ML20125D276

From kanterella
Jump to navigation Jump to search
Power Reactor Events and Issues
ML20125D276
Person / Time
Issue date: 10/31/1992
From:
NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
References
NUREG-BR-0171, NUREG-BR-0171-V01, NUREG-BR-171, NUREG-BR-171-V1, NUDOCS 9212150078
Download: ML20125D276 (19)
v • d • e


Text

- . . -

l POV/ER REACTDR 2."f*""-*

f"'%

EVENTS & ISSUES g *

      • ,+ -

i j

i Office for Analysis and Evaluation of Operational Data l).S. Nuclear Regulatory Commission

?

.sR-0171 R- PDR- :;jj i

AVAILABILITY NOTICE Availability of Reference Materials Cited in NRC Pub'ications Most documents cited in NRC publications will be available from one of the following sources:

1. The NRC Public Documer. f com, 2120 L E w W., Lower Level, Washington; DC 20555
2. The Superintendent of Documents, U.S. Govemment Printing Office, P.O. Box 37082, Washington, DC 20013-7082

]

3. The National Technical information Service, Springfield, VA 22161 Although the listing that fo,aws represents the majo ity of documents cited in NRC publica-tions, it is not intended to be exhaustive.

Referenced documents avDable for insoection and copying for a fee from the NRC Public i Document Room include NFC correspondence and interr al NRC memoranda: NRC bulletins, I circulars, information notict 2, inspection and investigatior' notices; licensee event reports; vendor reports and correspondence; Commission papers; ano applicant and Keensee docu-monts and correspondence.

The following documents in the NUREG series are available for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceed-ings, international agreement reports, grant pubhcations, and NRC booklets and brochures, Also available are regulatory guides, NRC regulations in the Code of Federal Regulations, and Nuclear Regulatory Commission Issuances.

Documents available from the National Technical Information Service include NUREG-series reports and technical reports pr3 pared by other Federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.

Documen:s available from public and special technical libreries include all open literature items, such as books, journal articles, and transactions. . Federal Register noticos, Federal and State legislation, and congressional reports can usually be' obtained from these hbraries, Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference proceedings are available for purchase from the organization sponsoring the I publication cited. .

Single copies of NRC draft reports are'available free, to the extent of supply, upon written request to the Office of Administration, Distribution and Mail Services Section, U.S. Nuclear Regulatory Commission, Washington, DC 20555.

l Copies of industry codes and standards used in a substantive manner in the NRC regulatory -

process are maintained at the NRC Library, 7920 Norfolk Avenue, Bethesda, Maryland, for use by the public, Codes and standards are usually copyrighted and may be purchased 4

from the originating organization or, if they are American National Standards, from the q

American National Standards Institute,1430 Broadway, New York, NY 10018. '

4 -

4

- .% . y , , ,..y.:.,h. ,,.-.n. , - - . . . - .w.. , , ., m  %. -. - - ,y+-

PG//ER REACTOR "4"? '""- m EVENTS & ISSUES +t' %

sb.....

Manuscript Completed: September 1992 Date Published: October 1992 Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission Washington, DC 20555 l

Power Reactor hents & Issues x

IN THIS ISSUE , .

Power Reactor Events & /ssues is published by the Office for Analysis and Evaluation of Operational Data (AEOD) of the Nuclear Regulatory Commission. The publication reviews selected operating events that have occurred at nuclear power plants and presents the results of NRC-spon-sored analysis of pertinent operating issues.

The feature article in this issue of Power Reactor Events & /ssues is based on AEOD Engineering Evaluation Report AEOD/E91-01,"A Review of Water Hammer Events After 1985," written by E, J. Brown. This study was initiated following several instances of water hammer involving the service water system at Arkansas Nuclear One. The task was to evaluate the need to reissue previous NRC guidance about water hammer or to suggest edditional measures to prevent or mitigate their occurrence.

This study concluded that the frequency of reported water hammer occur- -

rences continues to drop and no new phenomena were identified as causes of water hammer.-In addition, this study supports prior NRC conclusions regarding water hammer; however, some aspects that could impact safety and were identified in the study havo not been previously emphasized.

l Pagelii

1%er lleanor I:venn & Inun A REVIEW OF WATER HAMMER EVENTS AFTER 1985

SUMMARY

reduce the number of water hammer events were Water hammer is a rapid pressure change caused not supported by cost-benefit guidelines (Ref. 5, by a change in the velocity of a fluid in a closed 6). The reassessment determined that the volume (Ref.1). These pressure changes can frequency of water hammer events had actually create loads on piping and other components that decreased significantly since the initial review exceed their design limits. Water hammer performed for resolution of USI A 1. In addition, no new phenomena were identified as causes of incidents have been attributed to such causes as ^

rapid condensation of stcam pockets, steam. water hammer. (USI A-1 identified 148 reported water hammer events from 1969 to 1980, while driven slugs of water, pump startup with partially empty lines, and rapid valve motion. the reassessment after the San Onofre event ,

identified 40 reported water hammer eventt from Water hammer events involve not only piping 1981 to 1935.)

systems but components connected to them, as well as components used to support them Most While performing a Diagnostic Evaluation at of the damage caused by water hammer has Arkansas Nuclear One (ANO) during August and been relatively minor (o g-, damaging only pipe September 1989 (Ref. 7), NRC staff identified hangers and restraints); however, there have been several water hammer incidents that involved the some incidents that have resulted in the loss of service water system (Ref. 8). These water hammer events at ANO were not reportable as function of major components. Significant water Ucensco Event Reports (LERs) under 0 CFR 50, hammer effects have also been observed in 550.73. The Office for Analysis and Evaluation of instrumentation systems.

Operational Data (AEOD) reevaluated water After several water hammer incidents resulted in hammer events to assess the need to reissue piping and valve damage, the water hammer issue previous NRC guidance about water hammer or was determined to be an Unresolved Safety issue to suggest additional measures to prevent or ~

(USI), and classified as USl A 1, Water Hammer mitigate its occurrence. A summary of AEOD's -

reevaluation, based on AEOD Engineering (Ref. 2), USl A-1 was considered to be resolved by the publication of NUREG-0927, Revision 1 Evaluation Report AEOD/E91-01, A Review of Water Hammer Events After 1985 (Ref. 9),

Evaluation of Water Hammer Occurrence in Nuclear Power Plants, in March 1984 (Ref. 3). In performed by E. J. Brown, is given below.

resolving USI A-1, the NRC staff concluded that new requirements to reduce the number of water EVENT DESCRIPTIONS hammer events were not supported by cost-benefit guidelines. However, guidelines were AEOD's current assessment of water hammer provided concerning measures for preventing and identified 12 reported water hammer events mitigating water hammer. between January 1980 and March 1990 (Table 1).

These events were reviewed to determine whether The NRC reassessed resolution of the water they were associated with any new physical hammer issue after a water hammer event at San phenomena. This review also concentrated on Onofre Unit 1 in November 1985 damaged plant identifying common mode failure aspects and equipment and challenged the integrity of the lessons that may be useful to assist other plants plant's heat sink (Ref. 4). The reassessment, in preventing situations that could result in water completed in 1986, reconfinned the original hammer. Summaries of the 12 reported water conclusion that new or additional requirements to hammer events identified in E91-01 follow.

Pagei

-- - ... .- - . .- .. . . .-.- - - . _ _ ~...- - - . _ . - . . _ .. .

Mer Rewor Eirna & !ssues ;

y e ~

g.m ;w -

+ , a mg ; -- ,

E I v <

a m, r , , 3, ,

,,.g W al - t l'w E A5a V_X .jk VU%. 2 _A%

TABLE 1 - WATER HAMMER EVENTS AFTER 1985 Plant I Type System Event Description Susquehanna BWR RHR Shutdown Valve isolated while switching from Pump *D' to Unit 2 Cooling Mode Pump *A.* Partial pipe draindown - not refilled prior to restart. . .

Shearon Harris PWR Steam generator Rapid motion of isolation valve resulteo in snubber -

i blowdown damage. --

l Trojan PWR Accumulator fill Transferring water between accumulators with high '  ;

lines differential pressure. No procedure. '

South Texas PWR Auxiliary Feedwater Pressure fluctuations developed by crossover flow Unit 1 (AFW) Vent Lines control valve throttling.

Indian Point PWR Feedwater Cycling a flow control valve (FCV) caused pressure Unit 3 drop between FCV and isolation valve.

Oyster Creek BWR  ! solation Condens- Steam lines to IC partially filled with water. Suspect-er (IC) ed reverse flow through one-half of each IC.

Waterford PWR Steam Generator Cycled inside-containment isolation valve to verify.

Unit 3 Blowdown operability without closing outside isolation valve, e

Oconee Unit 3 PWR Main Steam Suspected water accumulation in drain line for the -

turbine bypass line to the A condenser.

Palisados PWR Accumulator Safety injection tank vented to 50 psig and back injection leakage from primary system resultedc ANO Unit 2 PWR Steam supply to Condensate buildup at low point resulted in water AFW slugging on Emergency Feedwater (EFW) pump startup.

Dresden Units BWR High Pressure Leaking HPCI irdection valve and check valve 2 and 3 Coolant injection caused FW leakage into HPCI system and void (HPCl) formation.

Dresden Unit 2 BWR HPC) FW leakage into HPCI system due to MOV failure _

to completely close after stroke timing test.

Susquehanna Unit 2 switchover to the "B" pump was completed, pump ..

'D' was shut down. However, at about the same On October 12, 1986, the unit was in the shut- time, outboard isolation valve F008 in the letdown down cooling mode with the 'O' residual heat line from the "B" recirculation pump to the suction removal (RHR) pump in service (Ref.10). The "B" of the "B" RHR pump closed, causing the "B" RHR RHR pump was then started (see Figure 1). After pump to trip. Operations personnel reset the logic

'2--

g 2,.'._b,-

~ Page 2 #

l. j

m . _ . _ - - . _ . . _. - _ . _ . _ . _ _ . _ - _ _ - _ _ _ > _ _ _ _ . _ _ _

Power Reactor bents & Issues p -

-, ;w 7

- m, <1 -y . -xm n <m, x

,n r y n; ca J - sn 7

, < m o

t. v 6 t m s /

and reopened valve F008. Hoopening valve F008 The corrective action was to increase the stroke

. resulted in water hammer and vane F008 again time ~of the valve.

closed. The water hammer was caused by partlai draindown of the system through a pathway to the condenserc This pathway is used to control Trojan the reactor water level while the RHR system is in service. The operators failed to fill and vent the On May 12,1987, the *A* accumulator 1-loch fill

' RHR system piping after the 'B' RHR pump line ruptured at the nozzle-to-pipe weld white tripped. transferring water from the 'A' accumulator to the

'D' accumulator (Ref.12). Approximately 2000 ga!!ons of water . were . released to - the 1 containment. T he 'A' accumulator was at 583 psig and the 'O' accumulator was depressurized. The Le a%'* Q weld failure was attributed to low-cycle, high-(* _

~

" stress fatigue cracking. After repairing the weld, the fillline again ruptured while transferring water i ~ ~-

, p.

irom the *A* accumulator to the *D* accumulator.

V uu r -

~N This time, the 'A' accumulator was at 650 psig.

The *D* accumulator was still depressurized. The I Q" 7 cause t'! the fill line rupture was excessive reverse -

g, flow through the packless diaphragm globe valve. l This reverse flow caased cyclic vibrations.

~~ == The root cause of the fill line ruptures was =

I inadequate procedures Procedures were in place i to transfer water betwoen accumulators via the gg. sample lines. No procedures existed . . for 44 transferring water between accumulators via the fil: lines.- Water was transferred between the-

-(.73-Dt ,D;-- accumulators via the filllines because the sample

~~ lines were ta0ged out for maintenance. A dynamic analysis, which modelled backflow through the fill line, showed that hydraulic loads far in excess of Figure 1 those necessary to fail the pipe would be imposed on the nozzle-to-pipe weld. In addition,

- Susquehanna Unit 2 RHR Shutdown Cooling a backflow test through a packless globe valve, similar to the vane in the accumulator fill line, resulted in a pipe failure at a flow of about 70 Shearon Harris Unit 1 gpm. Operating procedures were revised to -

prohibit water transfer between the accumuf atorsc On April 22,1987, plant personnel discovered a damaged snubber on a steam generator blow- South Texas Unit t down pipe (Ref.11). In addition, two broken pipe '

On November 5,1987, while in hot shutdown and supports and various pipe displacements were

- found. The steam generator blowdown piping is. prior to initial criticality, a 1-inch diameter, double .

safety-related because of- concerns of a' high vane vent line in the pump discharge piping of -

energy pipe break outside the reactor contain- the auxiliary feedwater (AFW) train *A* broke off mont. The cause of water hammer was attributed (Ref.13). A second failure occurred 3 days later to rapid motion of the blowdown isolation vane. In a similar manner in a double valve instrument

, ,, , -n .

.'O :' ' y n

.- is' ls .k .N k'; i.u $ h a n y

p

' . Page 3

l*gwr A'earsor I:veMs & hsues g- ,

.n y m e, -- - ,  ? -- - - -

n a

- = __

tap for the *D* train AFW pump discharge lino signal was generated by motion of the dischargo (see Figuro 2). The initial assessment cited the valvo (BFD-2 32) non-rotor-driven limit switch cause as water hammer resulting from improper (Crano/Tolodyno Model T 40-80), The limit switch venting of the - AFW cystem. The AFW system motion was caused by a water hammer shock however, continued to exporlonce vibration when generated by the cycling of the No. 32 MBFP in operation. Subsequent testing revealed that recirculation valve (FCV 1116) wh!!e the manual over-throttling the flow control valves in trains "A" isolation valvo (BFR 1-32) was closod. The water or *D* Introduced pressure fluctuations at a domi- hammer occurred because of the pressure drop .

nant frequency of 24 hertt This frequency between the recirculation valve (FCV-1116) and matched the natural frequency of the piping the manual isolation valve (BFR 132). The cause system. Mechanical stops were installed to of the water hammer was the recirculation valve prohibit excessive throttling of the valves. cycling while personnel were troubleshooting a _

faulty limit switch on the valvo.

Oyster Creek

,b s pre On September 28,1988, while operating at full w L 1 3 j, power, operators determined that both the 'A'

_y X g b, g and *B* isolation condensers (ICs) were operating n n n,.

In an unanalyzed condition (Ref.1S). Temperature gj ' '"'* Aj data sugges'od that the steam lines to the ICs ,

were at least partially filled with water, Analysis of

,  ; the temperature data suggestod existence of reverse flow through onmhalf of each IC and the n g} L E ww possibility of subcooled condensate buildup in the l pq '" ,g o steam lines to the 10. This raised concom over potential eff ects of increased piping loads, thermal x, stresses, and possible water hammer. The reactor p"o"m,,

was shut down. The event was the subject of an M a =*

NRC Augmented inspection Team review (Ref.16) pN , , . b* and extensive licensco followup.

p' g L The partial filling with water of the steam lines to g fy t._ tho _IC was outside the previously evaluated a

  • ** I operating conditions. A subsequent licenseo

@N $M* '

Technical Data Report concluded that water hammer had not occurred during this event and that other postulated transients would not result in Figure 2 water hammer (Ref.17). Operating procedures were implemented to limit water accumulation in -

South Texas Unit 1 AFW System the steam. lines.

Waterford Unit 3 Indian Point Unit 3 During a routine system walkdown on July 14,-

On March 31,1988, main boiler feedwater pump 1989, with the plant at full power, plant personnel (MBFP) No. 32 tripped with the reactor at full found a damaged pipe support on a steam power (Ref.14). The pump tripped in response to generator blowdown pipo (Ref.18). It was later a " discharge valve not fully open* signal. This determined that in this condition, the structural

. ~

A 4  % h Page 4 l

Power Reactor Dents & hsues l

integrity of the blowdown system (including the Oconee Unit 3 outside containment isolation vane and shield building penetration) could not be assured during After the main turbine tripped on March 6,1989.

a seismic event. the reactor tripped as expected (Ref.19). Follow-ing the trip, a water hammer in the main steam The support was observed to be undamaged turbine bypass lino to the *A* condenser damaged during a walkdown in June 1989. Between the two three pipe supports and deflected the bypass line system walkdowns, personnel performed the 12 to 18 inches. Subsequent investigations operating test " Engineered Safety Feature Actua- indicated that the most likely cause of the water tion Signal (ESFAS) Subgroup Relay Test." This hammer was water accumulation near an orifice test invoNed cycling the inside containment isola- in a drain line; however, no obstructions could be tion valve to verify its operability in response to a verified A Station Problem Report was initiated to ~'

containment isolation actuation signal or an emer- resolve problems with improper draining of the gency feedwater system actuation signal. The pipe (Ref. 20).

procedure did not require shutting the outboard containment isolation valve or minimizing the Palisades blowdown flow prior to opening the inboard iso la-tion vaNo (see Figure 3). Thus, a water hammer On April 25,1990, plant personnel found gross transient most likely occurred when the inboard deformation of some of the selsmic pipe supports isolation vane was opened. In piping between the "A" and *D* safety injection tanks (SITS) and the primary reactor system (Ref.

21). Minor damage was also found on piping supports for the "B' and "C" SIT piping. An evalua-tion found that a water hammer occurred on t October 1,1987. This water hammer could have

  • m,,,

Z "'" caused the damage; however, the exact cause a was not clear. At the time of the water hammer, the reactor was in hot shutdown and the primary coolant system temperature was greater than 500-g 0F. The SITS normally have nitrogen cover gaa at ,

200 psig, but the "A" tank had been vented to 50 psig. The licensee believes that this lower  ?

{

a.

, g o.,

pressure in combination with possibly leak.ng N * * * " " " " check valves from the primary system resulted in o flashing conditions and water hammer. Plant IniO procedures were changed to prchibit bleed down g of the SITS.

l mcceu.uw e' l """"~~~ Arkansas Nuclear One Unit 2 s , m - w., A mechanical snubber on the main steam supply line to the emergency feedwater (EFW) system turbine-driven pump was found inoperable Figure 3 because of severe degradation of the snubber intemals (Ref. 22). The damaged snubber was Waterford Unit 3 found while conducting inservice inspections in Steam Generator Blowdown System accordance with the plant's Technical Specifications during a refueling outage in Febr-uary 1988. The failure mode indicated that the PageS

' Power Reactor heats & hsues

_, s e . . - - ._ _

a - -.

- . . . - - ~ ,, u s snubber had been subjected to an overload condition during the just completed operating cycle. The most probable cause was determined . O g;b QM@4 o to be condensate buildup at a previously *

'"{ fph,,

unidentified low point in the EFW steam supply ""*

piping and subsequent " water slugging

  • upon starting the turbine-driven EFW pump. Poor h". ,g m, system maintenance and inadvertent bypassing of M+h steam traps for extended periods also contributed i bT,( ' 'N to the condensate buitdup. This same snubber ?4Ufa was found inoperable in the course of performing 2 2 n=m 2 nl?J,u inservice inspections during earlier refueling out- . . :i ages. During one of these refueling outages, 747_'Q"L-  !

snubber failure was attributed to excessive ' steam l slugging' resulting from a past problem with ""

overspeed trips of the turbine-driven EFW pump.

During another previous refueling outage, snubber i em failure was attributed to vibration and overload.

This fa!!ure mode was common to a large number Figure 4 of mechanical snubbers. Corrective action included modifying the EFW steam supply line t Dresden Units 2 and 3 remove the low point and inspecting, cleaning, Basic HPCI System, Normal Valve Line-up and rebuilding, as necessary, the steam traps.

A pipe temperature survey conducted in October Dresden Units 2 and 3 1989 revealed that the HPCI pump discharge pipe temperature was 246'F, while the temperature of A series of three events involving water hammer the piping between MOVs 2-2301-8 and 2-2301-9 in the high pressure coolant injection (HPCI) was 2750F, An evaluation determined that steam system occurred at Dresden Units 2 and 3 (Ref. voids could form in certain sections of the pipe 23). Two of these events were reviewed by an under these conditions inspection of the Unit 2 NRC Augmented inspection Team in October and HPCI discharge piping supports found deficien-November 1989 (Ref. 24). Preliminary indications cies in 47 percent (or 16) of the supports. The of a precursor to the initial Unit 2 damage in the Unit 2 HPCI systern valve lineup was changed so HPCI system were provided by increasing HPCI that the injection valve function was moved frc cubicle temperatures in May 1989. The pipe tem, MOV 24301-8 to MOV 2-2301-9 (see Figure 5).

perature at the HPCI pump was 1400F, while the Also, with MOV 2-2301-8 open, MOV 2-2301-10 piping between MOVs 2 2301-8 and 2-2301-9 was becomes an isolation valve that is now subject to 1600F (see Figure 4). Further pipe temperature feedwater pressure.

measurements taken in July 1989 found the HPCI

pump discharge piping at 175 F, while the pipe As a r_esult of the elevated temperatures
- between MOVs 2-2301-8 and 2-2301-9 was 2200F. discovered on the Unit 2 HPCI pump discharge l Closing MOV 2-2301-9 produced pipe tempera- piping, personnel investigated HPCI pump tures of 106"F after 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />. Pipe temperatures discharge pip!ng at Unit 3. Temperature measure-returned to 220 F after MOV 2-2301-9 was ments obtained with an infrared thermometer in reopened. This showed that check valve 2-2301-7 October 1989 found HPCI pump discharge pipe and MOV 2 2301-8 were both leaking feedwater temperatures ranging from 256 F between MOVs l' back to the HPCI system piping. 3-2301-8 and 3-2301-9. to 112 F at the dischar0e of the HPCI pump.This was evidence of possible steam void formations in the Unit 3 HPCI pump C -

Page 6

Power Reactor Everas & Issues n

l discharge piping. Further monitoring of the piping Dresden Unit 2 temperatum and pressure in early November 1989 revealed that the temperature in the piping was in March 1990, after completing routine HPCI increasing. The temperature increase was attribut- valve operabl!!ty surveillance tests and while ed to leakage past MOV 3-2301-10. This valve performing quarterly valve timing tests on the Unit was operated both electrically and manually in order to seat the valve. MOVs 3-2301-15 and -49 were also closed. After MOVs 3-2301-15 and -49 were closed, the pump discharge pressure ,nX

{ ([

mga "Ap increased from 650 psig to 1070 psig. This  % _Am ""*

pressure increase confirmed that the leakage y .

through MOV 3-2301-10 was successfully ,

stopped. .I?'. m.,

{mj T s. *D**

inspection of Unit 3 HPCI discharge pipe supports mb found deficiencies in 52 percent (or 21) of the supports. The Unit 3 HPCI system valve lineup 2 2** Z edyb was changed so that MOVs 3-2301-15 and -49 ,g g were normally closed and MOV 3-2301-48 was -H  ::

normally open (see Figure 6). @'{Q f

' ~#

{ nw

+

s + ,

Figure 6

-H-* p, ,

>C'#G ,

Dresden Units 2 and 3

,g

  • gg a HPCI Revised Alternate Valve Une-up

{ }-- %** g (Unit 2 Post-March 1990 Event and mn Unit 3 Post-October 1989 Event) ~

2 Z=+

2 da 2 HPCI pump discharge valves, the Unit 2 Shift

@a =p Supervisor heard banging noises coming from the

=,' HPCI pump discharge piping (Ref. 23). The valve

" " " timing test was terminated and the HPCI system valve lineup was returned to the configuration

== , shown in Figure 5. The pipe banging and motion
    • was monitored until it eventually ceased about 1-1/2 hours later. Valve manipulation and tempera-Figure 5 ture measurements along the HPCI pump discharge pipe led the licensee to conclude that Dresden Unit 2 HPCI Alternate Valve Line-up feedwater leakage back through HPCI test return (Post October 1989 Event) valve MOV 2-2301-10 was the root cause of this event.

Based on further investigation, the licensee postulated that MOV 2-2301-10 did not fully close after performing the HPCI system tests. However, Page 7

. Power Reactor bem & Issus

.y; n _

, . - - .g , - , . w nn n ,m, e -,

US.kl, , + r + & vA %.L -u-~ 2 a < >> 'S saGK

~

s ,< ~ vi.n Akr% $-$ 2.

rnay or may not fully close when given a close These interactions appear to' perturb monitored -

signal. This valve does not have a seal in feature parameters resulting in undesired isolation to complete the stroke after closure initiation. In signals. Shutdown cooling will be lost and water addition, the limit switches are set to provide a hammer may result if realignment procedures are torque switch bypass function in the open not adequate. This phenomenon does not seem direction untG the valve is 25 percent open. This to be well understood since several BWR plants limit switch also controls the indicated " valve have experienced this problem and have closed

  • light in the control room. Thus, an attempted different solutions with varying degrees operator removing a closure signal shortly after of success.

the control room panel lights indicate that the valve has closed could leave the valve nearly 25 The recent water hammer events seem related to percent open. A procedure was introduced to lack of implementation of the guidance issued in continue the closure signal 30 seconds after the

~-

the resolution to USl A-1 (Ref, 3). For instance, panel lights indicate MOV 2-2301-10 has closed. some causes cited were failure to fiti and vont The Unit 2 HPCI system valve lineup was again properly, rapid valve stroking, lack of guidance revised, and left to match the HPCI system valve about system configuration, low point water lineup for Unit 3 (see Figure 6). accumulation, depressurization of a system which could cause local flashing, and bypassing steam DISCUSSION traps. Thus, water hammer can result when plant staff are not vigilant conceming system conditions ,

From Table 1, it is evident that water hammer and changes in system arrangement.

events have occurred in both Boiling Water Reactors (BWRs) and Pressurized Water Reactors The events at Trojan and Dresden Units 2 and 3 (PWRs). The BWR events involved the shutdown  !!!ustrate this situation.The events involved system -

cooling mode of RHR, the isolation condenser, alignment changes in which the significance may and HPCI, These BWR systems have all been not have been fully appreciated. At Trojan, there identified with water hammer events in previous was an attempt to transfer water to accumulator studies. The PWR events involved main steam, "D' from accumulator "A" via the fill lines for each AFW steam supply, steam generator blowdown, accumulator because the sample lines, which had feedwater, accumulator (fill lines and injection to procedures for such application, were tagged out the reactor coolant system), and the AFW system for maintenance There were no procedures for vent lines. Most of these PWR systems have been transferring water between accumulators by use associated with water hammer in previous studies. of the fill lines. There were two attempts at water transfer, and two pipe ruptures. These attempted The causes of these recent water hammer events water transfers used an approach that had not are consistent with those reviewed in previous been reviewed for systems with a pressure differ-studies. However, previous studies did not ential of several hundred pounds per square inch.

emphasize the interactions between systems. For Thus, the need to fill the accumulator led to using example, the event at South Texas Unit 1 was- an unauthorized approach without appropriate initially believed to be water hammer caused by checks to satisfy safety considerations.

improper venting of the AFW system. Subsequent evaluation determined that the flow control valves The series of water hammer events in the HPCI were introducing pressure fluctuations when they system at Dresden Units 2 and 3 were all related were in a highly throttled position. The pressure to operating the plant with leaking isolation valves, fluctuations matched one of the natural frequen. - Plant operation with the leaking valves was initially cies of the piping system, causing pipe rupture. attempted by using monitoring techniques intended to' identify undesirable system The closure of shutdown cooling isolation valves temperature conditions. Subsequent efforts to has been identified with fluid systems interactions. operate with leaking valves involved . both -- _

p . , . .

6 e a ,- w L

a.;.p.;__

Page 8 l

1

l 1%er Reactor Events & Issues monitoring and realignment of MOVs. Water retum valve that did not have a seal-in feature for hammer events occurred in each instance. Simple the closure signal eventually lead to formation of monitoring was inadequate because temperatures local voids and water hammer in the HPCI could still increase to the point where local steam system. (The valve was 25 percent open when the voids could form and subsequent water hammer control panel light indicated closed. The control could occur. toom light indicated closed because the MOV control features used the same limit switch to set HPCI system valve realignment at Dresden Unit 2 the open torque switch by-pass and the closing involved closing the normally open isolation valve light indication; this approach may still be used in outside containment (making this valve an several plants.) Therefore, a relatively simple injection valve), opening the normally closed realignment of MOVs to provide HPCI injection isolation valve inside containment, and closing the evolved into a complex situation requiring detailed condensate storage tank return valve (making this knowledge of subcomponent control features and valve a pressure isolation valve). Rea!igning the settings and system (component) operational valves in this arrangement has several features of requirements (including IST tests) in order to interest: 1) assurance of MOV operability against protect the HPCI system from potentially full differential pressure (about 1200 psi), 2) a damaging water hammer events as well as relationship between the new alignment and valve providing assurance for injection capability.

position changes required for MOV tests, and

3) MOV control features such as signal sea!-in or FINDINGS torque switch bypass aspects.

Evaluation of the data identified from the search Also at Dresden, the normally closed HPCI for water hammer events for January 1986 injection valve (which leaked) was evaluated for through March 1990, resulted in the following operation against full differential pressure (1135 findings.

psi) as a result of IE Bulletin 85-03, Motor Operated Valve Common Modo failure During 1. There were 12 water hammer events identified Plant Transients Due to improper Switch Settings.

from January 1986 through March 1990. This issued November 15,1985 (Ref. 25). However, the indicates the event frequency is dropping MOVs used in the realigned configuration compared with previous studies (148 events '

(normally open isolation valve and condensate ~

from 1969 to 1980 and 40 events from 1981 return valve) were not evaluated for operation to 1985).  ?

against high differential pressures until NRC Bulletin 8543, Supplement 1, was issued on April 2. The causes of the 12 recent water hammer 27,1988 (Ref. 26). Supplement 1 addresses the inadvertent operation (closure or opening due to events reviewed in AEOD/E91-01 (Ref. 9) were similar to causes identified in previous mispositioning) of motor-operated valves. Thus, studies. Thus, there were no new phenomena a system realignment prior to April 1988 could identified.

have resulted in placing MOVs in a configuration for which operability requirements may not have 3. The 12 water hammer events appear related been adequately addressed.

to either a failure to implement the guidance issued in the resolution to USl A-1); a less Plant Technical Specifications at Dresden still than vigilant attitude concerning system required MOV stroke tests and HPCI system flow conditions, operations, or changes in system tests as part of the inservice test (IST) program. alignments that could result in water hammer; in order to conduct these tests, it was necessay or an insufficient understanding of system to use the leaking valves temporarily as isolation conditions, including component operational valves againct the feedsater system pressure. characteristics, that could cause water This aspect, in conjunction with the condensate hamrner.

Page 9

l I

. Q '*!

  1. ? M " W !'?!' , ~ .

y g, ~,g , ,. w e e n; g __m g,w ,m g%,,wMk, . , , ,,

kn '

. ' Ldm as M , . u ~ < 3' ' wz M dau a su & L a;c

4. The causes of the 12 water hammer events studies, there are some aspects that were not include fauure to fill and vent properly, too previously emphasized which impact safety. The -

rapid valve stroking, lack of guidance about specific areas include:

system configuration, water accumulation at low points, depressurization of a system 1. failure to irr.plement the guidance issued in which could lead to local flashing. and by- the resolution to USl A 1, passing steam traps.

2. hydrodynamic interaction between systems, S. Some water hammer events show that u hydrodynamic interactions between systems 3. system realignments involving MOVs that may may occur. involve complex issues concerning compo-nont operability as well as water hammer, and
6. System realignments involving MOVs may involve very complex issues that could affect 4. closure of letdown iso _lation valves in the shut-component operability or possibly result in down cooling mode of RHR for BWR plants.

water hammer events.

The first three areas were clearly evident in the 12

7. The water hammer event at Susquehanna 2 water hammer events reviewed. The fourth area, involved closure of the letdown isolation valve as demonstrated by the water hammer event at during attempts to use the shutdown cooling Susquehanna 2, shows that the closure of the '

mode of RHR. The RHR pump tripped and letdown isolation valves can cause water hammer shutdown cooling was lost when the isolation events. However, closure of the isolation valve will valvo closed. Several events similar to the also result in loss of shutdown cooling.

Susquehanna 2 event with respect to the loss of shutdown cooling were identified (Ref. 9). NRC issued Information Notice '(IN) 91-50 to auclear power plant licensees as a result of this CONCLUSIONS study. This IN discussed the events reviewed in this report, indicated that NUREG 1027,' Revision The assessment process for USl A-1 was based 1, (resolution of USl A-1), addressed many of the on a disciplined approach to review each event causes cited by these water hammer events, and and identify affected systems, determine event alerted the industry about hydrodynamic frequency, and establish the phenomena that interactions between systems as well as system -

caused the water hammer. realignment concems.

Based on the similar evaluation performed in AEOD/E91-01 (Ref. 9), it was found that the 12 REFERENCES reported water hammer events represent a reduction in frequency of occurrence of water 1. Generic Safety Issue Tracking and Evaluation hammer at operating plants and that no new Summary Description, " Water Hammer," Electric-physical phenomena were identified as causes of Power Research Institute, NSAC 81. July 19841 water hammer. Therefore, the recent operating experience is consistent with the resolution 2. A Pr/oritization of Generic Safety issues, "USl conclusions for both USl A-1 and the subsequent A-1, Water Hammer," NUREG/0933, U.S. Nuclear reassessment after the San Onofre event in 1985. Regulatory Commission, August 1987.

Thus, there appears to be no need to revise . .

NUREG 0927. 3. Evaluation of Water Hammer Occurrences in Nuclear Power Plants, NUREG-0927, Rev.1, U.S.

< Although these recent water hammer events are Nuclear Regulatory Commission, March 1984.-

comparable ' with those reviewed in previous

(

y ,

,ww-~ -y; ~

__ ,; g -

~

, m uu.u.*a am

_y u - >m . -

m . 5:c m w *o - s %e . < x #m s Page 10

l Power Reactor Evenu a Issues

4. Loss of Power and Water Hammer Event at San 15. LER 50 219/88421, " Plant Shutdown Due to Ono/re, Unit 7, on November 27,1985, NUREG. Both Isolation Condensors Being in an Unanaly-1190, U.S. Nuclear Regulatory commission, zed Condition Due toThermo-Hydraulic Operation January 1986. Outside Normal System Design," October 26, 1988, Oyster Creek.
5. Re-examination of Water Hammer Occurrenc-es, AEOD/E608, U.S. Nuclear Regulatory Com- 16. NRC Inspection Report 50-219/87-029, Sep-mission, July 14.1986. tomber 28,1987 for Oyster Creek.
6. U.S. Nuclear Regulatory Commission, H. R. 17. Oyster Creek Technical Data Report,950.

Denton to V. Stello, Jr.. " Review and Assessment of Water Hammer Occurrences Since CY 1981,* 18. LER 50-382/89415. " Containment isolation July 7,1986. Valve Inoperable Due to inadequate Design and Inadequate Procedure,' August 28,1989,

7. U.S. Nuclear Regulatory Commission, Diag. Waterford Unit 3.

nostic Evaluation Team Report for the Arkansas Nuclear One Units 1 and 2, December 1989. 19. LER 50-287/89402, " Reactor Trip Due to Turbine / Generator Trip," April 5,1989, Oconee

8. U.S. Nuclear Regulatory Commission, J. M. Unit 3.

Taylor to T. E. Muriey, et al., " Staff Actions Result.

Ing From the Diagnostic Evaluation at Arkansas 20. Oconee 3 Station Problem Report.

Nuclear One,* December 21,1989.

21. NRC Inspection Report 50-255/90-i4,.p m
9. E. J. Brown, A Review of Water Hammer 13,1990 for Palisades.

Events After 1985, AEOD/E91-01, U.S. Nuclear Regulatory Commission, February 1991. 22. LER 50-368/88 023, " Mechanical Snubber Failure on One Train of the Emergency Feedwater

10. LER 50-363/86-01541," Primary Containment System Due to Water Entralnment in Steam isolation Valve Closes Twice Due to a Spurious Piping Results in Operation Prohibited by High Flow Signal,' December 18,1986, Susqueha. Technical Specifications," June 1,1989, Arkansas nna Un:t 2. Nuclear One Unit 2. _

11.LER 50-400/87 029-01, ' Steam Generator 23.LER 50-237/89-029-01, " Elevated HPCI Blowdown Piping Snubber inoperable," July 31, Discharge Piping Temperature Due to Reactor 1987, Shearon Harris Unit 1 Feedwater System Back Leakage,' May 1,1990, Dresden Units 2 and 3.

12. LER 50-344/87-013-01, *Accurnulator Fill Une Rupture Due to Backflow Induced Vibration." July 24. NRC Inspection Reports 50-237/89423 and 10,1987, Trojan. 50-249/89-022, November 21,1989 for Dresden.
13. LER 50 498/87-016-01, " Auxiliary Feedwater 25. Motor Operated Valve Common Mode Failure System Report -Inyestigation of Hydraulic Design During Plant Transients Due to improper Switch Error,' March 15,1988, South Texas Unit 1. Settings, IE Bulletin 85-03, U.S. Nuclear Regulatory Commission, November 15,1985.
14. LER 50-286/88-002," Reactor Trip, Main Boiler Feed Pump Trip Due to Main Boiler Feed Pump 26. Motor-Operated Valvo Common Mode Discharge, Valve Limit Switch Actuation Caused Failures During Plant Transients Due to improper by Water Hammer Induced Vibration," April 21, Switch Settings, IE Bulletin 8543, Supplement 1, 1988, Indian Point Unit 3. U.S. Nuclear Regulatory Commission, April 17, 1988-Page 11

, , , . _ . _ _ . . ~ . _ ~ _ . _ . . -.. _. -__.=_-_ _ _ . . . _ . . . .

L i

i 1*ower Reactor Evena e issues gfc

  • Sy.:c,_

7 yr '*?gy(f:S fpygs@-y ,my ma >morpearrn a na,u aa;m r9g;1e,ug re=

n+. ,.w nr?,;Myg;p*<- a.3+. - ,'mb . a;.n

._ .x, .. ::,W

.an- .,a., ,. n . a na., c ., sa, ua.w, k .

ABOUT THE AUTHOR Earl J. Brown is the Chief of the Engineering Section, Reactor Operations - .

Analysis Branch, Division of Safety Programs of the Office for Analysis and -  !

Evaluation of Operational Data (AEOD). Dr. Brown joined the NRC in 1973 ; I as a Mechanical Engineer in the Structures and Components Branch in the former Office of Standards Development.' He joined AEOD shortly after its i formation in 1980 as a member of the Reactor Operations Analysis Branch. l Prior to joining NRC, Dr. Brown wa: with the Mechanical Design Group of the Nuclear Division of the former Combustion Engineering Corporation.- .l Dr. Brown received his B.S. In Mechanical Engineering from the University of New Hampshire, an M.S. In Engineering from Purdue University, and -

a Ph.D. In Applied Mechanics from Lehigh University.

b rp:w ~ m mu -nu >;gv e s-m-  ;;.v.; -

m- ^ yyo ,.ygn , g y7 n , mgy c ;,.gm.i.. , . , ,

e 5.i+ A ilj M ASe;, UEniv .,A, i LK.i.v (~B6M Jyi A g.; ;.;E&

  • ~
.,,1 J.w 3.:4;6 ,(@.f,. sM7;is .% a .Jydh?'C.hdy , ' .e, Qi.,;,;.;,'.*
Page 12 -

_ - _ .a - -

N4C FOHV 335 U.S. NUCtIAR f(CeULATOHY COMMISSION 1. Fit. FORT NUMBE R i (2 69) (Amoiped t>y NRC,C4f Volu l

NRCM 1102. Supp., fiev., and Addendum NJm-370t3792 t,ers, it arif.)

BIBLIOGRAPHIC DATA SHEET (s . me,umon. on in. .. verse NUREG/BR-0171 Vol.1 t Tmt Ano suuTmr Power Reactor Events & lasues 3 DAT{ ftLIORT PUBUSHED MONTH yggg October 1992

4. FIN Of4 GRANT NUMDCR S AUTHOHis) 6 TYPE OF f4CIORT
7. fTRIOD C04RtD (incluelv. Dates)
6. Pt.Rf ORMthG OAGANilATION - NAME AND ADD 4E 55 et NHC. proede Diamn. Or4 e or fiegion, U s Nudnar Regulatory Commission, and enMing a.17ess. If contiact;<, psovide name ard man g addrene.)

Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission Washington, DC 20555 9 STONSOf ttNG ORGAN'ZATION NAMC AND ADORE $$ (if NRC. type " Lame as abus*, if contractor, proverte NPC DMalon,14fice or Ngion.

U S. Nwicar WgoWtory Commlue:.n. and mamng address.)

Same as 8, above 10 SufMIME NT ARY NCIES it Ah5 TRACT (700 woeds or less)

Power Reactor Events & /ssues is published by the Office for Analysis and Evaluation of Operational Data (AEOD) of the U.S. Nuclear Regulatory Commission (NRC). This publication presents assessment of selected operating events at U.S. nuclear power plants. The objective of this periodical is to feedback the results of NRC-sponsored analysis of pertinent power reactor events and issues to the nuclear industry, specifically, to the Operations, Maintenance, and tne Training Managers at all the operating reactor sites. The feature article in this issue is based on AEOD Engineering Evaluation Report AEOD/E91-01,"A Review of Water Hammer Events After 1985,' written by E. J. Brown. This study was initiated following severalinstances of water hammer involving the service water system at Arkansas Nuclear One. The task was to evaluate the need to reissue previous NRC guidance about water hammer or to suggest additional measures to prevent or mitigate their occurrence. This study concluded that the frequency of reported water hammer occurrences continues to drop and no new phenomena were identified as causes of water hammer. This study supports prior conclusions regarding water hammer; however, some aspects that could impact safety and were Identified in the study have not been previously emphasized.

11 >D WORDS/TXSCRF' TORS (Ust words or poveses that w4ii assist researchers in laating the report) 13, AVAllA81UTY $1 ATEMENT water hammer events " """'m t

  • S W"N Unresolved Safety issuo

$'n'elassified USI A-1 (ms Repw hydrodynamic interactions Unitassified

, 1$ NUMBER OF PAGES )

16. PRICE

-\

Printed on recycled paper-Federal Recycling Program

-- yri<< e ,y p y -,,,.-p - , 9 ,e.. g - ,, -w _yp,. , ,. ,e,s._,. _.r,,,,. , .y,4 ,.w-..

,.W .ew- - ,

1[

l g

):

e

?

u

1. 4 1

p .t E

j . 4 I.

t Y, ^

3 5550?- CO' NOT' GNIHS AV

-?

.;gyngraciLRype90, .a N-RDf[f P}

r

- y.

101 +i v J IV C IN A I' 'l 'g 3;di0 t OFl

-CRN s /\ ly- ._ '

io315550'2I Y' d S

i

1

e ff *touq?% *

, UNITED STATES

!' 3 ,f, ( g j g .p NUCLEAR REGULATORY COMMISSION o, g WASHINoTON. D C. 20t65

4 /

9*sss e October 30, 1992 T0: RECIPIENTS

SUBJECT:

0FFICE FOR ANALYSIS AND EVALVATION Of OPERATIONAL DATA PUBLISHES POWER REACTOR EVENTS & ISSUES (NUREG/BR-0171)

Enclosed is the first issue of Power Reactor Events & 1ssues (PREI), NUREG/BR-0171, Vol. 1. This periodical replaces an earlier AE00 bi-monthly publication, Power Reactor Events, NUREG/BR-0051, that was discontinued in 1989. Our plans are to issue PREI semi-annually, with the capability of issuing special reports when events warrant a more timely issuance, in Spring of 1992, we sent the draft issue of PREI to the NRC program offices, regions and the industry-supported organizations for their comments on the potential usefulness of this publication. The highlights of this peer review were that such a publication would be beneficial to the plant staff provided it was not redundant to other NRC feedback documents, did not carry a regulatory tone, and if the information presented was factual and non-judgmental.

Each future issue of PREI will normally contain at least one article based on an AE0D study of U.S. reactor operating experience. We plan to.use this periodical to provide timely feedback of reactor operational experience, both domestic and foreign, without duplication of information contained in other NRC generic communications. Using a broad dissemination, our goal is to-feedback operating experience to the plant site management (i.e., operations, maintenance, and training organizations) as well as to the industry-supported organizations and the NRC staff.

In order to make the future issues of this periodical of interest to you, please use the response form attached to this letter to provide your comments regarding the usefulness of the information presented in PREI, We would also welcome suggestions for topics you would like to see covered in the future issues.

%M Thomas M. Novak, Director Division of Safety Programs Office for Analysis and Evaluation of Operational Data

Enclosure:

As stated l

j

e ht n

o t

s n

e m

m o

c r

e ht o

y . .

ne Alc i t

r s a

s' e e mr d ad e N A' h

t m

m o r

o r F f

g?

i ne sl s i c

i mtra s

a w

t a

h W

2 1

1 n 9 o i s Bs s iht B m' i

e N m s

u?n MopC ui o o y5 ot ya t r5 t m S o5 t 0 l

al a2 h r i go i f min MugC e w i RD o h r n f

l t apl et ao ic g

r n Tui i

Nh s j

aS a r

o RUW _

f  :

)

s s o

( c t c ips s t

i c ole n t

r t

i c e a dt e r m e t s

a m r

u ee o c _

t gr -

m a gu d e e u t n

Sf u I

t F

S e

-

Retrieved from "https://kanterella.com/w/index.php?title=ML20125D276&oldid=2919371"