ML18051B325
| ML18051B325 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs, Salem, Mcguire, Palisades, Maine Yankee, McGuire, 05000000, 05000319 |
| Issue date: | 03/18/1985 |
| From: | Ippolito T NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | Vollmer R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| Shared Package | |
| ML18051B326 | List: |
| References | |
| NUDOCS 8504020016 | |
| Download: ML18051B325 (10) | |
Text
~'*,\\*
March 18, 1985 Di strij?uti on~
/
AEOD SF ROAB CF RAOB SF Eleeds SRubin KSeyfrit Tippolito CHeltemes
~lb MEMORANDUM FOR:
Richard H. Vollmer, Deputy Director Office of Inspection and Enforcement Bf/t,,
FROM:
SUBJECT:
Thomas A. Ippolito, Deputy Director Office for Analysis and Evaluation of Operational Data TRANSMITTAL OF TECHNICAL REVIEW REPORT, COMPARATIVE ANALYSIS OF RECENT FEEDLINE WATER HAMMER EVENTS AT MAINE YANKEE, CALVERT CLIFFS, SALEM, MCGUIRE, ANO PALISADES Enclosed is an AEOD technical review concerning the recent feedline wa-ter hammer events. This review concludes that procedural inadequacies were the underlying cause of the events. It is suggested that IE consider issuing an Infonnation Notice that emphasizes the need for procedures that cover the restart of pumps following shutdowns for periods long.
enough to pennit draining of piping or creation of voids that could lead to water hammer.
If you have questions concerning this report or this suggestion, please contact Eric Leeds on extension 24445.
Enclosure:
As Stated cc w/enclosure:
Original signed b~
Thomas A. lppolitQ Thomas A. Ippolito, Deputy Director Office for Analysis and Evaluation of Operational Data
~1-317 SIB 255 3b~
- 3 -'
E. Jordan, IE E. Rossi, IE R. Baer, IE I
~BH4020016 e5031a t
s ADOC1i< 05000255 I PDR I I --r.
J 272 3 If
- -, : _.J
\\oFFICE~.~.Q~W.
....... ~ht.~R~.~~***.. ~.l.R9A.~.......... P.Q/~.rnD..J/*... O~io...................................................
su'.R:::;:.~.~~. i?.~~L'.. -~*~~b.i.n.~i~** :~~eY.f.r.:-~****_-** ~~~P.O.~~~.Q._... _.. **6P,?;~*~*:*** ********************* ********************
.. ~/.... I............. /..6. /............... /..k. I....... *.*....... <..........................................................................
NRCFoR.M31a1101so1NRCM0240 OFFICIAL RECORD cbPv
- u.s.GPo 1983-400-247
,,..:...,*-~:*.£.:~... 1.-_-.::..-~;~.7"*":.-;.;.~u.¥*¥-;,;,.wz~-.;..;:.._~.i--;-..;.:;"4i;...,;i..*fz.:.~~~-:..~::;.-~-;;;~.~c.;:::.~~?:12:e*:-::..~;z..~~:::**c';..,:ci.*::-:.;~*.::,-:.~~~::'.~~.~+/--!i°?i=;.~~~~i;~~~~:Ylii~::J:'d..~t~=I::l7.;;:..4S;...-t.~-.:.1e!~~l?#AW.,.~
. t::i
~~1:1**
~~
Mar'cn 18, 1985 MEMORANDUM FOR:
Karl V. Seyfrft, Chief Reactor Operations Analysis Branch Office for Analysis and Evaluation of Operational Data DISTRJ.2UTION:
PDR v AEOD CF AEOD SF ROAB SF ROAB CF Eleeds SRubin KSeyfrit Tippol ito CHeltemes THRU:
Stuart O. Rubin, Chief Reactor Systems Section 1 AEOD/T502 FROM:
SUBJECT:
Reactor Operations Analysis Branch Office for Analysis and Evaluation of Operational Data Eric J. Leeds~ Engineer Reactor Systems Section 1 Reactor Operations Analysis Branch Office for Analysis and Evaluation of Operational Data COMPARATIVE ANALYSIS OF RECENT FEEDLINE WATER HAMMER EVENTS AT MAINE YANKEE, CALVERT CLIFFS,.
