Attachment D.

2 In addition to those recommendations, the following actions I should be taken:

a. An assessment of the Halon CO and ventilation systems todeterminewhyexcessiveHal$nand CO leaked into 2

the Main Control Room.

b. A review of both the Halon and CO, system design and make recomendations for diversification and

' environmental hardening.

c. Inspection of penetrations above the Main Control Room and Main Control Room Annex and removal or replacement.

of damaged penetration seals.

Onsite Computers:

o The clocks associated with the various onsite computers, eg.,

P-250 plant computer, event recorder, security computer, SPDS emergency response facility computer and GETAR should be reviewed for possible synchronization to help with post event evaluations.

l

)

E VIII. LIST OF FIGURES Figure 1 Incident Response Group Organization Chart Figure 2 Secondary System Diagram Figure 3 Turbine Building General Area Figure 4 Photograph of Failure Site Figure 5 Photograph of Elbow After Removal Figure 6 Sketch of "Reassen61ed" Elbow Figure 7 Macrophotograph of Elbow Interior Surface Figure 8 Sketch of Elbow on J-Nozzle Configuration Figure 9 Photograph of Elbow Thinned Wall Cavities Figure 10 Macrophotograph of Ductile Tear Figure 11 &crophotograph of Localized Thinning and Membrane Tearing

IX. LIST OF ATTA09 G TS Attachment A - "Suninary of Significant Operating Event" -

Oraft dated December 18,1986 Attachment 8 - Estimated Feedwater Pump Suction Piping Pressures During Nonnal Shutdown Transient (later)

Attachment C - Feedwater Pump Suction Piping Wall Thickness Requirements (later)

Attachment 0 - Technical Report dated December 15, 1986.

Halon and CO, System Actuation and Discharge, Surry Power Station and associated documents Attachment E - Surry Unit 2 Damage Assessment and Additional Inspection Recommendations Attachment F - NDE Testing Program 1

l

?

i I

m .

o. ,

l X. LIST OF REFERENCES

1. IE Information Notice No. 86-106: Feedwater Line Break, dated December 16, 1986. y I

2. Surry Unit 2 J-Tube Erosion, prepared by L. Albertin and P. !

Kuchtrka, Westinghouse Electric Corporation. Steam Generator Technology Division, P. O. Box 855, Pittsburgh, Pennsylvania, September 26, 1983.

3. oorrosion/ Erosion of Steels in High Temperature Water and Wet Steam, j prepared by Ph. Berge and F. Khan, Electric 1te de France - Service j de la Production Thermique Groupe des Laboratoires - 21, alle'e  ;

Privee - Carrefour Pleyel; Sumary and Conclusions of .the l

( Specialists' meeting held at "Les Renardieres " May 1982. I J

Inspection and Testing of "C" Main Steam Trip Valve - Preliminary 4

Report, Gary Thompson.

l

5. Erosion / Corrosion in Nuclear Plant Steam Piping: Causes and Inspection Program Guidelines, EPRI NP-3944.

l t

)

l I

l l

_ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ - - _ . _ _ _ _ _ _ _