ML20209B412

Operational Experience Involving Turbine Overspeed Trips, Case Study Rept

ML20209B412
Person / Time
Site:	Davis Besse
Issue date:	08/31/1986
From:	Caroline Hsu NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML20209B180	List: ML20209B175 ML20209B232 ML20209B275 ML20209B342 ML20209B412 ML20212Q421
References
TASK-AE, TASK-C602 AEOD-C602, NUDOCS 8609080229
Download: ML20209B412 (83)

Text

. ._.

CASE STUDY REPORT AE0D/C602 OPERATIONAL EXPERIENCE INVOLVING >

TURBINE OVERSPEED TRIPS August 1986 Prepared by: Chuck Hsu Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulatory Commission l

l 1

Note: This report documents results of a study completed to date by the Office for Analysis and Evaluation of Operational Data with regard to operational situations. The findings, conclusions and recommendations contained in this report are provided in support of other ongoing NPC activities and do not represent the position or requirements of the responsible program offices of the Nuclear Regulatory Commission.

J 8609090229 860908 6 PDR ADOCK 0500 t

4 TABLE OF CONTENTS Pace EXECUTIVE

SUMMARY

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

1.0 INTRODUCTION

. . . . . . . . . . . . . . . . . . . . . . . . g . 3 2.0 DISCUSSION .......................... 4 2.1 Slow Governor Response During Quick Start ........ 4 2.2 Condensation in Steam Line .............. . 11 2.3 Errors in Reset Procedure ................ 13 2.4 Opening of Supply Valve With Trip Valve Open ...... 14 2.5 Stop Valve Stroking Time Too Fast ............ 17 2.6 Turbine Spinning . . . . . . . . . . . . . . . . . . . . . 18 2.7 Pump Air or Vapor Binding .............. 19 2.8 Trip and Reset Problems of Trip Valves and Trip Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . 19 2.9 Worn or Loosened Nuts and Linkages . . . . . . . . . . . . 23 2.10 Misadjustment, Miscalibration or Drifting Controller ....................... 24 2.11 011 Leak and Contamination ............... 25 2.12 Electrical Component Failures ............. 26 2.13 Summary ........................ 27 3.0 FINDINGS AND CONCLUSIONS ................... 30 4.0 RECOMMENDATIONS . . . . . . . . . . . . . . . . . . . . . . . . 33

5.0 REFERENCES

.......................... 35

4 LIST OF TABLES Page Table A Number of AFW Turbine Overspeed Trip -

Events from January 1972 through 5

September 1985 ................

Table B Number of HPCI and RCIC Turbine Overspeed Trip Events from January 1972 through 6

September 1985 . . . . . . . . . . . . . . . . .

Turbine Overspeed Trip Events ......... 7 Table C Table D Number of Turbine Overspeed Trips in Terms of Major Attributed Causes and Calendar 8

Years .....................

Table 1 Events of Slow Governor Reponse During Quick 36 Start .....................

41 Table 2 Events of Condensation in Steam Line . . . . . .

...... 45 Table 3 Events of Errors in Reset Procedure Table 4 Events of Opening of Supply Valve With Trip 46 Valve Open ..................

Events of Stop Valve Stroking Time Too Fast .. 48 Table 5 Events of Turbine Spinning . . . . . . . . . . . 49 Table 6 Events of Pump Air or Vapor Binding ...... 51 Table 7 Table 8 Events of Trip and Reset Problems of Trip Valve and Overspeed Trip Mechanism . . . . . . . . . . 52 Events of Worn or Loosened Nuts and Linkages . . E6 Table 9 Table 10 Events of Misadjustment, Miscalibration or 61 Drif ting Controller ..............

Events of Hydraulic 011 Leak or Contamination . 67 Table 11 Table 12 Events of Governar Electrical Component 71 Failures . . . . . . . . . . . . . . . . . . . .

........ 75 Table 13 Events of Unknown Failure Mode l

d I__

LIST OF FIGURES Page Figure 1 - Turbine EG Governor Control System ......... 10 -

Figure 2 Schematic Diagram of PG-PL Governor . . . . . . . . . 15

! t Figure 3 T&T Valve and Overspeed Trip Mechanism (OTM) .... 21 i i s

t i

f l 5 4

4 1

I l

l l

4

, , . - . ,, --.-,- ..- - , ,. n ...- - ,.,,,_w . , . , . . . . . ~ _ . . _ , . , - _ . _ , . . _ , - , . , , - - . . _ - . . , - - - . , - _ _ . _ _ , - . . - - . . ., . - _ , ,, . _ _ _ , - ,

4 EXECUTIVE

SUMMARY

A study was performed to primarily review and evaluate past operating experience involving overspeed trips of PWR auxiliary feedwater (AFW) turbine-driven pumps. In addition, the study was extended to include a review of turbine overspeed trip operating experience on BWR high pressure-coolant injection (HPCI) and reactor core isolation cooling (RCIC) turbinc-driven pumps. The study identifies major attributed causes for these tript.

Recommendations for preventing or reducing the frequency of these turbine trips were also developed. The study was performed in response to action item 8(f) of the actions directed by the ED0 to respond to the NRC staff investigation of the June 9, 1985 event at Davis-Besse.

Almost all of the turbines in use in AFW, RCIC and HPCI systems at operating plants are manufactured by the Terry Corporation (Terry), Steam Turbine Division and are equipped with governor:; manufactured by the Woodward Governor Company (Woodward). The operational conditions for these three systers are also quite similar. Accordingly, although the study was prompted by concerns about the effect of turbine overspeed trips on the reliability of PWR AFW systems, the review also included RCIC and HPCI A total of events to aidinvolving 128 events in identifying causes for AFW turbine overspeed trips.

overspeed trips of steam turbines associated with AFW, HPCI and RCIC systems were reviewed. These events occurred from January 1972 to September 1985 and were identified primarily from Licensee Event Reports (LERs).

Review of these events indicates that the dominant attributed causes of AFW turbine overspeed trips are speed control problems asscciated with governors, and trip and reset problems associated with trip valves and overspeed trip mechanisms. These problems are primarily the result of inadequate performance by plant personnel, inadequate procedures, and insufficient system design considerations. The governor speed control problems are (1) slow response of the Woodward Model EG governor during quick startup, (2) entrapped oil in the speed setting cylinder of the Woodward Model PG-PL governor, (3) incorrect governor setting and (4) water induction into the turbine. The trip and reset problems stem from the complexity of reset operations and a lack of trip position indication. Although some of these concerns about trip and reset problems have been previously identified in IE Information Notice 84-66, these problems continue to exist.

To prcvent or reduce the frequency of these turbine overspeed Specifically, thetrip problems, report provides several recommendations were also developed.

the following recommendations:

1. Licensees of FWR plants utilizing a Woodward Model EG governor for the AFW turbine should be requested to consider implementing stean bypass modifications to the AFW system to improve the turbine reliability during startup.

2. In view of the number of turbine overspeed trips resulting from incorrect governor speed settings, licensee.s should be asked to review the adequacy of the existing vendor-supplied calibration procedures used for the control system of AFW turbines.

4

3. To assure that condensate in the steam supply line of the AFW turbine is removed before reaching the turbine, all licensees should be required to review and verify that: (1) the steam supply line steam trap operability administrative controls are adequate; (2) the capacity of the steam traps is sufficient to remove instantaneous, rapid condensation resulting from heating the cold steam line during turbine startup; and (3) the steam supply line piping is in a configuration that will minimize the formation of condensate during a turbine cold start. 5

4. To minimize trip and reset problems involving the trip and throttle (T&T) valves and the overspeed trip mechanisms (OTMs), licensees should be required to review the adequacy of the existing procedural instructions and the training programs regarding reset operation of T&T valves and 0TMs. Local indication of the position of T&T valves and OTMs, as well as control room indication for operability / availability of these devices, should be verified and/or provided as appropriate.

5. IE should issue an information notice to alert licensees of operating reactors of the findings that led to the above recommendations.

Specifically, the information notice should address events involving turbine overspeed trips resulting from entrapped oil in the governor speed setting cylinder, including the conditions which resulted in the oil becoming entrapped. The information notice, to the extent possible, should point out that the problem could be avoided by: (1) establishing improved administrative controls to bleed off the entrapped oil, (2) installing a remotely controllable dump valve in the governor hydraulic circuit, or (3) providing indications that would annunciate in the control room when a turbine is spinning.

As a result of this study it is requested that NRR and IE consider these recommendations in the ongoing programs to improve AFW system reliability (particularly in response to action items directed by the ED0 in connection with NUREG-1154 (Reference 3)).

1 4

1.0 INTRODUCTION

The auxiliary feedwater (AFW) system of pressurized water reactors (PWRs) is required to provide feedwater to the steam generator to remove decay heat following a loss of main feedwater or following a loss of ac power. A total increase the probability loss of the AFW system when needed will significantly(HPCI) of core damage. The high pressure coolant injection system and the reactor core isolation cooling (RCIC) system for BWRs are installed to mitigate a small break loss of coolant accident and a total loss of normal fledwater, respectively. In view of the importance of these systems, an evaluation was conducted of past operating experience involving overspeed trips of steam turbines in the AFW system of PWRs and the HPCI and RCIC systems of BWRs. The evaluation includes the identification of causes for these overspeed trips.

The overspeed trips of AFW turbines were found to be somewhat widespread and one of the major causes for loss of operability or avai_ lability of AFW systems. A typical AFW system usually consists of one turbine-driven pump and one or two motor-driven pumps. In the event of loss of all ac power or failure of the trains containing the motor-driven pumps, a turbine overspeed trip would result in the entire AFW system being unavailable. Moreover, there were at least three units where the AFW systems relied solely on turbine-driven pumps. These units were Davis-Besse and Turkey Point 3 and 4. At Turkey Point 3 and 4, there are three turbine-driven pumps shared by two units. Davis-Besse has recently installed a third (full flow) motor-driven pump, however, for diversity.

Both of these plants have experienced a loss of all AFW as a result of coincident overspeed trips of their AFW turbines.

Almost all of the turbines which are used in AFW, RCIC, and HPCI systems at operating plants are manufactured by the Terry Corporation and are equipped with Woodward governors. The service conditions for these three systems are quite similar ard, as such, review of HPCI and RCIC events can aid in identifying of causes for overspeed trips on AFW turbines. In addition, General Electric Company, as the vendor for boiling water reactor (BWR) units, has performed a number of studies for overspeed trip problems involving the HPCI and PCIC turbines, and has provided recommendations based on the results of these studies. Some of these studies and recommendations, which may be applicable to AFW turbines, are identified in this report.

This study identified a total of 128 events involving overspeed trips of steam turbines associated with the AFW, HPCI and RCIC systems. The events occurred at numerous operating nuclear power plants during the period from January 1972 to September 1985. Information relating to these events was obtained from Licensee Event Reports (LERs), licensee failure analysis reports, regional inspection reports, and other relevant documents (Ref. 1). The LERs were identified by a search of the Sequence Coding and Search System (SCSS) and the RECON database system. Other events were identified in a review and search of information obtained from the Nuclear Plant Reliability Data System (NPPDS).

This study was initiated as a response to Action Item 8(f) of the NRC staff actions resulting from the investigation of the June 9,1985 event at i Davis-Besse (Refs. 2 and 3). Near the completion of the preliminary study, IE issued Information Notice 86-14, "PWR Auxiliary Feedwater Pump Turbine Control Problems," as a response to a separate staff action item. The IE l information notice addressed a few, but not all of the overspeed trip cause categories identified in this report.

' s 2.0 DISCUSSION A total of 128 events involving turbine overspeed trips are covered in this .

study. Among these, 61 events occurred at pressurized water reactor (PWR) plants and involved AFW turbines and 67 events occurred at boiling water reactor (BWR) plants in which 47 involved RCIC turbines and 20 involved HPCI turbines. Tables A and B present the number of events identified for each year from January 1972 through September 1985 for PWR and BWR plants, respectively.

A review of these events shows that these turbine trips can be grouped into 12 known and one unknown cause categories. The known categories are: (1) slow governor response during quick start; (2) condensation in the steam line; (3) errors in reset procedure; (4) opening the steam supply valve with the trip valve open; (5) stop valve stroking time too fast; (6) turbine spinning; (7) pump air or vapor binding; (8) trip and reset problems associated with trip valves and overspeed trip mechanisms; (9) worn or loosened nuts and linkages; (10) misadjustment, miscalibration or drifting of controllers; (11) hydraulic oil leak and contamination; and (12) electrical component failures. Turbine overspeed trips due to the stop valve stroking time problem, identified in category (5), are unique to the HPCI turbines in BWRs. The 13 categories are presented in Table C along with the related systems and the conditions under which the turbine overspeed trips occurred. The three trip conditions identified are: on demand operation, surveillance testing, and equipment maintenance activity. In addition, the number of overspeed trips for these cause categories reported each year from January 1972 through September 1985 are presented in Table D.

Based on the identified attributed cause categories, a brief description of the 128 events involving turbine overspeed trips are listed in Tables 1 through 13.

Unless otherwise noted in these tables, the turbines and governors involved are Terry turbines and Woodward governors. The types of Woodward governors are also listed in these tables. Each of the 13 categories are discussed separately and in detail in the following subsections. Findings and conclusions associated with each cause are provided, where appropriate.

2.1 Slow Governor Response During Quick Start In the normal standby readiness state, the steam turbine's trip-and-throttle (T&T) valve (which closes on an overspeed trip) and governor valve (which regulates steam flow to control the speed of the turbine) are open. The turbine is started by opening the steam inlet valve located upstream of the T&T and governor valves. The governor and governor valve will start to respond only when the hydraulic control oil pressure has been established by the turbine speed. Initially, the governor will see a speed error signal between the actual speed and the set speed, and will call for opening of the already open governor valve. Once the turbine reaches and begins to exceed the set speed, then the speed error signal causes the governor valve to begin to close.