SALEM, MCGUIRE, ANO PALISADES Enclosed is a-:technic-al review concerning the recent feedline water hammer events. This review concludes that procedural inadequacies were the underlying cause of the events. It is recommended that these events be summarized in the PRE *to re-emphasize the importance of operating procedures in preventing water hammer events. It f s suggested that IE consider issuing an Information Notice that empha-sizes the need for procedures that cover the restart of pumps following shutdowns for periods long enough to pennit draining of piping or creation of voids that could lead to water hammer.
Enclosure:
As Stated I ;)l Eric J. Leeds, Engineer, Reactor Systems Section 1 Reactor Operations Analysis Branch Office for Analysis and Evaluation of Operational Data orne< *. ~R~.~.... ~k ;RQ~~..... ~.~.:.~QA~.JJ..... QR;,~~~.............................................................. '
SURNAME. J.~~...... ~.... :... ~R~.~.i.n..~...
~?.~Y.fX.i.t......... HRP.Q........................................... ;........................... *..
DATE*- *.... -.Q.?.~...... ?L~t.~.~....... }{.;0./.?.~......... ~(~/.?.~................................................................ ~:.. *I NRCFORM3181101801NRCM0240 OFFICIAL RECORD COPY
- U.S.GPO 1983-400-2,4:(
-"";".._:=.~-;~7:_-:'"".:"r: \\: _ :_-:-::--- :.:!* :_:..:..~.::: :.:-.:..\\.:r;,;....:-:-- :::.- ~.~:.. ;.:.. ::-..... :...=-.;;_,_:..:_..~~:.*. :.;:.;:;;...=--:=.,,-.';:;::;:;:-;;:_~-~::!:.C2:;:.;-;::-* ~ ~-~_..... **~~::*.i."'""'\\;;"'::r~i-"":~~~~=~-!~~~~*E;:~~~~z*~-r-~j.lzg;~..,;~~-~
. n. '
AEOD TECHNICAL REVIEW*
UNIT*
DOCKET NOS. :
LICENSEE:,_
Multiple PWRs Multiple PWRs Multiple PWRs TR REPORT N_O. :AE~s~s-so2 DATE: March 18, EVALUATOR/CONTACT:
E. J. Leeds
SUBJECT:
COMPARATIVE ANALYSIS OF RECENT FEEDLINE WATER HAMMER EVENTS AT MAINE YANKEE, CALVERT CLIFFS, SALEM, McGUIRE AND PALISADES EVENT DATE:.
Multiple
SUMMARY
A review was conducted to detennine if a common root cause contributed to the recent PWR feedline water hammer events. Several of the events reviewed indicate that feedwater system procedural inadequacies were the underlying cause of the water hammer occurences. Specific hardware and operating pro-
. cedures were previously identified during the staff evaluation of Unresolved Safety Issue A-1, Water Hammer, as important measures needed to reduce the incidence of severe water hammer in PWR feedwater and steam generator systems.
This technical review concludes that the recommended actions formulated by the staff are adequate to prevent or mitigate the damage caused by water hammer and that there. is limited safety significance involved in these water hammer events.
8504020025 850318 1
1 l.*
PDR ADOCK 05000255
___ s_ ___... __.
_P_PR _____ _
- This document supports ongoing AEOD and NRC activities and does not represent the posi.tion or requirements of the responsible NRC program office.
. "-) DISCUSSION In the study of Unresolved Safety Issue (US!) A-1, Water Hammer, the staff found that 40 of the 67 reported PWR water hammer events involved either the feedwater or the steam generator system. The staff evaluation report for US! A-1 concluded that operational experience was al so a factor in the frequency of water hammer events. Approximately.one-half of the 150 reported water hammer events since 1969 have occurred during preopera-tional testing or during the first year of commercial operation. Because design and operating changes have been instituted throughout the industry to prevent water hammers, the frequency of water hammer events in general, and feedline water hammer events' in particular, have declined. For PWRs, after a peak of 12 reported water hammer events of all types in 1976, the number of events declined to four in 1980 and two in 1981.
However, since the completion of US! A-1, five feedline water hammer events have been reported.
At the completion of US! A-1 in December 1982, the staff's generic study concluded that the total elimination of water hammers was not possible because of inherent design and operational conditions of nuclear power plant systems.
The purpose of this technical review is to *determine if:
(1) The lessons learned from the staff's generic review address the cause of these recent events and; (2) if a common root cause contributed to the five feedline water hammer events. A discussion of the circumstances and evaluation of each of the five events follows:
Operating Experience
- 1.
On January 25, 1983, the Maine Yankee plant experienced a reactor trip from full power.
Prior to the reactor trip, the plant had been operating with the lODi capacity turbine driven main feedwater pump.
The steam drive*n main feedwater pump had been installed during the previous outage to provide diverse feedwater capab.ility. Prior to the installation of the steam driven pump, the motive force was provided by electrically driven feed pumps which continued operation following a reactor/turbine trip.
In this event both motor driven main feedwater pumps were unavail-able.
The main turbine tripped as designed following the reactor trip.
The turbine driven main feedwater pump, driven by high pressure extraction st~am, became unavailable when the main turbine tripped, causing Maine Yankee's first reactor trip with total loss of feedwater.
When auxiliary feedwater (AFW) flow was manually initiated on low steam generator (SG) level, a loud noise was heard in the plant machine shop located directly beneath the main feedwater lines. Shortly after initiating AFW flow, operators noticed a rise in containment humidity, an increase in contain~
ment sump level, and the actuation of a containment fire detection alarm.
The visual inspection made upon containment entry revealed substantial leakage from the main feedwater line in the vicinity of the No. 2 SG inlet nozzle. A plant cooldown was initiated and the licensee declared an unusual event.
- When the main turbine tripped and the turbine driven feedwater pump became unavailable, the plant experienced a loss of nonnal feedwater flow to the SGs.
The reduction in steam demand after the turbine trip caused SG level to drop due to void collapse.
SG level fell below the level of the feedring, allowing the water in the feedring to drain through the seventy-six 1-inch diameter discharge nozzles
- in the bottom of the ring. A water hammer occurred when the relatively
- cool AFW flow was manually initiated on low SG level causing the steam in the partially filled feedring to conden~e. The water hammer caused a pre-existing crack in the main feedwater line to the No. 2 SG to become a through-wall crack. During the analysis of the Maine Yankee event pre-existing cracks were discovered in the lines feeding two of the thre~ SGs.
These cracks were caused by fatigue from excessive the nna 1 eye 1i ng
- Because the Maine Yankee plant had never experienced a SG water hammer event in its previous operating history, the SG feedrings had never been modified with the 1J-tube 1 hardware fix recommended in the US! A-1 generic report to prevent SG water hammer.
Prior to the addition of the 100%
capacity turbine driven main feedwater pump, the feedwater system con-sisted of two motor driven main feedwater pumps and the plant had never experienced a reactor trip from 100% power without main feedwater avail-able. - As a result of the water hammer event, the licensee sealed the seventy six 1-inch discharge nozzles on the bottom of the feedrings and installed twenty eight 3-inch J-tubes on the top of the feedrings.
The J.:..tubes will extend the time required for the feedring to drain once it is uncovered.
Automatic initiation of auxiliary feedwater along with operational instructions and procedures were also instituted to reduce the probability of another water hammer occurrence and to reduce the thennal cycling of the main feedwater lines to prevent fatigue cracking.
- 2.
On April 22, 1984, the Calvert Cliffs Unit 2 plant was operating in Mode 3 (1400 psig, 406° F).
The operators were in the process of realigning the main feedwater system following the testing of a newly installed motor driven auxiliary feedwater pump.
During the auxiliary feedwater pump* test, SG level had been lowered below the feedring, main feedwa ter fl ow had been secured, and the main feedwa ter i so 1 ati on valve was shut due to leakage past the feedwater regulating valve. Approxi-mately thirty minutes later, the main feedwater isolation valve was reopened to feed the SG using the main feedwater system.
Upon opening the main feedwater isolation valve a severe water hammer occurred.
The water hammer damaged a main feedwater isolation valve, a main feedwater regulating valve and a main feedwater regulating bypass valve.
,* During the thirty minutes that SG water level remained below the level of the feedring, water in the feedring drained through the slip fit joint where the feedring enters the thennal sleeve. Apparently, leakage past the feedwater regulating valve was significant enough that when the main
- feedwater isolation valve was opened a severe water hammer occurred.
According to written procedure, the AFW system is required to be used whenever SG water level is below - 26 inches, which corresponds to a con-servative level for the top of the feedring.
At Calvert Cliffs, the AFW pumps feed through a separate nozzle located lower in the SG than the main feedring.
Had AFW flow been initiated as required by procedure, the main feedring would have been reflooded prior to restoring main feed-water flow and the water hammer event would not have occurred. Thus, the event could be attributed to a failure to follow established procedures.
The licensee had previously installed J-tubes on the SG feedrings in late 1978 for Unit 2 and mid-1979 for Unit 1. The licensee also implemented the other modifications designed to prevent or mitigate the effects of SG water hammer as described in Branch Technical Posi-tion (BTP) ASB 10-2.
However, the modification did not prevent the SG water hammer from occurring. It appears that the J-tube modification only extended the time required for the feedring to drain once ~ncovered.
The feedring apparently drained through the fitting at the thernJal sleeve.
Thus, the J-tube modification will not prevent a SG water hammer in all circumstances.
The licensee's actions to prevent recurrence of this event included strengthening the procedural precautions already in place (to prevent water hammer) and increasing operator awareness of water hammer phenomena through training.
- 3.
On April 6, 1984, the Salem Unit 2 plant was in hot shutdown (Mode 3) and in.the process of stroke time testing each of the feedwater regulat-ing valves and bypass valves. Testing had been satisfactorily completed on two sets of valves.
SG pressure was approximately 1000 psig. With the feedwater system secured, one operating condensate pump was circulating condensate to provide secondary system cleanup and maintaining approximately 500 psig in the main feedwater header.
When the feedwater regulating valve in loop 3 was opened for its stroke time test a loud rumbling noise was heard.
Upon hearing the noise, the maintenance personnel closed the feedwater regulating valve and the rumbling noise stopped.
The transient lasted for approximately 20 - 30 seconds.
An inspection revealed that damage to the feedwater regulating valve and the feedwater piping support systems had occurred.
At Salem-2, there are two valves in the main feedwater line between each feedwater regulating valve and its associated SG; a normally
- l open isolation valve and a motor operated stop checkvalve. The check valve prevents reverse flow from the SG whenever.the main feedwater pumps are secured. It was detennined that the stop chetk valve in loop 3 failed to close against the 1000 psig loop 3 SG pressure. The licensee believes that the check valve failed to close because of crud buildup. Magnetite was found in the bowl section of the valve in the post-event inspection.
When the feedwater regulating valve in loop 3 was opened for its stroke time test, reverse flow occurred due to the high differential pressure across the valve, 1000 psig on the downstream (SG} side and*SOO psig on the upstream (condensate pump} side. This reverse flow slammed shut the stuck open check valve generating a water hammer compression wave upstream of the valve and a rarefaction wave downstream into the SG.
The check valve and all damaged equipment were repaired.
On the subsequent plant startup on April 23, 1984, a turbine trip and reactor trip occurred due to high-high level in the loop 3 SG.
The licensee attributed the event to sluggish response of the feedwater level control system at low power operation and minor binding of the feedwater control bypass valve. After taking corrective action, the licensee conducted another startup on April 27, which again resulted.
in a reactor trip due to high-high level in the loop 3 SG.
A thorough investigation' conducted by the licensee revealed that the loop 3 feed-water flow nozzle had been displaced approximately 24 inches from its original location. The feedwater flow nozzle provides the pressure drop.
for flow measurements used for indication, level control and protection signals. Apparently, the pins holding the nozzle in place had been broken.
The licensee attributed this to the earlier April 6, feedline water hammer event.
To prevent a similar occurrence, the licensee modified the plant test procedures to require that the stop check valves be locked in the* closed position using the valve motor operators, prior to opening the main feed-water regulating valves when perfonning the stroke time test. Also, a review of the feedwater system's piping support was to be perfonned to detennine whether modifications were necessary to prevent similar water hammer damage.
- 4.
On August 21, 1984, McGuire Unit 1 was operating at lOOi power and supply-ing steam to the Unit 2 main feedwater pump turbines in preparation for a restart of Unit 2.
The McGuire switch yard computer malfunctioned, opening multiple 230 KV switch yard breakers resulting in a loss of offsite power and a trip of the Unit 1 reactor.
The Unit 1 reactor coolant system cool-
. l
- down was greater than expected after the trip because Unit 1 was supplying steam to Unit 2.
To prevent an excessive cool down and depressurization of the RCS, the main steam isolation valves (MSIVs} were manually closed and the steam generator power operated relief valves were used to control steam line pressure and remove decay heat from the reactor. The loss of offsite power and subsequent closing of the MSIVs resulted in a loss of all condensate and feedwater pumps.
When power was restored and the pumps restarted, water hammers occurred in the condensate and feedwater systems.
The licensee detennined that following the trip, the feedwater pump recir-
~ulation valves failed as designed to the full open position. Once the condensate and feedwater pumps were lost, the full open recirculation valves allowed water to drain back to the condenser creating voids in the condensate system and draining the section of piping between the main feedwater pumps and their discharge check valves.
The depressurization
- of the feedlines caused by the tripped feedwater pumps also allowed the*
400° F feedwater to fl ash to steam.
When power was restored and* the pumps restarted, the steam and water flow in the voided sections of piping caused water hammers in the condensate and feedwater systems.
Inspection revealed that the water hammers caused no damage to any pipe hangers although minor damage to the condensate booster pump discharge pressure gauges occurred.
The licensee is modifying the condensate and feedwater procedures to minimize the possibility of a water hammer in these systems following a loss of offsite power.
- s.
On March 13, 1984, at the Palisades plant, the licensee reported that an SG inspection revealed extensive damage to the AFW feedring piping in both SGs.
An inspection conducted just prior to the discovery of the internal SG damage also revealed several loose and damaged AFW piping hangers.
Although no specific event was identified as causing the damage, it appeared that the damage was the result of a water hammer.
However, subsequent review and analysis detennined that the internal SG damage was _caused by flow induced vibration resulting in cyclic fatigue failure *. The internal SG damage is thought to have contributed to the AFW piping hanger damage by allowing the system to become susceptible to water hammer events.
The licensee is planning to completely remove the present AFW feedring piping and replace it with an entirely different design to prevent a similar occurrence. Since the event at Palisades was not a water hammer occurrence, it has not been examined further in this report.
- . l
- Of the five events, only the water hammer event at McGuire Unit 1 may be of potential generic concern.
At this unit, the feedwater pump recirculation valves are in a line which tees off the main feedwater lines between the main feedwater pumps and the feedwater pump discharge check valves. This configuration allowed water to drain to the condenser creating voids in the condensate system and draining the short section of piping between the main feedwater pumps and the discharge check valves on a loss of offsite power.
If the feedwater pump recirculation valves were located downstream of the feedwater pump discharge check valves, water in the feedwater piping downstream of the main feedwater pump discharge check valves might drain to the condenser when the recircula-tion valves fail open after a loss of offsite power.
The resultant voiding of the main feedwater piping could conceivably contribute to a water hammer event if the piping was not filled prior to starting the main feedwater pumps.
To determine if any plants had this postu-lated secondary piping configuration, we reviewed the feedwater piping at four PWRs:
Sequoyah, Catawba Unit 1~ Salem Unit 1, and Palisades.
None of these plants had a feedwater system that would have allowed a large portion of the feedwater piping to drain to the condenser after a loss of offsite power, i.e., their recirculation piping was between the main feedwater pumps and the pump check valves.
Findings The root cause of each water hammer event can be summarized as follows:
- 1.
At Main Yankee, the SG water hammer was traced to the draining of the feedrings following the loss of main feedwater flow.
The installation of J-tubes on the feedrings, automatic initiation of AFW flow and operating procedure modifications were made to prevent recurrence.
- 2.
The water hammer. event at Calvert Cliffs Unit 2 occurred when operators failed to observe the written operating procedure to feed the SGs using the AFW system whenever SG water level is below -26 inches, a conserva-tively chosen level for the top of the feedring.
- 3.
At Salem Unit 2, personnel following inadequate test procedures caused a water hammer which resulted in two reactor trips.
- 4.
At McGuire Unit 1, a loss of offsite power resulted in the voiding of sections of condensate and feedwater system piping. Operating procedures allowed operators to restart secondary system pumps when power was re-stored which caused condensate and feedwater system water hammers.
In all four events, plant procedures failed to prevent water hammers and were subsequently m_odi fi ed to prevent future recurrence.
.. "J.
- CONCLUSION The USI A-1 staff's study concluded that the total elimination of water hammer is not feasible due to design and operational conditions wherein steam, water and voids can co-exist within a system.
However, it appears that procedural inadequacies continue to contribute to the frequency of water hammer events. This technical review concludes that the staff's recommended actions are adequate to prevent or mitigate the damage caused by water hammer and that there is limited safety significance involved in these water hammer events.
Suggestion*
It is suggested that IE consider issuing an Information Notice that emphasizes the need for procedures that cover the restart of pumps following shutdowns for periods long enough to permit draining of piping or creation of voids that could lead to water hammer.