Only then will the turbine be fully under the governor's control. If the governor and the governor valve do not respond fast enough on turbine startup, the steam flow to the turbine will likely cause the turbine speed to overshoot the set speed. If the overspeed is high enough, the turbine may reach the

l_ 1 t

a 2 13 2 2 3 11 4 3 2 21 13 1 2 1 2 1 3 1 2 1 1 1 3 1 6 o 1 T

1 1 1 6 5 1 8 2 9 _

1 1 3 4 1 I

8 9

1 1 1 6 3 1 1 11 8

9 s 1 t

n 1 4

2 2 e5 8 1 v8 E9 9 1 1 p

i r 2 9

re 1 8 2 1 2 1 1 Tb 9 m 1 de et 5 ep pe 0 1 1 1 2 1 1 1 1 sS 8 1 6 r 9 eh 1 vg r Ou a 7 o e 9 1 3 1 1 1 er Y 7 9

nh it r 1 b a r2 d 2 u7 n 8 1 1

l9 1 e 7 l 9 W a 1 Fy C Ar 4 a 7 1 1 f u 7 2 ona 9 1

rJ 2 e 6 1 1 b m mo u r 7

9 1

Nf A 5 7

e 9 l 1 b 3 a 4 2 1, T 7 9

1 3

7 9

1 2

7 9

1 1 2 d d 3 3

1 2 n n 2 3 t r 1 1 2 a a n e r 2 e 2 h l l o 2 i m k e a a ec s s c e e 1 e om e 1 2 v k s e n n s d a I I S e r r P u e i os n n e re f f h i S r s s P, o e1 2 n e e 1 2 o o a c a a CB o A Ada VeB i i o n n y u n y C 12 l s s s l - y y t t r r h m m O O o L a e .

nn a t n n a . se e s h h s on i i c e e u jk C f oo t n a a st Civrl lr ll t t r lil i a a n l l n n q . o r l o

r u . o ii o o a ao r r a a a a aet t

a k k T r r y . aa a i l

P AACDDFFMNNPPPPPR$SS5S$TTVWZZ

l t

a 434353 341 2'22243 5 1 3 3 7 6

4 o 1

T 5 H H H 8 2 9 R 1 R 2 R 2 4 4

8 9

1 R R R 2

3 H H 8 ,

9 1 P R 3 R 4 ,

2 8

9 R 1 R R s 2 t

n H H e 1 H H v 8 2 4 E 9 R

1 R P R p 3 i

r H T 0 H H 8 2 d 9 e5 1 R R R e8 2 p9 s1 r 9 H H H er 7 2 ve 9 R

Ob 1 R R R m 3 ee nt H ip r 8 H H be a 7 2 rS e 9 R R u Y 1 R R Th 2 4 g r Cu a H I o d 7 H Cr n 7 Rh e 9 R R t l 1 R d a 2 n2 C . .

a7 H cc 9 6 H H I1 7 2 tt C 9 ee P y 1 R R R  ;;

Hr 2 3 a ss

. pp fu ii on 5 rr a 7 tt rJ 9 R e 1 R R R R R b m 1 5 CI I C mo ur 1 1 CP Nf 4 H H RH 7

oo B

9 1 R R 2

R ww tt e

l b 3 H H H  ::

a 7 2 RH T 9 22 1 R R pp ii 2 H H rr 7 tt 9 CI 1

I C CP RH ee I 2 3 2 3 e nn 1 2 1 e OO y y y d 1 mm k r r r 1 2 l k o o s s a n r r r o c f o t t e e n a  ::

e e e k k n i l 1 l t t 1 i i n Y RH F F F c c r r u l o o t t a i i A 1 t G e e B B mi i i h t s s s s w w r a l c C C e n t n n n s s e e h pd l i h h r u o l  :

n w w w n n p n c z n a c c g d t d q m a e a o o o u u o a t t a S n a a l a a s r t t l r r r r r o u a i r a o e e i u u u e o o P B B B BB C DHFGLMPPPQQSV T N l

e

• Table C Turbine Overspeed Trip Events Attributed Cause System Condition of Event Equipment AFW RCIC HPCI On Demand Surveilla,nce Maintenance Category Operation Testing Activity AFW RCIC HPCI AFV RCIC HPCI

1. Slow Governor Response 3 5 3 During Quick Start 6 6 1

2. Condensate in the Steam 9 6 3 Line

3. Error in Reset 1

Procedures 1

4. Openino of Supply Valve With Trip Valve Open 2 2

5. Stop Valve Stroking 1 2 1 Time Too Fast 2

6. Turbine Spinning 3 1

7. Pump Vapor / Air Binding 1 1

8. Trip and Reset Problems 4 3 5 of Trip Valve and OTM 12

9. Worn or Loosened Nuts 5 5 4 4 5 and Linkages 4 9

10. Misadjustment, Mis-calibration or Drifting 5 2 1 3 6 4 Controller E 7 1

11. Oil Leak or Contamina- 6 13 4 tion 6 13 4

12. Electrical Component 2 4 2 1 4 6 Failures 7 8 5 2 2 2 1 6 3 1

13. Unknown 11 3 38 36 17 4 61 47 20 19 91 4 Total Events 128 33

,l l)')i)lj;)Jl .

ii l

a 2 9 1 2 2 3 1 2 8 7 '3 3 5 8 t 1 1 1 1 2 1 1 2 1

o T

5 2 8 2 1 1 1 1 2 1 3 1

9 1

4 2 5 8 1 2 9

1 3 1 1 8 5 1 3 1 1

9 1

2 6 8 1 2 1 2 9

f 1 o

2 3 4 3 1 2 6 ss 1 1 1

nr 1 8 r a 9 ee 1 TY nr ia 0

8 4 1 3 2 1 2 2 4 9 1

d 9 sn pe 1

il ra 9 1 3 2 1 1 2 TC 7 4 1 9

dd 1 en r ea p a 8 2 3 1 1 1 8 ss e 7 re Y 9 es 1 vu r Oa a C d 7 7 e n 7 2 1 1 3 nd e 9 i e l 1 bt a ru C ub 6 2 1 7 Ti 7 1 1 2 r 9 ft 1 ot A

r 5 2 0 1 2 5 er 7 1 1 bo 9 mj 1 ua NM 4 3 6 7 1 2 D 9 1

e l

b 3 1 3 a 7 2 T 9 1

2 1 1 7

9 1

e e g s v n f t n l n i o o ae g d - r n pt Vp n nm s e e sr O i i e s i l n ea y k gB l s de M l o Rt n pe o n bM eg o p y S i t pv r i r oT na ,rr m r r e ul tt n i r0 ek tot n o o ok n s Sa Ss n A P/ sns noi en oni n d oC g nc o e V a i /

eF pr te oL moC at l e ri ie R f t eu tn eo pv S o ev L ti aa a vQ aisLn ir ugrao i el ao ea rn p sl d st ok nc s C o ua nai i en Ga n d nT VT n V RV oa jri e mr rw e db tL at uo l s win da em r ei i /

o cnh p eb pp pn saif t cl n a t

u a loru n ot er r roet pi ti o muu rm r ro iiuia rt ri sli ol lannei k o CSDCSEPOWSTTPTTWNMcDOCEFU T j ll ll lll  !

i r

-9 overspeed trip setpoint. The 12 events involving overspeed trip due to slow governor response, with a brief description of the corrective actions taken and the type of governor involved, are listed in Table 1. Six plants identified in Table 1 have reported incidents of turbine overspeed trips attributed to a slow response of the governor valves. In all of the incidents, the overspeed trip occurred almost immediately following the start signal. The apparent cause for the slow governor responses was either governor valve binding or ths sensitivity of the speed control to a quick start transient. g The turbine overspeed trips at Zion 1 (1-2, AFW)* and LaSalle 1 (1-3, RCIC) were attributed to governor valve binding while closing in an unsuccessful attempt to contrcl turbine speed. The binding condition at LaSalle 1 resulted from slight wear and binding of the governor valve linkages. The binding condition at Zion 1 resulted from a sticky governor valve which appeared to be caused by a icng idle period. The slow governor response, resulting from the governor vahe binding, occurred on turbines equipped with Model EG and PG Woodward governors. The incidents at Arkansas 2 (1-1, AFW), Susquehanna 1 (1-4,1-5 and 1-6, RCIC), Palo Verde (1-7, AFW), and Grand Gulf 1 (1-8, RCIC) were related to the sensitivity of the governor speed control characteristics during a quick start transient. The sensitivity problem occurred only on turbines equipped with Model EG governors.

The turbines utilizing the Model EG governors are brought to rated speed by a ramp generator. The turbine speed is then controlled by a flow controller which senses pump discharge flow. Figure 1 illustrates a typical turbine EG governor control system. The ramp function is initiated when the steam supply valve leaves the fully closed position. At the same time, the turbine governor valve should close and slowly open in response to the ramp signal. However, with the initial admission of full steam flow, the turbine acceleration is greater than 3000 rpm /sec. This requires an almost instantaneous governor response, and the governor valve must travel from the fully open to the fully closed position in less than 1 second to prevent turbine speed from reaching the overspeed trip setpoint. With this characteristic, the design of EG governors appears to have a high sensitivity to minor system deficiencies which should not, by themselves, cause the turbine to overspeed. The governor valves at the four plants having EG governors were determined to have not closed fast enough on turbine startup, thereby permitting an overspeed trip.

The corrective actions taken for this problem were identical at these four plants where it occurred. The startup method was changed so that steam is admitted to the turbine through a small bypass line prior to the admission of full steam flow. This allows the turbine speed to increase to the idle speed, which builds up oil pressure necessary for proper governor control prior to the admission of full steam flow. The bypass line installed around the steam supply valve incorporated a startup bypass valve and a flow restricting orifice. The bypass valve was made part of the automatic initiation logic, and will open before the normal steam supply valve opens upon receipt of a start signal for the system. With the introduction of the bypass valve, the system response times were increased. The increased response times at these four plants were still within the allowable times required in their technical specifications.

• Data in parenthesis refers to table and item numbers, and the system associated with the turbine (e.g., Table 1, item 2, auxiliary feedwater system).

i

+

DE p _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ > _ _ _ _ _ _ _ _ _ _ _TU_RB_IN_E_S_P_EED q p _S I_RE_D _S P_EE_D l i I I I U U I l l I I i I

i rF---47 -

' i Steam Supply EG-M l I II I

Valve O

CONTROL I BOX l L$J l l

l 6

MAIN STEAM - I mT2 L_ _ _l (FROM STEAM GENERATORS) I O ELECTRIC SIGNAL i g I gTO EG-R ACTUATOR 1

• I I TURBINE TRIP AND e i

l TitROTTLE VALVE , s I __h EG-R '

l [o .,

ACTUATOR l v v l

/ _j ' a i c 3  : < -

,o

, N l **

l l TURBINE LUBE OIL SYSTM

. Auxiliary Auxiliary TURsINE EG-R OIL PUMP n "

Feedwater Pump Feedwater GOVERNOR REMOTE (INTEGRAL WITH )

SERV 0 EG-R - DRIVEN i Turbine VALVE OFF TURBINE SHAFT)

LEGEND: _

0 MAIN STEAM LINE f l CIL FILTER I I HYDRAULIC OIL LINE OIL SUMP FOR

---

ELECTRIC CONTROL SIGNAI S Figure 1 EG-R ACTUATOR Auxiliary Feedwater Pump a Turbine EG Governor Control System rROM j # ,

TURBINE i

• OIL SUMP

a ,

The startup bypass modification was recommended for all RCIC systems to improve startup performance in General Electric Service Information Letter (SIL) 377 issued on June 1, 1982 (Ref. 4) to all BWR owners. This SIL indicated that the startup bypass would result in increased system reliability and availability, and would result in less wear on system components.

In summary, ten events of turbine overspeed trip due to slow response of an EG governor have occurred at four plants (two CE PWRs and two BWRs). Jhesteam supply valve at these plants are used in starting the turbines. The design allows full steam flow to the turbines during turbine startup. Thus, control of full steam flow during the initial startup transient is dependent on the governor valve, which may not be able to respond quickly enough in some cases.

These four plants have used a steam bypass modification to solve the problem.

This problem was observed on RCIC turbines at BWR units and relevant information was fed back to all BWR plant operators in SIL 377 by General Electric (GE). In this SIL, GE recommended that BWR owners consider modifying the RCIC system with a steam bypass to improve RCIC startup performance. There are at least 10 other PWR plants which also utilize an EG governor. Some of these plants depend solely on the steam supply valve to start the turbines.

For these plants, the potential exists that the governor may not respond fast enough to control the turbine startup acceleration and result in a turbine overspeed trip. The solution implemented at the four plants has generic applicability to AFW turbines. A review of feedback correspondence, however, indicates that the problem encountered at the four plants has not been formally communicated to the other potentially affected PWR plants.

2.2 Condensate in the Steam Line Table 2 identifies a total of nine turbine overspeed trip events that occurred at four PWR plants as a result of the presence of undrained condensate in the steam supply lines to the turbines. The undrained condensate had accumulated because of either residual condensation of steam during cooldown after use of the system or rapid condensation of steam which was introduced into the cold supply line upon turbine startup.

In June of 1977, the AFW pump turbine at Crystal River 3 tripped on overspeed upon initial startup. Although a new governor was installed, the problem persisted. The cause of the overspeed trip was then attributed to condensate in the steam supply line, which disrupted the proper operation of the governor valve. Condensate, containing significantly less energy than an equivalent mass of steam, tended to slow the turbine and caused the governor to open the governor valve further. When the condensate cleared and steam followed, the governor could not close the governor valve fast enough to prevent a turbine overspeed trip. To prevent the buildup of condensate from recurring, the drain system for the steam supply line was modified to increase its capacity.

Subsequent multiple cold starts indicated that the problem had apparently been corrected. However, on July 17, 1977, the AFW turbine tripped again on overspeed upon auto start following a reactor trip. The trip was reset and the immediate restart was successful. However, it tripped again during several subsequent operability tests. The cause of the trips was determined to be l

condensate buildup in the steam supply line upstream of the steam supply valve.

The licensee corrected the problem by installing a bypass line around the steam supply valve to facilitate removal of condensate from the steam line upstream of the supply valve.

1 1

--.--1 --

r- ,- _, - , - - .y

. o Among five AFW turbine overspeed trips resulting from condensation at Arkansas Nuclear One (AN0), one occurred at ANO-1 and the other four at AN0-2. All of these trips occurred when the turbines were started in order to perform a safety function either automatically or manually. Water slugs in the steam lines were due to the steam traps having been isolated during the previous plant heatup or maintenance activities. In order to prevent recurrence, the licensee initiated a program to have the steam lines blown down once per shift, and implemented additional administrative controls to prevent improper isolation of the traps (i.e., the operating procedures were modified to include the steam trap isolation valves in the valve lineup).

The AFW system at the Davis-Besse plant currently relies on two turbine-driven pumps and an electric-driven emergency feedwater pump for the auxiliary feedwater system. The system at the time of the June 9, 1985 event consisted of two separate pumps and turbines. Following a loss of main feedwater flow, both AFW turbines started and tripped on overspeed approximately 25 seconds after initial roll on June 9,1985, and resulted in a total loss of feedwater flow to the steam generators. It was determined that a slug of water, formed in the steam piping to the turbine due to residual or rapid condensation of steam while heating the long, cold steam supply line, caused the overspeed trips. The licensee's special report (Ref. 5) on the overspeed trip problem indicated that water induction to the turbine can indeed cause an overspeed trip. The relationship between water slug and turbine overspeed is based on Terry Turbine's knowledge of turbine design as well as the result of some special tests performed in 1969. The testing involved injecting quantities of water into the turbine to verify the turbine's ability to withstand the water.

The test results indicated that the turbine speed began to decrease as the water slug came through the governor valve. The governor then tried to maintain turbine speed by opening the governor valve further. When the water slug cleared and steam followed, the valve was open too far to enable the governor to react rapidly enough to control turbine speed before reaching the overspeed trip setpoint.

To further support the conclusion that water in the steam supply lines had caused the overspeed trips, the licensee performed a review of past AFW turbine quick starts. The results of this review clearly showed a difference in the stability of turbine speed when operated with a cold versus hot steam supply line. When the turbines were started with a cold steam supply line, the turbine speed was very erratic during acceleration to rated speed. However, when the turbines were started with a hot steam supply line, the speed was very smooth with a constant acceleration to rated speed. The erratic turbine speed was attributed to the formation of water slugs when the steam was introduced into the cold steam supply lines during early heatup of the lines.

To eliminate or reduce water formation in the steam supply lines, the licensee for Davis-Besse has decided to relocate the steam supply valves closer to the turbines. This new configuration will provide the ability to maintain most of the steam piping in a hot and pressurized condition such that the amount of water delivered to the turbines during a start would be reduced to a necligible quantity. The licensee will also review the capacity of steam traps in the supply lines for adequacy. New steam traps will be added as required.

During a plant hot functional test in July 1985, the AFW turbine at Palo Verde Unit 2 tripped on overspeed when starting from a cold condition after receiving

. o an actuation signal. This failure was not repeated on subsequent starts when the system was hot. Condensate accumulation in the long steam supply line due to an inoperable steam trap was the cause of the turbine trip. The licensee's corrective action was to include periodic verification of steam trap operability in the surveillance testing of the AFW pumps.

~

In this respect, Terry has specified in its instruction manual that water slugs cannot be carried over into the turbine. If wet or saturated steam >is used, it is very important that the piping be arranged such that condensate Gill not be carried over into the turbine. Condensate is not only harmful to the turbine, but also adversely a'fects control of the turbine speed. The manual indicates that a steam separator of the proper size with a steam trap of ample capacity should be installed before the turbine inlet.

In summary, water induction into a turbine can cause an overspeed trip. Four PWR plants have experienced overspeed trips due to water in the steam supply lines. The source sas condensate which had accumulated in the steam lines due to inadequate or innperable provisions for condensate removal. An analysis conducted at Davis-besse showed that the existing steam piping configuration of the AFW system would generate a significant amount of water due to steam condensation and that this water would reach the turbine during the initial turbine startup transient. This would occur even though the steam lines were provided with steam traps. The solution to this problem was to either increase the capacity of the condensatt separation and removal process or to minimize condensate formation by keeping most of the steam line in a hot and pressurized condition. Administrative controls were also to be implemented to ensure operability and availability of the steam traps that were already installed in the steam supply lines. Adequate condensate separation and removal from steam lines is one requirement for the design of steam lines. However, due to the actual piping configuration, it is possible that the condensate may not be separated and/or removed fast enough to prevent it from reaching the turbine during a cold start. This can cause the turbine to trip on overspeed. The actual condition causing the everspeed trip is often not det?rmined because the subsequent system start will likely be successful. This occurs because the cold steam line has been heated up and the condensate removed. Accordingly, the cause of the turbine trip may remain undetected and the turbine may still trip on overspeed during the next demand. A review of the existing feedback correspondence indicates that neither PWR nor BWR licensees have been formally informed of this potential problem area. Accordingly, a number of the 15 events in the " unknown" category may be due to this phenomenon without the awareness of the plant operators.

2.3 Errors in Reset Procedure Table 3 contains one event involving an overspeed trip of an AFW turbine, due to an error in the resetting procedure for the unit. Turbine-driven auxiliary feedwater pumps equipped with a Woodward Model PG-PL governor will overspeed if restarted within a short period of time after shutting down because of the significant length of time required for hydraulic oil to drain from the governor's speed setting cylinder. The time needed for the oil to drain depends on internal component clearances. It may take as long as 30 minutes unless the speed setting knob is exercised after shutdown. The turbine speed c

at which the governor will control is determined by the volume of hydraulic oil trapped in the speed setting cylinder. During startup, the oil is initially

i supplied to the area with the rate properly controlled by the acceleration control feature (ramp bushing) to prevent the unit from overshooting the desired speed. With entrapped oil, the acceleration will not be properly controlled on starting which could cause the turbine to overspeed. A schematic diagram of the PG-PL governor is shown in Figure 2.

The event that occurred at Turkey Point 3 on July 22, 1985, involved both AFW turbines tripping on overspeed upon restart within 30 minutes after> shutdown.

The AFW turbines at this plant were equipped with Woodward Model PG;PL gover-nors. During the event, the normal pump shutdown procedures were inadequately implemented. The licensee's shutdown' procedure for the AFW pump specifies that the trip and throttle (T&T) valve for each turbine be opened prior to exercising the speed setting knob. However, these instructions lacked clarity.

During the incident, the speed setting knobs of the two AFW turbines were exercised prior to opening the T&T valve for the two turbines. Subsequently, when the T&T valves were opened, the turbine rolled and due to oil pump rotation, oil was admitted into the speed setting cylinder. Consequently, the two AFW pumps tripped on mechanical overspeed when next called upon to operate.

The licensee has revised the shutdown procedures to clearly specify the proper sequences for these actions.

The turbine vendor, Terry, has advised plants utilizing Woodward Model PG-PL governors of this condition, and has recommended a proper reset procedure The reset (Ref. 6) for bleeding the entrapped oil after equipmentSince shutdown.

exercising the speed requires manual exercise of the speed setting knob.

setting knob can only be performed locally, this may present a potential problem for ensuring immediate availability of the AFW system (if there is no other AFW pump operable), if an automatic start of the AFW pump is required within a few minutes after shutdown. Terry states that entrapped oil can be remotely or automatically drained by installing a dump valve in the governor hydraulic circuit. Two types of dump valves are available, these being solenoid operated (remote; i.e., from the control room) and pressure actuated (automatic).

Summarizing, turbines with Woodward Model PG-PL governors can overspeed if restarted within 30 minutes after shutdown unless the speed setting knob is exercised to bleed out the entrapped hydraulic fluid in the governor speed setting cylinder. Since exercising the speed setting knob and subsequent resetting can only be performed locally, this may present a potential problem for ensuring immediate availability of the AFW system (if there is no other AFW pump operable) as an automatic start of the AFW pump may be required within a In addition, this action may be performed under few minutes af ter shutdown.

stress and it is possible that the entrapped oil may not be fully dumped or the speed control may not be accurately reset.

l 2.4 Opening the Steam Supply Valve With the Trip Valve Open i

A review of operational data shows that two plants (Trojan and Wolf Creek 1) use the T&T valve to start the turbine. Each of these plants has l

' experienced a turbine overspeed trip event as a result of opening the steam supply valve while the T&T valve was in the open position. These events occurred during manual starting of the AFW pumps and are identified in Table 4. One event occurred at Trojan in 1983, and the other at Wolf Creek 1 in 1985. The event which occurred at Trejan was caused by a procedural inadequacy in that

- 15 -

N f-

s  :

=

5 A I

,,, s u - g

'glo 7 %7^'h N -

O-3

.s ' p' ' p ' \ t 'M Skl g a.u, . l. i . .

}

E il l i f -

i u

s e.: a 6 O O s .:!

E 1

gf

!s R V y

y 5 d

• O d $g, l1s rg k

t, I o to  ; t

- i h ;e

,. /

.,, i 1 {:

• 9 2a: a !f; a: asa i

-![ .

E "; 59582 ,3 i @ i 15 Ik f  !!!' l$ =

- l! an!!!

,1:  !

!s U ii

,1 e: t > 5 1 ii lj s g g i5 ? is i T "'

$2 g - er :f

! e 81 3y a,;

[ f !y 5,5 i li

! e

!! r il r sr rss : e e 1

\\lg l(I 4

1 g

-- o /.-  ! /~h s, Fy

,d w5 l" kSr% M,miy-n? d " $ t%

!s,:! e x-

=;j s "=s o

= (

o </ .

9

$ ~~p e o

w ai , w ;

e 3 ms t ,

5:

Z}222-?d l

ris-1717 /w77/7 -

i- L 48 i -

1 1 3 i , W:  : -

. ru m , - .

^

I \ "! ..

, O *' !a

!i L1  !!

j lIN E !tII  !.  : sz s o-a s :gg r -

. 2 sg !r s s< g < )

i e

s (

j,* li

.v. -

( 6  :

L- m C . . . . . . .w. . . . anawn n r1

' ' #~1_ + M1C- -@ ..

"; n

.n i w, ,=V w -

3 ,

.a R . M 51

- )

5-l.

I-q Ig--g 1 k

. A, og IE g, 8 nshmm t.e: ;I I g!

j

!!k u I s.

e 6

i l

the procedures concerning restart of the pump did not adequately address the problems which could occur when an immediate restart was attempted.

The automatic start signal opens the T&T valve over a 20-second time period (versus less than 5-second stroke time for the supply valve) to prevent turbine overspeed. Until repositioned locally, the T&T valve remains open when the pump is shut down from the control room by shutting the steam suppip valve.

The operator at Trojan manually stopped the AFW pump from the contrp1 room which involves only closing the steam supply valve. Therefore, when the operator attempted to restart the pump by opening the supply valve, the pump tripped on overspeed, since the T&T valve was already fully open; an overspeed trip would occur each time the operator used only the steam supply valve to manually start the pump.

Using the supply valve to manually stop the turbine can make the ArW pump unavailable if the pump is started before the T&T valve is closed. Failure to close the T&T valve before starting the turbine will cause the pump to trip on overspeed. The overspeed trip disables the pump until the trip is reset locally at the pump.

The overspeed trip of the AFW turbine at Wolf Creek I was also the result of operator error when the operator manually started the pump from the control room, and opened the steam supply valve with the T&T valve open. This opera-tion was out of proper sequence. The T&T valve for the AFW turbines at Trojan and Wolf Creek I are designed to open slowly when an auto-start signal is present, in order to prevent a turbine overspeed. The T&T valve should not be opened before opening the supply valve when manually starting the pumps.

The 1&T valves for the AFW turbines at these two PWR plants are designed to open slowly during turbine startup (a 20-second time period versus less than 5-second stroke time for the steam supply valves). In these events, the turbine overspeed trip occurred as a result of opening the steam supply valve while the T&T valve was in the open position. This indicated that the governors may not respond fast enough to control full steam flow from opening of the steam supply valve when the T&T valve is not modulating steam flow during turbine startup. The governors of the AFW turbines at these plants were Woodward Model EG governors. These two events also indicate that these governors tend to respond too slowly during turbine startup.

After the mechanical overspeed trips at Trojan and Wolf Creek 1 actuated, they needed to be reset locally. Even though the AFW turbine at Wolf Creek 1 is equipped with an electronic overspeed trip, which is normally set at 110 percent of rated speed, the mechanical overspeed trip, which is normally set at 125 percent rated speed, also actuated. The electronic overspeed trip can be reset remotely from the control room, whereas the mechanical trip cannot. The electronic trip wcs installed with the expectation that it would help avoid trips of the mechanical overspeed trip device. Operating experience, however, has shown that the acceleration of the turbine is generally so fast that the electronic trip will not slow the turbine dcwn before the mechanical overspeed trip actuates. This problem was observed earlier on RCIC turbines at BWR plants and was discussed in SIL No. 382, dated October 1982 (Ref. 7) by General Electric. This SIL recommended removing the electronic overspeed trip feature from the RCIC turbine control system. A review of feedback correspondence

O o indicates that the PWR operators have not been formally informed of the speed overshooting experience of the electronic trip.

2.5 Stop Valve Stroking Time Too Fast Operating experience of turbine overspeed trips caused by steam stop valve problems appears to be confined to BWR HPCI turbines. Table 5 lists only two incidents of HPCI turbine overspeed trips resulting from rapid, erratic opening of the stop valve during turbine startup. One of these incidents occurred at Peach Bottom 3 and the other at Brunswick 1. The overspeed trip at Peach Bottom 3 occurred during an automatic startup of the turbine. The licensee has stated that the overspeed trip was due to rapid stop valve opening. The rapid opening of the stop valve was a result of valve damage. The turbine trip at Brunswick 1 occurred during a routine performance test. The failure analysis performed by the licensee also indicated that the stop valve had opened too fast (6 seconds versus 12 seconds), allowing rated steam flow to enter the turbine before the governor valve could respond to control the high acceleration rate. The HPCI turbines provided by Terry Corporation are equipped with Woodward Model EG governors. The opening transient for the stop valve should be smooth, with no erratic movement. Rapid opening of the stop valve will significantly increase a startup acceleration transient, by not allowing the ramp generator to gain speed control before the overspeed condition is reached. The opening rate will vary as hydraulic oil is cycled between the stop valve hydraulic cylinder and the governor valve power piston.

Erratic movemeat of the stop valve is indicative of improper steam balance chamber pressure, or inadequate hydraulic pressure or flow.

Problems associated with rapid, erratic opening of the HPCI turbine stop valve were discussed in General Electric SIL 352 dated February 1981 (Ref. 8). This SIL indicated that this problem was identified primarily with the system " cold quick start" transient. An oil drum air accumulation problem was identified and was believed to be associated with down-time periods. To solve the problem, all HPCI turbines were modified to include a loop seal in the hydraulic cylinder 4-inch " dump" line. With the completion of this modification, the magnitude of the stop valve opening problems was reduced significantly.

However, continued problems at two sites resulted in damage to the stop valve seat, the stop valve stem, and the hydraulic cylinder seals. Further investigations at these sites showed that the stop valve balance chamber pressure had not been adjusted. Generally, the steam balance chamber adjustments were made during initial plant startup with no further adjustment required. The erratic opening problem of the two stop valves was corrected I

with a steam balance chamber adjustment. SIL 352 also recommended that BWR

• owners with HPCI Terry turbines monitor the opening transient of the stop valve during HPCI surveillance testing. For this monitoring, a transient recording of stop valve position and balance chamber pressure was to be obtained. Any changes in the response during subsequent surveillance recordings required an immediate inspection of the hydraulic circuitry and the steam balance chamber I adjustment. The actions taken in response to SIL 352 appear to have corrected the rapid, erratic opening problem of the HPCI turbine stop valve, since our search of the LER and other operational experience databases identified no additional similar problems.

l t

The hydraulic overspeed trip circuitry with automatic reset is a unique ,

overspeed trip mechanism for the HPCI turbine. The stop valve is a part of l this hydraulic circuitry and, therefore, the stop valve stroking time problem is not applicable to the AFW turbine. However, the turbine overspeed trips of these events indicate that Woodward Model EG governor may not be able to respond fast enough in controlling the full steam flow during the initial startup transient as the case described in Section 2.1. The HPCI turbines provided by Terry Turbine are equipped with Woodward Model EG goverpors. The stop valve for a HPCI turbine is located downstream of the steam supply valve and upstream of the governor valve. The stop valve is designed to open slowly during the initial stage of turbine startup so that the speed peak on the first acceleration transient will not exceed the rated speed of the turbine. The second acceleration transient is controlled by the ramp generator in the governor. This acceleration transient will ramp to the speed demand value of the flow controller. Two plants have experienced HPCI turbine overspeed trips due to rapid opening of the stop valve during turbine startup. In these cases, the control of full steam flow during the initial startup transient was dependent on the governor valve, which could not respond quickly enough to control the speed before it reached the setpoint. These events indicate that the EG governor tends to respond too slowly during turbine startup.

2.6 Turbine Spinning The events identified in Table 6, which occurred at St. Lucie 2 and Crystal River 3, are similar in that the overspeed trips of both events were the result

' of the turbine spinning due to steam leakage. At St. Lucie 2, the "C" AFW turbine tripped after receipt of an auto start signal; and at Crystal River 3, the AFW turbine tripped several times during monthly surveillance testing. The overspeed trips were found to be caused by steam leakage past the steam supply valve. This caused the turbine to roll at a low RPM before the turbine was started. The turbines at both plants were equipped with Woodward Model PG-PL governors.

The governor senses turbine speed through gearing which rotates the mechanical fly weights, and provides power to supply hydraulic oil to the governor. The governor acceleration control feature is set to function properly on a start signal when the turbine starts from a stop, and there is initially no oil pres-sure in the governor's speed setting cylinder. With the shaft rolling at low RPM prior to turbine start, hydraulic oil may flow to and build up oil pressure in the governor's speed setting cylinder. Oil pressure in the cylinder prevents a prompt governor response to close the governor valve and control the initial acceleration upon turbine starting. At the time this case study was completed, the licensee for Crystal River was intending to install e modified governor with an auto-bleed feature that will relieve oil pressure in the governor's speed setting cylinder to prevent a turbine trip on overspeed if the shaft is rolling prior to pump start.

On September 8, 1985, the AFW turbine at V.C. Summer was discovered to be rolling at approximately 200-250 RPM with no start signal present and the steam supply valve indicating closed. It was determined that the turbine roll was caused by seat leakage past the steam supply valve. The licensee believed that

inadvertent rolling of the AFW turbine might cause a turbine trip on overspeed and declared the AFW turbine to be inoperable.

Summarizing, turbines equipped with Woodward Model PG-PL governors may overspeed if the associated pump is already spinning at low RPM at turbine startup. Three events of AFW turbine overspeed trip due to turbine, spinning before turbine startup have occurred at three PWR plants. In these cases, spinning of the turbines was caused by steam leakage past the steam > supply valve. Turbine spinning at low RPM prior to start may result in the buildup of oil pressure in the governor's speed setting cylinder, thus disrupting prompt governor response upon turbine startup. Although control room indications for turbine spinning are provided at some plants, the oil pressure buildup in the speed setting cylinder resulting from turbine spinning does not decay immediately unless the speed setting knob is locally exercised and reset. This may present a potential problem for ensuring immediate availability of the AFW pump. The review of a Terry document (Ref. 6) identified 10 operating plants, other than St. Lucie 2, Crystal River, and V. C. Summer, that also use PG-PL governors in their AFW turbines. Terry indicated that an auto-bleed device can be installed to release oil pressure in the speed setting cylinder to prevent a turbine trip on overspeed if the shaft is rolling prior to turbine start.

However, this device is an optional item and not all PWR plants utilizing this model governor are equipped with the device.

2.7 Pump Air or Vapor Binding Air or vapor accumulation in the suction line of the AFW pump may result in a turbine trip on overspeed upon a pump start. An event at Zion 2 (7-1, AFW) involved a turbine overspeed trip resulting from air accumulation in the pump suction line. The suction line for the AFW pump at Zion 2 has a high point upstream of the pump inlet which allowed air to accumulate. When the pump was started, the accumulated air was carried into the pump casing. The pump became air bound, which resulted in a rapid reduction in pump impeller resistance and a fast acceleration of the turbine. The governor appeared to be unable to respond fast enough to prevent the overspeed trip. Subsequent licensee tests indicated that each time the AFW pump was required to run after an idle period of several hours, it would overspeed and trip. This clearly demonstrated that air accumulation caused the turbine overspeed trip. In addition to an existing casing vent, the licensee installed a continuous high-point vent line to the pump common suction line and the problem was corrected.

2.8 Trip and Reset Problems of Trip Valves and Overspeed Trip Mechanisms The 12 events identified in Table 8 show that 11 PWR plants have experienced trip and reret problems with the trip valve and overspeed trip mechanism (OTM) of their AFW turbines. In these events, the AFW turbine trip valves were thought to have been reset and ready to operate when, in fact,Inthey were in the each case, the tripped position (as if an overspeed trip had taken place).

spurious trip condition was noted either during routine inspection or as the result of an investigation of the failure of the turbine-driven pump to respond to an on-demand actuation. The trip valve or trip mechanism of an AFW turbine could be in a tripped condition without an indication to the operator in the control room. Thus, an automatic or remote manual start of an AFW pump would be prevented until the trip condition was discovered and reset. The lack of

reliable control room indication regarding availability of the AFW pumps represents a significant generic problem.

The undetected unavailability of the AFW pumps associated with these events could be attributed to one or more of the following:

• Lack of position indication in the control room for the T&T vaive and the OTM; g Position of overspeed trip latching mechanism can be misleading; Reset operation is complex, and not well understood by some operators; and The overspeed trip mechanism can be accidentally tripped by personnel working around it.

Control room indications for the positions of the trip valve and the overspeed trip lever are optional items and not all PWR plants are equipped with such devices for the AFW turbines. For the turbine to be available, the latch must be engaged and the T&T valve must be fully open. However, control room indica-tion at some PWR plants indicates when the trip valve is in a tripped (closed) condition, but does not identify the latched condition. There are cases where the trip valve is unlatched and the valve is indicated to be in an untripped condition. Under these conditions, the turbine would be unavailable. This type of installation does not provide an unambiguous indication of the trip valve position to the operator in the control room, and hence the turbine unavailability would be undetected.

At Rancho Seco and Salem 2, the operator found that the position of the overspeed trip lever can be misleading. The design of the linkage does not include any labels, operator aids, or easily identifiable position indication.

Thus, by observing the overspeed mechanism, it is not clear whether it is in tripped condition or not. A positive position indication for the overspeed trip linkage would ensure local verification of a reset condition. Without clear indication, it is difficult for the operator to determine the reset or tripped position of the overspeed trip mechanism. The licensee of Rancho Seco has installed a pointer device on the overspeed trip linkage to aid the plant operators during visual checks. Local mechanical indication for the trip valve position was also added at Salem 2.

The condition causing T&T valves and overspeed trip mechanisms to remain undetected in the tripped or unlatched position may be associated with the resetting process. Operators at some of these plants have noted that the spurious trip condition was a result of not properly resetting the mechanism after an overspeed trip during a maintenance or testing activity. At several plants, the mechanism had to be reset several times before the situation was corrected. Improved procedures for resetting and relatching the overspeed trip mechanism and special training for proper resetting operations were included in some of the corrective actions taken. Discussions with Terry personnel indicate that relatching the T&T valve is not a simple straightforward operation. Relatching the T&T valve involves two separate actions; first the trip tappet must be reset; and second, the trip hook must be engaged with the trip valve's latch-up lever so that the valve stem can be raised to open the valve. (Refer to Figure 3 for the location of each component of the T&T valve

l t

a m l

TRIP HOOK REPOSITION SPRif(G 9 '6

~

. E% Trip Hook u.as e ir!!

we 4 17 Nf1G&l V*fd

- Tno c):p 3 h Leer 4 -

ore s- Head p o

Lewf n *:

ff sa Tappet Turbine Trip Mectsanism

'.1

} -

steam Staam In > To

& Turdine Trip Vatve Figure 3 T&T Valve and Overspeed Trip Mechanism (OTM)

and OTM.) In order for the latch-up lever to be engaoed with the trip hook, it is necessary to rotate the handwheel in the close direction until engagement occurs. This is required even though the valve is already in the closed posi-tion. In case of a mechanical overspeed trip, reset of the T&T valve and 0TM can only be performed locally. For other trips of the T&T valve equipped with the electric reset feature, the T&T valve can be reset remotely through the optional electrical solenoid.

SomeofthetripandresetproblemswerepreviouslyidentifiedinII Information Notice 84-66 (Ref. 14). However, these problems continue to exist.

Operating personnel at Davis-Besse also experienced difficulty in relatching and opening the AFW turbine T&T valves during the event of June 9,1985. The licensee's subsequent review of the operating records revealed that numerous spurious trips relating to the T&T valve reset problem had occurred previously.

In view of this finding, the licensee conducted a thorough investigation (Ref. 9) in which it was determined that the most probable cause of these spurious trips of the AFW turbine was the operator's failure to reset the overspeed trip mechanism prior to latching and opening the T&T valve. The licensee found that it was possible to open a T&T valve without resetting the overspeed trip mechanism. This would result in the valve latch mechanism being in a metastable condition with the trip hook being held in place only by friction. The latch was thus unstable, and could be disengaged with a slight disturbance of the mechanism (e.g., T&T valve operation or mechanical shock).

The licensee identified several root causes for the improper resetting by the operators.

These were:

The reset or tripped condition of the overspeed trip mechanism is difficult for the operators to detennine, since there is no local position indication on the mechanism.

The procedural resetting instructions did not clearly identify that T&T valve relatching also required the overspeed trip tappet to be reset.

Training on the relatching operation included little " hands-on" operation of the mechanisms.

To prevent a recurrence of similar problems, the licensee of Davis-Besse has planned actions to correct the existing deficiencies related to the identified root causes. The licensee determined that it would be necessary to:

Install local indication for positions of trip and reset of the overspeed trip mechanism (OTM).

Install local position indication on T&T valve.

Provide means for operators working in the AFW turbine room to verify position indications for the T&T valve and 0TM and to conrnunicate with the operators in the control room.

Modify the applicable procedures to incorporate the proper reset sequence for the OTM and the T&T valve and ensure that the T&T valve and 0TM are properly reset af ter completion of testing.

. o

• Install a handle on the connecting rod of the OTM to facilitate easy operation of this mechanism during resetting.

• Revise the reset instruction to include detailed theory of operation for the OTM and T&T valve. The instruction will be included in annual operator requalification. ,

Provide " hands-on" training on operation of the T&T valve and the OTM to the operators. The training will also become a requirement fof operator qualification.

2.9 Worn or loosened Nuts and Linkages A significant number of turbine overspeed trips identified in this study were associated with worn or loosened nuts and linkages. Most of the degraded nuts and linkages which resulted in turbine trips were associated with the T&T valve and overspeed trip assembly. A few of them were associated with the governor.

A total of 18 events resulted from worn or loosened nuts and linkages. These events are identified in Table 9. In almost every case, a premature overspeed trip occurred. A review of these events shows that the degradation of nuts and linkages can involve one or more of the following: worn tappet nut face, loose cable connection, loosened set screws and worn pistons of the tappet assembly, improperly latched T&T valve, and items resulting from improper assembly.

Six events that were related to worn tappet nut faces occurred at Browns Ferry 2 (9-6, RCIC), Quad Cities 1 and 2 (9-12, 9-7, RCIC), Millstone 2 (9-15, AFW), North Anna 2 (9-16, AFW), and Duane Arnold (9-18, RCIC). Normal wear of the tappet nuts caused improper alignment of the tappet nut-to-head lever, thereby reducing the holding force on the head levers. The reduction in holding force was such that normal operational vibration would release the head lever and cause the turbine to trip prematurely. The current Terry maintenance manuals recommend that the overspeed trip device linkages and clearances be checked once every year for wear. However, the number of turbine trips resulting from tappet nut degradation suggests that not all licensees are following the recommended frequency for checking the linkages.

Two overspeed trip events at Hatch 1 (9-8, and 9-9, RCIC) were related to improper latching of the T&T valve. As described in Section 2.8, the T&T valve could be improperly latched without operator awareness. When the valve is not properly latched, the trip hook is held in place with friction such that the vibration resulting from turbine startup could unlatch the valve and trip the turbine.

Three events that were related to the trip assembly of the HPCI turbine occurred at Brunswick 2 (9-1, HPCI), Browns Ferry 1 (9-2, HPCI), and Vermont Yankee (9-4,HPCI). The turbine trips at both Brunswick 2 and Vermont Yankee were attributed to loosened set screws, while the trip at Browns Ferry 1 was caused by a worn piston. The piston is locked on the threaded stem of the tappet and ball holder by a set screw inside the hydraulic trip assembly of the HPCI turbine. A loosened set screw will cause the trip set point to drift, while a worn piston would result in insufficient hydraulic pressure in the oil supplied to the pilot valve on the turbine stop valve such that full hydraulic pressure would not be developed to hold the stop valve in the open position.

The degradation of set screws and pistons along with other design deficiencies

for the overspeed trip assembly were identified in GE SIL 353, dated February 18, 1981 (Ref. 10). In an effort to improve the reliability of the overspeed trip assembly on HPCI turbines, Terry made several design changes and successfully tested an improved overspeed trip assembly. The improvements included a new piston and the use of a cotter pin to replace the set screw. In SIL 392, dated April 1983 (Ref. 11), GE advised BWR owners of the availability of the improved mechanical overspeed trip assembly for HPCI turbines.

lhe events that were attributed to loosened cable connectors occurr d at three BWR plants. These plants were Brunswick 1 (9-14, RCIC), Browns Ferry 3 (9-3, HPCI), and Pilgrim 1 (9-5, HPCI). The loosened cable connectors were installed between the governor control module (EGM) and the electro-hydraulic servo (EGR) in Woodward Model EG governors, and were manufactured by the Amphenol Company. This degradation may indicate a generic problem. Following the incident, Browns Ferry revised the equipment inspection manual to include verification of the tightness of connectors.

Four events that could be related to improper assembly occurred at Prairie Island 2 (9-10, AFW), Quad Cities 1 (9-11, RCIC), Browns Ferry 2 (9-13, RCIC) and Sequoyah 1 (9-17, AFW). The turbine trip at Prairie Island 2 was attributed to loosened linkage between the governor and the pressure compensator. The turbine trip at Quad Cities 1 was caused by a failure of the overspeed trip mechanism. A ball and tappet assembly within the trip mechanism was found to contain dirt and crud which prevented the ball and tappet assembly from operating properly, thereby causing a premature RCIC turbine overspeed trip. The trip at Browns Ferry 2 resulted from a loosened nut on the turbine wheel, which allowed the shaft to strike the pick-up, causing a speed control malfunction. The trip at Sequoyah I was due to normal unit vibration which loosened a locking nut on a trimpot of the control circuit, causing the set-point to drif t. All three of these degradations were corrected by tightening the nuts or linkages.

2.10 Misadjustment, Miscalibration o- Drifting Controller A total of 17 events are identified in Table 10 in which turbine overspeed trips occurred as a result of a misadjustment or miscalibration of the governor speed controls and trip linkages, and/or drifting controllers. Among these, three events occurred when a turbine actuated in response to a safety demand.

The remaining events occurred during surveillance testing. A review of these events suggests that those which involve a misadjustment or miscalibration may be grouped into three categories. These categories are: (1) incorrect governor speed setting, (2) misadjustment of the trip linkage assemblies, and (3) misadjustment of the governor valve linkages.

Eight events were identified that related to an incorrect governor speed setting. These events occurred at Pilgrim (10-1, HPCI), Brunswick 2 (10-4, RCIC), Salem 1 (10-7, AFW), Peach Bottom 2 and 3 (10-5, 10-6, HPCI) Farley 1 (10-9, AFW), and North Anna 1 (10-10 and 10-13, AFW). In most of these events, the governor speed was set too high without the operators being aware of the errors. Normally, the governor speed setting should be verified prior to performing surveillance testing. However, six of these events occurred during surveillance testing. It appeared that the operators were not aware of the incorrect speed settings. The other two events in which the turbines actuated on automatic safety signals were determined to be caused by the operators'

failures to follow procedures and check speed settings. Therefore, events involving incorrect governor speed settings may be attributed to an inadequacy in the adjustment and calibration procedures, an operator's failure to follow procedures, or a combination of the two.

Questions concerning procedural inadequacy in defining step sequances for governor control system calibration of the Terry turbines used in the HPCI and RCIC systems were raised in GE SIL 351 (Ref.12) issued in February >1981. This SIL provided BWR owners with revised and more detailed calibration procedures for the components used in the control systems of HPCI and RCIC turbines. A better defined step sequence for governor speed setting also was included in this SIL. This indicates that a procedural inadequacy in governor speed setting also may exist for the Terry turbines used in PWR AFW systems.

Three events which related to misadjustment of the trip linkage assembly were identified. These events occurred at Peach Bottom 2 (10-2, RCIC), Browns Ferry 2 (10-8, HPCI), and Cooper (10-11, HPCI). Also, two other events were identified which related to the governor valve linkages being out of adjustment. These events occurred at Browns Ferry 1 (10-3, RCIC) and Vermont Yankee (10-12, RCIC). The failure to have a component adjusted properly was attributed to poor maintenance activities due to inadequate human performance during such activities.

Four events were attributed to drifting controllers. These events occurred at Cooper (10-14, RCIC), Duane Arnold (10-16 and 10-17, RCIC) and Sequoyah I (10-15,AFW). The turbine trips at Cooper and Duane Arnold were caused by drifting flow controllers, and the turbine trip at Sequoyah was a result of drifting in the electronics associated with the speed controls.

Turbine overspeed trips have occurred as a result of misadjusted or miscali-brated governor speed controls and trip linkages, and/or drifting controllers.

Events involving incorrect governor speed settings may be attributed to an inadequacy in the adjustment and calibration procedures, an operator's failure to follow gocedure, or a combination of the two. Procedural inadequacy in governor speed setting, identified in GE SIL 351 for BWR RCIC and HPCI turbines, may also exist for Terry turbines used in PWR AFW systems.

Misadjustment of trip linkage assemblies can be attributed to poor maintenance activities due to inadequate human performance.

2.11 Oil Leak and Contamination A total of 23 events involving turbine overspeed trips attributed to an oil leak or contamination in the governor speed control systems or the hydraulic overspeed trip assemblies are identified. Oil leakage or contauination in the hydraulic system of a turbine could prevent the system from developing sufficient oil pressure to perforn its functions properly when the turbine is in operation. Eighteen of the 23 events were attributed to oil leaks and the others were attributed to oil contamination. All of these events occurred during surveillance testina. The 18 events involving oil leakage occurred at Brunswick 2 (11-2, RCIC), Prairie Island 1 (11-4, AFW), Fitzpatrick (11-5, RCIC), Browns Ferr i

(11-10, AFW), Palisades 11-11,(yAFW),

3 (11-6, RCIC),

and Peach Arkansas Botton 2 (11-9, 3 (11-7 and AFW), San On 11-8, HPCI).

! The dominant cause of oil leaks as illustrated by these events was component and seal failures. With the exception of the ruptured diaphragm in the i

{

Robertshaw valve installed in the overspeed hydraulic system of the HPCI turbine at Peach Botton 3, the causes of component failures appear to have no generic implications. In general, oil leakage can be detected by verifying the oil level in the turbines, and the correct oil level should be verified prior to surveillance testing. However, all of the turbine trips noted above occurred during surveillance testing; therefore, it appears that oil levels were not checked prior to surveillance testing. The licensee for Prairie Island revised the appropriate procedure to include oil level verification after each putnp shutdown. Failure of personnel to verify correct ofi level for the turbines could be attributed to inadequate procedures. However, the problem of the ruptured diaphragm associated with the Robertshaw valve was discusred in GE SIL 358 (Ref. 13) issued in June 1981. This SIL reccmmended a new type of material for the diaphragm in these valves.

The five events involving oil contamination occurred at Monticello (11-1, HPCI), Vermont Yankee (11-3, HPCI), LaSalle 1 (11-12, RCIC), North Anna 1 (11-13, AFW) and North Anna 2 (11-14, AFW). The oil contamination was attributed to poor maintenance activities on the turbines.

2.12 Electrical Component Failures Thirteen events in which turbine overspeed trips were attributed to electrical component failures are identified in Table 12. Most of the failures of electrical components in these events result from overvoltage spikes, erroneous The grounds or connections, environmental conditions, and/or misadjustments.

electrical components involved were transistors, resistors, diodes and relay lamps. These components are in the governor speed controls and electronic overspeed trip circuitry. Eleven of the 13 events occurred during surveillance testing, and the remaining two events were reported because of a turbine spurious overspeed trip. These two events resulted from failure of the relay lamp which also actuated the trip alarm.

Four of the 13 events were related to a failure of a transistor or a resistor in the EGM controls. These events occi "ed at Brunswick 2 (12-1, 12-5 and 12-8, RCIC) and San Onofre 2 (12-10, AFW). These failures in the EGM controls resulted in a loss of turbine speed control and led to a turbine overspeed trip. The licensee attributed these component failures to moisture intrusion into the EGM control and a reduced voltage condition.

Three events that were related to erroneous grounds or connections were identi-fied. These events occurred at Browns Ferry 3 (12-2, HPCI), Quad Cities 1 (12-9, RCIC), and Cooper (12-12, HPCI). At Browns Ferry 3, a grounded connector on the governor of the HPCI turbine caused it to fail to control the turbine speed. At Quad Cities 1, a ground in the RCIC logic circuit caused low voltage at the electronic overspeed trip sensor, which resulted in an electrical overspeed signal at a speed below the overspeed trip setpoint. At Cooper, reversed electrical connections between the governor control units (EGM and EGR) caused the governor valve to fail full open, resulting in a loss of governor speed control. The licensee attributed the wire reversal to a modification activity involving the electronics of the Woodward governor system during the previous refueling outage.

Two of the 13 events were related to a diode failure. These events occurred at D.C. Cook 2 (12-4, AFW) and Browns Ferry 3 (12-7, RCIC). The failure of the

diode at D.C. Cook 2 was due to an electrical short, and the failu e of the diode at Browns Ferry 3 was due to end of component life (i.e., apparently a random failure). ,

Two other events involved overspeed trips of the AFW turbine, and occurred at Trojan (12-3, AFW) and Farley 1 (12-6, AFW). At Trojan, the cause of the turbine overspeed trip was a faulty governor speed sensor circuit card. The cause of the failure was unknown and appeared to be unrelated to any maintenance activity. The AFW turbine overspeed trip at Farley I wa5 caused by an out of adjustment limit switch on the steam supply valve. The limit switch allowed the ramp signal in the governor control circuit to demand full opening of the governor valve prior to opening the steam supply valve.

Two events that were related to failure of a relay lamp occurred at Quad Cities 1 in 1985 (12-11 and 12-12, RCIC). The relay lamp is associated with the electronic overspeed trip meter relay on the RCIC turbine. The meter relay is designed to register on overspeed alarm and trip the turbine upon the opening of the circuit. Failure of the relay lamp will cause the circuit to open. There have been previous problems involving the RCIC overspeed trip relay lamp at this plant. The recurring problems of this overspeed trip relay prompted a recommendation by General Electric to remove it. The electronic overspeed trip problem had been observed at other BWR plants, and was discussed in GE's SIL 382, date1 October 1982 (Ref. 7). The SIL recommended removing the electronic overspeed trip system from the RCIC turbines.

2.13 Summary The information in the preceding sections provides the major causes of past events involving overspeed trips of steam turbines associated with AFW, HPCI and RCIC systems. Some of the recommendations provided by GE to BWR owners for preventing overspeed trips of the HPCI and RCIC turbines were also presented.

Almost all the turbines in use for the HPCI, RCIC, and AFW systems are manufactured by Terry and equipped with Woodward governors; therefore, recommendations by GE regarding the RCIC and HPCI turbines also may be applicable to the AFW turbines used in PWRs. ' Conversely, the hydraulic overspeed trip circuitry with automatic reset is a unique overspeed trip mechanism for the HPCI turbines. Accordingly, the stop valve stroking time concern described in Section 2.5, the loosened parts in the HPCI trip assembly in Section 2.9, and the rupture diaphragm associated with the Robertshaw valve in Section 2.11 are parts of this hydraulic circuitry and, therefore, are not applicable to the AFW turbine. Excluding these items, the major causes for AFW turbine overspeed trips are summarized as follows in the order in which the overspeed trips were analyzed in the preceding sections.

(1) Slow governor response during quick start including:

Binding and sticky governor valve and linkage.

• Sensitivity of the EG governor speed control during the initial acceleration transient from idle speed.

(2) Condensate in steam lines due to:

• Undrained condensate due to the steam trap being isolated.

Insufficient capacity of steam traps.

• Inadequate procedures for verifying the availability and operability of steam traps.

(3) Turbines with PG-PL Woodward governors will overspeed if restarted within 30 minutes after shutdown unless the speed setting knob is locally exercised to bleed-out the entrapped hydraulic fluid frcm the governor speed setting cylinder.

(4) The T&T valves for the AFW turbines at some plants are designed to open slowly when an auto-start signal occurs. This design feature is intended to prevent the turbine from overspeeding. There have been cases in which the start valve was opened manually to start the turbine while the T&T valve was in the open position. The cause of such improper operations is a combination of operator error and procedural and training inadequacies.

(5) Turbires equipped with Woodward Model PG-PL governors will overspeed if the associated AFW turbine and pump are spinning at low RPM during startup or if air or vapor is drawn into the pump casing. The cause of these problems were identified as:

• Steam leakage past the steam supply valve, causing the turbine to roll.

• Air or vapor accumulation in the suction line of the AFW pump, causing the pump to inject the air or vapor into the pump casing during startup resultino in an actual turbine overspeed.

(6) Trip and reset problems of trip valves and trip mechanisms. These problems were found to be the primary causes of spurious turbine overspeed trips. Such problems could be attributed to one or a combination of the

• following factors:

• Lack of position indication in the control room for T&T valves and overspeed trip mechanisms.

Position of overspeed trip latching mechanism can be misleading.

Reset operation of trip valves and trip mechanisms is complex and is not well understood by some operators. Local indication of reset is inadequate or not available.

• The overspeed trip mechanism can be tripped accidentally by personnel working around it.

(7) Worn or loosened nuts and linkages caused by:

Worn tappet nut.

• LMse cable connector between EGM and EGR in Woodward EG governors.

Improper assembly resulting from poor maintenance activities.

f

29 -

(8) Misadjustment, miscalibratica or drifting controller due to:

Incorrect governor speed setting.

Misadjustment in the trip linkage assemblies and the governor valve linkage. .

Drifting flow controller. g (9) Oil leak and contamination involving:

Failure to verify correct oil level in the turbine.

Oil contamination due to poor maintenance activities and lack of a program to regularly check for oil impurity.

(10) Electrical component failures involving:

Transistors, resistors, diodes, relay lamps, and limit switches.

Contributing factors such as erroneous grounds, incorrect adjustment, and moisture.

1 2

-e-- .- ->- nn--r ,e -.-r,,,-- .n- , - , - - - , - , - - , - - , . .-- u-n- - - . . - --,

3.0 FINDINGS AND CONCLUSIONS Based on the preceding discussion and related follow-up activities conducted for this study, the following findings and conclusions are provided:

(1) The major causes for AFW turbine overspeed trips are speed control problems associated with governors, and problems related to relatching and resetting of trip and throttle (T&T) valves and overspeed tripgmechanisms (OTMs). These problems are mostly the result of inappropriate performance by plant personnel, inadequate procedures or insufficient design consideration.

(2) The Woodward Model EG governor (electro-hydraulic model) may not respond fast ene"gh to control turbine speed if full steam flow is admitted to the turbine at the very beginning of turbine startup. With the admission of rated steam flow, the turbine acceleration is so great that almost instantaneous governor response is required. Quite often the oil pressure required to close the governor valve is not developed fast enough to close this valve during initial startup. This allows entry of a large amount of high energy steam to the turbine. This ungoverned steam flow will likely cause the turbine speed to overshoot before the governor can respond to control the acceleration. Turbine overspeed trips due to slow response of the governor typically occur within the first few seconds of turbine startup.

(3) Ten events of turbine overspeed trips due to slow responses of the EG governor have occurred at four plants.

(a) The governor valves for the turbines at these plants were observed to have not closed fast enough during startup of the turbines. The steam supply valves at these plants are used in starting the turbines. Such a design allows full steam flow to the turbines during turbine startup. Thus, control of full steam flow during the initial startup transient is totally dependent on the governor valve which may not be able to respond quickly enough in some cases.

(b) The solution for the slow response problems associated with EG governors at the four plants was identical. This solution involved installing a bypass line around the steam supply valves to the turbines. The turbine startup method was changed so that a sufficient amount of steam is admitted to the turbine through the bypass line to establish idle speed and place the turbine on governor control prior to the admission of full steam flow. This modification also introduced a time delay in the opening of the steam supply valve and consequently increased the response time required for the system to reach rated flow. The magnitude of the time delay is plant-specific.

The response time after modification at Falo Verde was still within the allowable time of 20 seconds, while at Arkansas 2, the response time was increased from 21.4 seconds to 50 seconds, which was also within acceptable limits. However, for such a modification, the potential exists for rated flow to not be achieved within the time limit required by the technical specifications, i

(c) The EG governor startup transient also was observed for RCIC turbines at BWR plants, and was discussed in SIL 377 issued by GE in June 1982 (Ref. 4). This SIL recommended improving startup performance by installing a bypass line around the steam supply valve.

(4) HPCI turbines manufactured by the Terry Coorporation are equipped with Woodward Model EG goverr. ors. The stop valve for a HPCI turbine is located downstream of the steam supply valve and upstream of the goverpor valve.

The stop valve is designed to open slowly during the initial stage of tur-bine starting. Two plants identified in Section 2.5 have experienced HPCI turbine overspeed trips due to rapid opening of the stop valve during turbine startup. In these cases, the control of full steam flow during the initial startup transient was dependent on the governor valve, which could not respond quickly enough to control the speed before it reached the setpoint. The rapid opening of the stop valve was due to a valve malfunction.

(5) Two events identified in Section 2.4 indicated that the TheWoodward T&T valves EG for governor responded too slowly during turbine startup.

I the AFW turbines at these two plants are designed to open slowly during turbine startup (a 20-second time period versus less than 5-second stroke time for the steam supply valves). In these events, the turbine overspeed trips occurred as a result of opening the steam supply valve while the T&T valve was in the open position. In these cases, the governor may not have responded fast enough to control full steam flow from opening of the steam supply valve.

Four plants (6) Water induction into a turbine can cause an overspeed trip.

have experienced overspeed trips due to water in the steam supply lines.

The cause of condensate accumulation in the steam lines was inadequate or inoperable provisions for condensate removal. An analysis conducted for Davis-Besse has shown that the existing steam piping configuration of the AFW system would generate significant amounts of water due to condensation and that this water would reach the turbine during the initial turbine startup transient.

(7) A significant number of turbine overspeed In most trips were trips, of these overspeed related thetogovernor incorrect governor speed setting.

speed setting was too high without the operator being aware of this error.

Nonnally, the governor speed setting should be verified prior to performing surveillance testing; hewever, many of these events occurred during surveillance testing. This indicates that the incorrect settings can be attributed to inadequacies in the adjustment and calibration procedures, operator failure to follow procedure, or a combination of the two.

Turbines with Woodward Model PG-PL governors can overspeed if restarted (8) wf thin 30 minutes after shutdown unless the speed setting knob is exercised to bleed out the entrapped hydraulic fluid in the governor speed setting cylinder. Since exercising the speed setting knob and subsequent resetting can only be performed locally, this may present a potential problem for ensuring immediate availability of the AFW system (if there is no other AFW pump operable) because an automatic start In of addition, the AFW pump sincemay this be required within a few minutes after shutdown.

action may be performed under stress, it is possible that the entrapped oil may not be fully dumped or the speed control may not be accurately reset by the equipment operator.

(9) Turbines equipped with Woodward Model PG-PL governors may overspeed if the associated pump is already spinning at low RPM at turbine startup. In these cases, spinning of the turbine was caused by steam leakage past the steam supply valve. Turbine spinning at low RPM prior to startup nay result in the buildup of oil pressure in the governor's speed sett1ng cylinder, thus disrupting prompt governor response upon turbine startup.

Indication of turbine spinning is provided in the control room at some plants. Oil pressure builoup in the speed setting cylinder does not decay immediately unless the speed setting knob is locally exercised and reset.

This may present a potential problem for ensuring immediate availability of the AFW pump. Crystal River plans to install a modified governor with an auto-bleed feature that will release oil pressure in the speed setting cylinder to prevent a pump trip on overspeed if the shaft is rolling prior to pump start.

(10) Twelve events at 11 PWR plants have resulted from a trip and reset problem with the AFW turbine T&T valve and 0TH. In these events, the AFW turbine was in an overspeed trip condition without any indication to the operators in the control room. This lack of indication regarding availa-bility of the AFW pump represents a significant generic problem. Such problems could be attributed to one or a combination of the following items:

• Lack of position indication in the control room for T&T valves and 0TMs.

" Local reset indication for the T&T valves and 0TMs is inadequate or not available.

• Reset operation of T&T valves and OTMs is complex and is not well understood by some operators.

• The OTM can be tripped accidentally by personnel working around it.

Some of these concerns have been previously identified in IE Information Notice For example, 84-66(Ref.14). However, these problems continue to exist.

operating personnel at Davis-Besse experienced difficulty in relatching and opening the T&T valves associated with the AFW turbines during the event of June 9, 1985. In a subsequent review of the operating records, the licensee found that numerous spurious trips relating to the T&T valve reset problem had occurred without corrective actions being taken. To prevent recurrence of similar problems, Toledo Edison (the Davis-Besse licensee) plans the following actions: (1) install local trip and reset position indication for the T&T valves and the OTMs; (2) provide means for operators in the turbine room to verify the correct position for the T&T valve and OTM, and to communicate with the operators in the control room; (3) modify the applicable procedures to i reflect the proper reset sequence; and (4) provide detailed instructions and hands-on training to operators.

4.0 RECOMMENDATIONS As a result of this study, we conclude that tnere are areas that appear to require specific improvements and modifications for preventing or reducing unwanted turbine overspeed trips. For this reason, the following recommenda-tions should be considered with the j;eneric actions on reliability.of the AFW pump turbines that IE and NRR are taking in response to the ED0 memorandum dated August 5, 1985, regarding the staff actions resulting from th,e investigation of the June 9, 1985 event at Davis-Besse.

(1) Licensees of PWR plants utilizing a Woodwara EG governor on an AFW turbine should consider modifications to the AFW system to improve the turbine reliability during startup. The problem of overspeed trips during startup has been experienced by RCIC turbines in BWR plants. GE has recommended that BWR owners consider modifying the RCIC system with a steam bypass to improve RCIC startup transients. This GE recommendation should be reviewed for its applicability and feasibility to all PWR plants utilizing Woodward EG governors on the AFW turbines. The review should include the impact of the increased response time on system reliability, the criter';a for achieving an optimum bypass startup, and whether the technical specification response time could be modified to accommodate the increase resulting from a bypass modification.

(2) In view of the number of turbine overspeed trips resulting from incorrect governor speed settings, licensees should review the adequacy of the existing vendor-supplied calibration procedure used for the AFW turbine control system. Many of these turbine trip events occurred during surveillance testing. This suggests that the operators were not aware of the incorrect settings or had difficulty in determining the correct settings due to inadequate procedures.

(3) Licensees should review the adequacy of administrative controls to ensure the availability and operability of steam traps installed in the steam supply lines. Water induction into a turbine could cause an overspeed trip. It is important that condensate in the stean supply line should be removed before reaching the turbine. The capacity of steam traps should be sufficient to remove rapid condensate buildup, resulting from heating a cold steam line during turbine startup. The configuration of the steam supply line piping should be such as to minimize condensation and to facilitate easy removal of condensate during turbine startup operation.

(4) Licensees should be required to review the adequacy of the existing procedural instructions as well as their training programt regarding reset operations of T&T valves and 0TMs. Clear local indication of the position of T&T valves and 0TMs are needed. Additionally, consideration should be given to either routine local (e.g., once per shift) position verification or control room indication for a " latch" condition of T&T valves to ensure proper reset conditions and turbine availability. Therefore, licensees should also review the availabili+.y and the adequacy of these indications and provide modifications as appropriate.

(5) An IE information notice should be issued to alert licensees of the findings that led to the above recommendations. Specifically, the information notice should address events involving turbine overspeed trips resulting from entrapped oil in the governor speed setting cylinder. There were two conditions under which the entrapped oil in the speed setting cylinder occurred. These were: (1) the procedure for turbine shutdown and speed setting knob exercising was inadequately implemented, and (2) turbine spinning at low RPM prior to or at startup. Turbine spinning yas caused by steam leakage past the steam supply valve. The conditions tould be improved by: (1) establishing more stringent administrative controls for bleeding out the entrapped oil, (2) installing a remotely controllable dump valve in the governor hydraulic circuit, or (3) providing instrumenta-tion so that turbine spinning would be indicated in the control room.

l l

l

5.0 REFERENCES

1. M. Trojovsky, EG&G IDAHO, Inc., " Data Summaries of Licensee Event Reports of Pumps at U.S. Commercial Nuclear Power Plants," USNRC Report NUREG/CR-1205, January 1982. -

2. Memorandum from W. J. Dircks, NRC, to Directors of NRR, IE, RE5 and AE00, and Region III Administrator, " Staff Actions Resulting from the Investigation of the June 9 Davis-Besse Event (NUREG-1154)," August 5, 1985.

3. U.S. Nuclear Regulatory Commission, " Loss of Main and Auxiliary Feedwater Event at the Davis-Besse plant on June 9, 1985," U.S. NRC Report NUREG-1154, July 1985.

4. General Electric, SIL NO. 377, "RCIC Startup Transient Improvement with Steam Bypass," June 1982.

5. D. Wilczynski, Toledo Edison Co., Overspeed Trips of the Auxiliary Feed Pump Turbines on 6/9/85 at Toledo Edison's Davis-Besse Nuclear Power Station," Rev. 3, November 8, 1985.

6. Letter from R. R. Theroax, Terry Corporation, to R. W. Woodruff, NRC,

Subject:

Reset Procedures for PG-PL Governor Equipped with Ramp Bushing, dated February 27, 1979.

7. General Electric, SIL. N0. 382, " Removal of RCIC Electronic Overspeed Trip," October 1982.

8. General Electric, SIL NO. 352, "HPCI Turbine Stop Valve Steam Balance Chamber Pressure Adjustnent," February 18, 1981.

9. R. J. Gradomski, Toledo Edison Co., "AFPT Overspeed Trip and Throttle Valve Problem," Rev. 3, October 15, 1985.

10. General Electric, SIL N0. 353, "HPCI Turbine Mechanical Overspeed Trip,"

February 18, 1981.

11. General Electric, SIL N0. 392, " Improved HPCI Turbine / Mechanical Overspeed Trip Design," April 1983.

12. General Electric, SIL N0. 351, "HPCI and RCIC Turbine Control System Calibration," February 18, 1981.

13. General Electric, SIL N0. 358, " Replacement Diaphragms for Robertshaw Valve (Model No. VC-210)," June 1981.

14. IE Information Notice 84-66, " Undetected Unavailability of the Turbine Driven Auxiliary Feedwater Train," dated August 17, 1984.

o m

Table 1 Events of Slow Governor Response During Quick Start (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (1-1,AFW) Before the installation of a bypass Arkansas-2 EG All of these overspeed trips occurred almost immediately following the start line around the steam supply valve,80-022 which eventually corrected the per-80-036 signal. Some of these starts were actuated on demand, and others were sistent turbine overspeed trips, a 80-039 80-050 actuated during operability tests. The number of interim measures were taken.

turbine governor valve was visually Although some of the measures taken 4 incidents in 1980 did correct the problem, the licensee AFW observed either to have never moved during the transient or to have begun to determined that bypass modification can move only at about the same time as the provide better reliability for the turbine speed reached the overspeed trip system. Some of these measures were:

setpoint.

(1) Replaced governor actuator hyraulic fluid sump.

(2) Increased tubing internal size from the governor actuator to the remote servo.

(3) Installed a sight glass on the hydraulic sump to allow monitor-ing of oil level in the sump duria and between startups.

(4) Reduced governor valve travel from 7/8 inch to 5/8 inch per Terry Turbine recommendation.

(5) Reduced turbine idle speed from 1100 rpm to 800 rpm.

o (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (6) Delayed the initiation point for the ramp signal by 6 seconds.

(I-2,AFW)

Zion-1 PG-PL Following the reactor trip, two of the Atter the trip mechanism was76-042 three AFW pumps did not start to operate reset, pump operated satisfactorily.

8/8/76 as designed. The turbine-driven pump AFW tripped on overspeed after an automatic start. The trip probably was caused by a sticky governor valve not responding fast enough to control the turbine speed.

(1-3,RCIC)

LaSalle-1 EG At 10% power level, during a performance The linkages were reassembled to 82-089 of the vessel injection cold, quick start, vendor specification.

8/15/85 test turbine tripped on overspeed. The RCIC turbine governor valve was binding while traveling to the closed position in its unsuccessful attempt to control turbine speed. The governor valve linkages indicated slight wear and binding.

(1-4,RCIC)

Susquehanna-1 EG In response to a low RPV water level RCIC quick start testing 83-051 signal, the turbine tripped in an attempt frequency was increased to 5/22/83 to start. Analysis of the transient monthly to ttend, control RCIC monitoring system showed that slow system performance.

operation of the governor valve caused the overspeed trip; the governor valve took two seconds, instead of one second, from full open to full close on the start of the turbine. A visual inspection of the oil showed small metallic particles which

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions were attributed to the initial testing and wear in the turbine components. The unit was successfully manually started immediately after the overspeed.

(1-5,RCIC)

Susquehanna-1 EG The turbine tripped during an operability The governor valve was reworked to 83-103 test of the RCIC system. The turbine updated vendor specifications and had tripped two days earlier when the the system successfully retested.

7/7/83 RCIC system initiated on low RPV level follow-ing a scram. Based on Terry Turbine, clearances between the governor valve and bonnet guide sleeve were measured and found restrictive. With the tolerances too low, the governor valve closure response time could be increased, resulting in an overspeed trip.

(1-6,RCIC)

Susquehanna-1 EG RCIC initiated on low RPV level, and the (1) The governor valve linkage turbine tripped 3 seconds later on travel was reduced by one-quarter 83-120 8/28/83 electrical overspeed. Slow response of inch to ensure closure before an RCIC the governor valve; did not close in time overspeed condition could occur.

during system start.

control oil, installation of con-trol oil filters, installation of.

blowdown drain in the steam supply line, and verification of steam line drain pot level indication.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (3) Installation of a bypass line around the steam supply valve is planned.

(1-7,AFW)

Palo Verde-1 EG The Terry AFW Turbine tripped when start- Installed a 1" bypass line around Summer 1983 ing from a cold, ambient condition during the steam supply valve to the turbine.

AFW startup testing. Its resolution involved This bypass line incorporated a 5/8" troubleshooting with the assistance of orifice and a solenoid operated valve.

the turbine and governor vendors. Adjust-ments'were made to the speed control system; however, the results of the adjustments proved to be unsuccessful based on retesting.

The problem was eventually corrected by installing a bypass line around the steam supply valve to the turbine.

(1-8,RCIC)

Grand Gulf-1 EG Following cleaning of the stop valve, the (1) As an interim measure, a mechan-

,83-170 turbine began to trip on overspeed during ical stop was installed to limit 11/1/83 start attempts. It was determined that the governor valve position to 40%

RCIC the governor valve was not closing fast of full open. This prevented the enough on turbine start to prevent turbine from overspeeding until oil turbine speed from reaching the trip pressure built up to control the setpoint. The slow closing was due to governor valves.

oil pressure not building up fast enough.

(2) As a final solution, a startup bypass valve was installed around the steam supply valve.

(Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (1-9,RCIC) (1) A modification was made to the Grand Gulf-1 EG The reactor scrammed on a low water level governor valve to restrict it from following condensate booster pump and 84-030 reactor feed pump trips.. Operators were full opening and causing a turbine 5/25/84 unable to manually start RCIC to restore overspeed trip.

RCIC the level prior to the scram. Following condensate booster pump and reactor feed (2) The turbine trip valve was lubri-cated and retested satisfactorily.

pump trips, the reactor water level began to decrease; operator tried twice to manually start RCIC, but both times the turbine tripped. When RCIC was initially started, it operated properly and momen-tarily stabilized water level before tripping on overspeed. It was reset and again tripped on oversped. The turbine evidently oversped on initial startup, but the trip valve did not close instantaneously, thereby allowing RCIC to operate for a short time.

Table 2 Events of Condensate in Steam Line (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (2-1,AFW)

Crystal River-3 PG-PL On initial startup, the turbine of AFW Modified the drain system to ensure 6/2/77 pump EFP-2 tripped on overspeed. the drain capacity.

AFW Condensation in steam supply prevented the governor valve from closing fast enough to prevent an overspeed trip.

(2-2,AFW)

Crystal River-3 PG-PL Following a reactor trip, the AFW pump A bypass line around the steam 77-092 EFP-2 autostarted and tripped on supply valve was installed to prevent 7/17/77 overspeed. Although the turbine had condensate buildup; condensate blowdown AFW started on the second attempt, it can be achieved through this bypass line.

tripped on several subsequent operability tests. The trips were attributed to the condensation accumulated in the steam supply line upstream of the throttle valve.

There were two similar occurrences to EFP-2 on 3/7/77 and 4/16/77.

(2-3,AFW)

. Arkansas-1 EG During zero power physics testing, AFW (1) Steam traps were placed in-l 79-008 pump P7A was actuated on demand signal service and pump P7A operability 6/21/79 and tripped on overspeed. Following a was proven.

AFW reset, the turbine was manually started and tripped on overspeed again. The over- (2) Procedures'were modified to speed trips were due to water in stcam include the traps in the valve supply line. The steam traps upstream line-up.

of the valves were isolated during the previous plant heat up.

l

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (2-4,AFW)

Arkansas-2 EG During mode 1 operation at 50% power (1) Operations initiated a steam 79-055 level, the main feedwater controller line blowdown once per shift.

9/23/79 malfunctioned, requiring auxiliary AFW feedwater. The turbine of AFW pump (2) The steam trap application 2P7A tripped on overspeed upon receipt was evaluated for adequacy.

of a demand signal, resulting in no flow from 2P7A. The trip was due to water in the steam supply line. The water was blown from the steam line and 2P7A was successfully started three times. ,

(2-5,AFW)

Arkansas-2 EG Immediately following a plant trip, AFW (1) Drained the pump casing and 79-081 pump 2P7A was manually started and the steam header.

9/23/79 turbine tripped on overspeed. 2P7A was AFW then restarted three times to verify operability. The apparent cause was water in the steam supply header to the turbine. It seemed that the steam trap did not work.

(2-6,AFW)

Arkansas-2 EG Immediately following a plant trip, Additional administrative controls on 79-104 AFW pump 2P7A started and then tripped trap isolati,qp were instituted to 12/30/79 on turbine overspeed. Three subsequent prevent recurrence.

AFW starts were successful. A steam trap was isolated due to a blow-thru problem, resulting in condensate buildup in the steam supply line which caused the overspeed trip.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (2-7,AFW)

Arkansas-2 EG The AFW pump 2P7A tripped on overspeed The insulation was replaced and the 81-038 following a manual start during a unit trap was unisolated.

10/11/81 trip recovery. The trip was attributed AFW to condensation in the steam lines which resulted from insulation not being rein-stalled and a steam trap not being uniso-lated after a maintenance activity.

(2-8,AFW)

Davis-Besse PGG Following a reactor scram resulting from (1) Turbine inlet isolation valves85-013 PG-PL loss of main feedwater flow, the #1 located next to the turbines will 6/9/85 turbine-driven AFW pump was started auto- be replaced with control valves.

AFW matically by the steam and feedwater The existing control valves will be rupture control system (SFRCS) on low normally open to pressurize the level in the steam generators. Five supply lines up to the new control seconds after the initial SFRCS actuation, valves.

the reactor operator inadvertently initiated SFRCS for ; team generator low (2) Steam trap capacity of the pressure. This improper manual SFRCS supply lines will be reviewed for initiation started the #2 turbine-driven adequacy. New traps will be added AFW pump and aligned the system running as required.

on the cross connect steam supply lines.

Shortly thereafter, both AFW pumps tripped (3) Control valve opening time will on overspeed, and all feedwater flow to be adjustable.to allow control of the steam generators was lost. Within steam flow during initial stages of about 12 minutes of the AFW pump overspeed turbine starting.

trips, the AFW system and startup feed-water system were recovered and were used to cool down the reactor coolant system.

It was determined that water slugs in steam

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions piping or rapid condensation of steam while heating long, cold steam supply pipelines caused the overspeed trips.

l This was the first time that the AFW turbines were run solely on the cross connect steam supply lines.

(2-9, AFW)

Palo Verde-2 EG During performance of a hot functional (1) Steam trap operability will be July 1985 test, the auxiliary feedwater turbine verified pericdically as a part AFW tripped on overspeed when starting of the system surveillance test, on an actuation signal from a cold condition. This failure was not repeated (2) Verified that steam trap can on subsequent starts when the system was function as design.

hot. Investigation revealed that the steam trap on the steam header supplying steam to the AFW turbine was inoperable, resulting in condensate buildup in the steam supply line. The condensate was forced into the turbine upon startup and caused the overspeed trip.

1

Table 3 Events of Errors in Reset Procedure (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (3-1,AFW)

Turkey Point-3 PG-PL On July 22, 1985, at 12:40 a.m., the AFW (1) Specific guidance has been added 85-021 system received an automatic start to the shutdown procedure for 7/22/85 signal due to a low level in the 3B SG. proper shutdown sequence of the AFW The A and C AFW pumps promptly tripped AFW pump.

on mechanical overspeed, and the B AFW pump trip-and-throttle valve cycled (2) The operator involved was counseled closed because its electronic overspeed on the importance of verifying that setpoint was exceeded. The cause of the the pumps are at rest prior to A and C AFW pump trips was that, in the adjusting the manual governor speed shutdown shortly prior to this autostart, setting knob.

the governor speed setting knob was not cycled after the pumps stopped rotating; instead, it was cycled before the pumps stopped rotating. Therefore, before the governor oil could bleed down, the auto-start signal, at 12:40 a.m., was received and the pumps oversped. The improper operation of the B AFW pump was due to an incorrectly adjusted setpoint. The failure of the A and C AFW pumps was due to inadequate shutdown procedural guidance.

Table 4 Events of Opening of Supply Valve With Trip Valve Open (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (4-1,AFW) (1) Added clarification to operating Trojan EG A high-high level in the C SG initiated an automatic reactor trip and isolated procedures on the correct method to 83-002 shutdown and restart manually.

main feedwater. Then the two AFW pumps 1/22/83 started automatically as designed. The AFW operator, however, thought that the main (2) Provided training to all opera-feed pump was still running and manually tion personnel.

tripped the AFW pumps by shutting off the supply valve from the control room. When ,

i both the B and C steam generator levels decreased to the low-low level setpoint, the operator attempted to restart the AFW pumps. The diesel-driven AFW pump would not start and the turbine-driven AFW pump started but tripped shortly afterwards on overspeed. There was no feedwater flow to tne SGS for 7 minutes. The cause of this trip was a result of failure to reposition the T and T valve before manually opening the start valve on restart.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (4-2,AFW) Issued ins **"ctions for proper Wolf Creek 1 EG With the plant at approximately 13% power, (1) subsequent to the main turbine trip, the sequenci,ig of valves during manual 85-041 starting of the TDAFW pump to all 6/13/85 operating main feedwater pump A tripped, AFW The operator manually started the motor operation personnel.

driven AFW pumps and reestablished feed-water flow to the steam generators. When (2) Post a sign on the control board SG water levels continued to decrease, the next to the starting switch to TDAFW pump was manually started to provide indicate proper operational additional feedwater flow, but it tripped sequence, on overspeed when the operator opened the T and T valve before opening the steam supply valve. The trip was the result of operator error (not using the proper sequence during manual starting). The mechanical overspeed trip was quickly reset returning the TDAFW pump to " operable" status.

4 #

A Table 5 Events of Stop Valve Stroking Time Too Fast (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (5-1,HPCI) The valve was repaired and the Peach Bottom-3 EG Following a scram, low vessel level initiated HPCI operation. High turbine operated successfully.

4/29/76 HPCI steam flow tripped the turbine during startup. The high steam flow was caused by a rapid opening of the turbine stop valve.

(5-2,HPCI)

Brunswick-1 EG While attempting to perform HPCI The spring loaded bypass oil flow 80-029 initiation response test, the stop orifice was adjusted for a proper valve opened too fast, 6 seconds valve stroke time.

4/2/80 HPCI versus 12, not allowing the flow controller to gain speed control before the overspeed condition was reached.

I

Table 6 Events of Turbine Spinning (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (6-1,AFW) The leakage was corrected and the St. Lucie-2 PG-PL The unit tripped due to a generator exciter field failure. The C AFW pump spinning stopped.84-011 11/21/84 turbine received its start signal and AFW tripped on overspeed upon starting.

The cause of the overspeed trip was l

pump spinning as a result of steam leakage past the leaking steam supply valve before the turbine start.

(6-2,AFW) The valve ASV-5 was repaired and, Crystal River-3 PG-PL AFW pump EFP-2 tripped on overspeed several times during a monthly thus, stopped the shaft from rolling.

12/84 AFW surveillance test. The cause for

' the overspeed trip was identified to be steam leakage past steam supply valve ASV-5 causing the turbine to continue to roll at a low RPM before the unit was started.

(6-3,AFW) (1) The steam supply valve was re-V. C. Summer PG-PL The TDAFW pump was discovered to be rolling at approximately 200-250 RPM paired and returned to service.85-026 9/8/85 with no TDAFW start signal present ,, ,

and the steam supply valve indicating (2) Evaluated the potential for a AFW closed. The TDAFW pump was declared turbine overspeed trip when the inoperable. It appeared that excessive TDAFW pump is rolling at low RPM leakage through the steam supply valve prior to turbine start.

allowed the turbine to roll at low RPM.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions The licensee conducted an evaluation of the potential problem concerning the governor's ability to control turbine speed during a start when previous shaft motion was present as in this case.

k g"P e

Table 7 Events of Pump Air or Vapor Binding (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (7-1,AFW)

Zion-2 PG-PL A low-low steam generator level initiated Installed a high-point vent in 2/15/74 automatic start of 2A (turbine-driven) the suction piping to ensure the AFW and 20 (motor-driven) AFW pump. removal of any entrapped air.

Pump 2C started satisfactorily.

Discharge pressure of 2A pump developed; however, the turbine tripped on overspeed. The pump was then restarted twice and, in each case, tripped on overspeed. It was determined that air accumulation in the suction line of the AFW pump was aspirated into the pump, causing fast acceleration of the turbine in which the governor response appeared to be too slow to prevent the turbine overspeed trip.

gY e

Table 8 Events of Trip and Reset Problems of Trip Valve and Overspeed Trip Mechanism (Item)

Plant, LER Number Type of Event Date, Systera Governor Description of Occurrence Corrective Actions (8-1,AFW)

Prairie Island-2 PG-D The trip valve of No. 22 AFW turbine A design change was initiated to 79-029 was found tripped during a surveillance prnvide valve trip annunciation in 11/2/79 test. Further inspection found that the control room.

AFW the trip valve of No. 11 AFW turbine was also in a tripped position.

(8-2,AFW)

Prairie Island-2 PG-D A workman accidentally tripped the A design change is being expedited--

80-013 T and T valve for the No. 22 AFW involving the annunciator for the trip 4/11/80 turbine. The valve was immediately valve and turbine trip mechanisms.

AFW reset, but over the next several hours Both must be correctly reset to clear was found tripped and was reset three the alarm.

times.

(8-3,AFW)

D. C. Cook-2 PG-PL A solenoid is used to actuate the trip Licensee plans to change the latching 80-017 crank to unlatch the trip valve and mechanism such that it will be 3/21/80 permit it to trip by de-energizing the unlatched by energizing the AFW solenoid. Several solenoid coil failures solenoid.

have been experienced leaving the turbine tripped without indication.

(8-4,AFW) . ,

Rancho Seco PG-PL Overspeed trip lever was founo tripped. (1) Daily local check.80-046 The cause was unknown. The licensee's 11/4/80 further review revealed that the posi- (2) Added a pointer device on the AFW tion of the trip lever can be misleading. turbine trip mechanism to make a To determine whether the device is more positive indication.

tripped, requires close examination.

O (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (8-5,AtW)

Sequoyah-2 EG The AFW turbine failed to start on a Modified procedures to require local 82-002 safety injection actuation. Subsequent visual verification of resetting.

1/12/82 investigation revealed that the T and T AFW valve had not been properly reset after an overspeed trip. The valve is remotely reset from the control room, but does not include control room indication to verify the latch has reset properly.

(8-6,AFW)

Salem-1 PG-PL The AFW turbine failed to start following (1) Daily local verification of the 83-034 a low low steam generator level signal latched position.

8/11/83 because the T and T valve was in the trip-AFW ped position. The valve had apparently (2) Design change was initiated to been tripped and left in that position provide a positive indication in following previous maintenance and the control room.

testing activities. The valve trip position limit switch was out of adjustment causing the " trip" indication in the control room to be inoperable.

(8-7,AFW)

Salem-2 PG-PL The T and T valve of the AFW turbine (1) Installgd local position 83-056 was found in the tripped position. indication.

10/5/83 The valve had apparently been left AFW unlatched following completion of (2) Planning to install indication routine surveillance testing an hour of the " latched" condition in the before. Investigation revealed that control room.

the operator who relatched the trip valve during the surveillance testing (3) Conducted on-shift training in

(I tem)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions had not completly reset it due to the operation of the trip valve.

operation and lack of local position indication.

(8-8,AFW)

San Onofre-3 EG The AFW pump failed to start on an 83-099 automatic actuation signal following a 10/31/83 reactor trip. The operator's investi-AFW gation found the pump turbine in a tripped condition. The pump had previously been satisfactorily tested on October 30, 1983. The cause of the pump failure was unknown. The licensee indicated that although instrumenta-tion is available to signal in the control room when the turbine trips, it is possible to relatch the trip mechanism in such a way that the over-speed trip is reset, but the valve is not and the condition is not indicated in the control room.

1 (8-9,AFW) Indoctrination training for contractor Palisades Elliott During plant normal operation, the 81-007 Turbine overspeed trip device was found in workers is bging. revised to strecs J

the tripped condition. Subsequent necessity to exercise caution when 1/22/81 working in vital areas.

AFW investigation revealed that a contractor's worker accidently bumped the overspeed trip and caused it to trip.

1 (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (8-10,AFW)

Point Beach 2 PG-D A contractor accidentally tripped the All contractors were reminded to use 81-05 turbine overspeed trip valve. At the caution when working around safety-6/25/81 time of the trip, the 3D diesel related equipment.

AFW generator was out of service, meaning one auxiliary feed pump would not operate in an accident condition coincident with loss of ac power.

(8-11, AFW)

Farley 2 EG During routine shutdown operation for The trip linkage was reset and the 82-044 a scheduled refueling outage, the pump became operable.

10/23/82 overspeed trip linkage was found AFW in the tripped position. No apparent cause for the pump trip could be determined.

Table 9 Events of Worn or Loosened Nut and Linkage (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (9-1,HPCI) With assistance of Terry Turbine, Brunswick-2 EG The turbine tripped at 3500 rpm the two screws were placed in their 79-033 instead of the 5000 rpm setpoint during testing for system operability. The current position.

5/18/79 loosened set screw and spring adjust-HPCI ing screw in the trip assembly caused the setpoint to drift. The set screw was five turns out of position. The spring adjust screw was six turns out of position.

(9-2,HPCI) (1) New improved replacement parts Browns Ferry-1 EG The HPCI failed to meet the flow re- were installed.80-045 quirements due to inadvertent tripping 6/2/80 of the stop valve. Full hydraulic pres-sure sufficient to open the stop valve (2) Unit 2 and 3 have also been HPCI changed, did not develop due to a worn mechanical overspeed trip piston.

(9-3,HPCI) (1) The connector was tightened.

Browns Ferry-3 EG The turbine tripped on autostart during fast start test. Loose cable connection 81-024 at the electro-hydraulic governor regulator (2) Inspection manual will be re-5/14/81 (EGR) actuator prevented proper EGR vised to,, verify the tightness HPCI operation. of connectors.

(9-4,HPCI) The turbine tripped ten minutes after The ball and tappet holder was Vermont Yankee EG it started during vibration testing. replaced and piston was reset using 83-018 Normal vibration loosened the set a cotter pin in place of the set 8/9/83 screw which holds the piston on the screw as recommended by GE SIL 353 HPCI ball and tappet holder which is part of the hydraulic trip mechanism.

4 (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (9-5,HPCI) Replaced the broken connector.

Pilgrim 1 EG While parforming a routine operability 85-008 test, the turbine tripped on overspeed.

3/31/85 The cause was the result of a faulty HPCI connector. The cable between EGM and EGR actuator became disconnected when a retaining ring at the EGR side was broken.

It was considered an isolated case. The connector was manufactured by Amphenol.

(9-6,HPCI) Tappet nut was rotated for proper Browns Ferry-2 EG The RCIC turbine tripped while in opera-tion to remove decay heat. The flat alignment and checked for tightness.

I/16/75 surface on the tappet nut head rotated RCIC out of alignment with the head lever.

Vibration from turbine operation caused the tappet nut to disengage with the 4

head lever and trip the turbine.

(9-7,RCIC) The tappet nut was filed to have l EG The RCIC turbine kept tripping at 3500 rpm Quad Cities-2 while perfonning monthly RCIC surveillance sufficient contact surface.

8/30/75 RCIC testing. The face of the tappet nut, which contacts the head lever, was worn " -

sufficiently to give the premature over-speed trip.

(9-8,RCIC)

Hatch-1 EG Following a scram, the turbine was manually 76-019 started and was tripped by mechanical over-2/22/76 speed. The trip was believed to be caused RCIC by the trip valve not being properly latched and the vibration induced during turbine

l' (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions startup which caused the valve to unlatch and trip the turbine before actual overspeed occurred.

(9-9,RCIC)

Hatch 1 EG Turbine start following'a scram was 3/5/76 tripped by mechanical overspeed trip.

RCIC Restart was successful after a reset.

The trip valve may not have been latched properly and vibration on startup caused the valve to unlatch.

(9-10,AFW) (1) Linkage was tightened.

Prairie Island-2 PG During surveillance testing at hot shut-77-038 down, No. 22 AFW turbine tripped on overspeed during two start attempts. (2) The annual test will be run 11/10/77 monthly for three months.

AFW Loose linkage on governor to pressure compensator was the cause.

(9-11,RCIC) The ball and tappet assembly was Quad Cities-1 EG While performing the quarterly. flow rate i

test, the turbine exhibited repetitive replaced.78-015 5/5/78 premature trips. The trips were due to

RCIC a dirty ball and tappet assembly.

(9-12,RCIC) The seating surface of the tappet j

~

Quad Cities-1 EG The turbine tripped prematurely during nut was filed flat to achieve sharp 78-024 the system flow rate test. It was discovered that the knife-edge on the edges.

8/7/78 tappet nut head became rounded so that RCIC a slight vibration would release the head lever and trip the turbine.

l l

r .

(Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence 1

(9-13,RCIC) The turbine tripped on overspeed during Replaced wheel nut, bearing, and Browns Ferry-2 EG flow test. There was a loss of speed actuator.

/9-013 signal in the control room. Improper 5/30/79 initial assembly allowed the turbine wheel RCIC nut to loosen destroying threads and causing axial loading of bearings.

Bearing wear allowed the shaft-to-strike the pick-up causing speed control mal-function.

(9-14,RCIC) The turbine tripped on mechanical over- Replaced connector and clearance on Brunswick-1 EG speed following an automatic initiation magnetic pickup was reset for a proper 79-089 fit.

signal following a scram. Subsequent 11/20/79 investigation revealed a crack in RCIC the Amphenol connector to the tachometer pickup causing a loose connection.

(9-15,AFW) The turbine tripped prematurely below Adjusted the nut to improve contact Millstone-2 PG-PL with the head lever.80-025 the setpoint during operability testing.

Normal wear of the tappet nut reduced 7/12/80 holding force on the head lever such AFW ,.

that normal vibration caused dis-engagement and tripped the turbine.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (9-16,AFW)

North Anna-2 PG With Unit 2 at 100% power, the AFW (1) The worn parts were machined 81-019 turbine overspeed trip valve tripped to obtain proper linkage clear-2/14/81 twice. Improper alignment of tappet ances and were reinstalled.

AFW nut to head lever due to wear caused the unexpected trips. (2) The new parts are on order.

(3) The linkage will be checked yearly as recommended by Terry Turbine.

(9-17,AFW)

Sequoyah-1 EG The AFW turbine tripped on electrical The trimpot was readjusted.81-035 overspeed after an inadvertent pump 4/3/81 start. The trip occurred when a locking AFW nut on a trimpot in the control circuit vibrated loose and allowed the setpoint to drift.

(9-18,RCIC)

Duane Arnold EG The RCIC turbine tripped on its first (1) New parts were on order.85-035 manual start while running a surveillance 9/4/85 test. The cause was determined to be a (2) The interim measure was to RCIC worn linkage of the overspeed trip mecha- machine,,the, parts to restore nism. The contact area between the tappet contact area.

nut and the trip lever was found to be worn, diminishing the contact area between them.

the tappet, itself, was also found slightly loose. This lack of contact area was the cause of the mechanical overspeed trip being susceptible to vibration induced actuation.

Table 10 Events of Misadjustment, Miscalibration or Drifting Controller (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (10-1,HPCI) While an operator was attempting to Pilgrim-1 EG start the HPCI to perform the pump 1/16/73 operability test, the turbine tripped HPCI on overspeed condition. The test procedure used was inadequate.

(10-2,RCIC) The RCIC turbine was operated on The turbine was inspected and the Peach Bottom-2 EG several occasions to support startup linkage was readjusted.

12/24/73 testing of the RCIC system. On three RCIC different occasions the turbine tripped for no apparent reason. The cause was an improper adjustment of the mechanical linkage between the trip valve and over-speed trip nechanism, placing the trip valve in a par cially latched position s sch that normal turbine vibration could cause the valve to trip closed.

(10-3,RCIC) The turbine tripped on overspeed during Adjust servo to align with the Browns Ferry-1 EG routine initiation surveillance testing. governor lever.

11/15/74 .. .

RCIC The remote servo (which controls the

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions governor valve) was out of alignment with the governor lever. Binding of the governor valve linkage slowed the response time of the remote servo and allowed the turbine to overspeed.

(10-4,RCIC)

Brunswick-2 EG The turbine tripped on overspeed after 77-027 being started to restore reactor level 5/7/77 following loss of feedwater pumps. The RCIC cause was failure to follow procedures.

i The operator did not check the manual i

speed setting on the RCIC flow controller to ensure that it was set to minimum speed before opening the steam supply valve as stated in the procedures.

(10-5,HPCI) The control was recalibrated with Peach Bottom-2 EG During testing of HPCI auto start capability at 1000 psig following a proper limits.78-009 2/7/78 unit startup, the HPCI turbine tripped HPCI on overspeed. The turbine speed control had been miscalibrated. Following maintenance, the speed control was calibrated using a manufacturer's general procedure which included an incorrect hand-entered calibration .

limit.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (10-6,HPCI) Governor was calibrated and opera-Peach Bottom-3 EG A loss of feedwater resulted in a reactor scram. HPCI initiated, but bility verified.78-021 11/6/78 delivered unstable flow. Within HPCI minutes the turbine tripped. The governor went out of calibration resulting in unstable flow and overspeed trip.

(10-7,AFW) The governor speed control was reset Salem-1 PG-PL During recovery from a reactor trip, the AFW pump started as required, to the specified value.78-083 12/13/78 then tripped on overspeed. The speed AFW control setting of the governor was found too high, 4200 rpm instead of 3835 rpm as specified by the manufacturer.

(10-8,HPCI) Ball and tappet assembly were Browns Ferry-2 EG The turbine tripped several times on mechanical overspeed during surveillance replaced. The trip weight was79-015 properly set.

6/7/ 79 testing. The trip weight had been HPCI improperly set, causing damage to the ball and tappet assembly.

(10-9,AFW) (1) The manGal speed control was Farley-1 EG The turbine tripped on overspeed during surveillance testing. The overspeed was adjusted and verified.79-051 11/3/79 due to a manual speed misadjustment as a result of lack of position markings. (2) A modification will be incorpo-AFW rated to add a knob with a pointer and a scale to allow resetting to the proper value after maintenance in testing.

I

O (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (10-10,AFW) The turbine trip on overspeed was (1) The governor valve was reset; North Anna-1 PG the governor valve setting adjusted 80-078 apparently due to an incorrect to its proper setting.

9/8/80 governor setting. The governor valve was reset. Next, while verifying AFW initial conditions for a periodic test, (2) Indoctrinating operations the governor setting was found in- personnel to the importance of correct again, proper setting.

(10-11,HPCI) The valve was throttled.

Cooper EG During testing, the HPCI turbine 81-014 tripped on overspeed prior to reaching rated system flow. The occurrence was 6/8/81 caused by an improper valve line-up HPCI procedure. The oil supply valve in the hydraulic trip system should be throttled to maintain about a 20 psig limit on oil supply instead of keeping full open as stated in the procedure. Excessive oil pressure as a result of full opening of the oil supply valve caused the premature turbine trip.

(10-12,RCIC) The turbine tripped on mechanical over- The pilot valve plunger was properly Vermont Yankee EG speed during monthly surveillance test. adjusted.81-024 The EGR had failed to regulate the turbine ,, ,

8/19/81 governor valve. This was caused by the RCIC pilot valve plunger on the EGR actuator assembly being out of adjustment.

i

9 (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (10-13,AFW) The turbine tripped on overspeed during The existing governor was damaged North Anna-1 PG testing. The governor was incorrectly during subsequent adjustment. A 83-011 set and caused the trip. The speed new governor was installed and 3/4/83 setting was too high. Subsequent properly adjusted.

AFW investigation revealed that the gover-nor high-low speed stop.was adjusted back to a position thought to be its "as found" position following completion of the previous testing.

(10-14,RCIC) The RCIC turbine tripped on overspeed (1) The high limiter on the flow Cooper-1 EG controller was readjusted to 50 mAe during startup testing. The flow control-5/30/74 ler was found to have drifted upward from RCIC (2) To prevent recurrence, an elec-50 mA to 54 mA. The additional four mA tronic clamping circuit was instala provided sufficient input to the turbine led to limit the controlled output:

speed control system to cause an over- to 50 mA maximum.

' speed condition.

(10-15,AFW) The turbine tripped on overspeed during Adjustment.

Sequoyah-1 EG a surveillance test. The apparent cause 81-045 6/7/81 was a drift in the electronics of the speed controls for the turbine, t

AFW (10-16,RCIC) Recalibrated high limit setpoint on i

During surveillance testing, the turbine j Duane Arnold EG repeatedly tripped on overspeed on fast the flow controller.81-022 starts. The cause was determined to be 6/7/81 upward drift of flow controller high limit RCIC setpoint. The flow controller is a GE Model 540-01.

W- __ _ --- _ _ _ _ _ - _ _ _ _ _

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (10-17,RCIC)

Duane Arnold EG The turbine tripped on overspeed on fast (1) The flow controller was recali-start under auto control during the per- brated for better operation.

7/20/85 formance of monthly operability test.85-028 (2) The ramp generator was recali-RCIC The causes were found to be an upward drift of flow controller high limit brated to within acceptable setpoint, and the ramp generator limits.

function reaching completion in approxi-mately half its required time. The ramp generator time discrepancy was attributed to instrument drift. The controller was calibrated once per refueling outage, including the outage that had just con-cluded.

I

.v .

4

t Table 11 Events of Hydraulic Oil Leak or Contamination (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (11-1,HPCI) Drain and clean the oil system.

Monticello EG Overspeed trip during testing following maintenance work. Plastic fragments 7/31/72 were found on disassembly of the oil HPCI inlet port to the pilot valve of the turbine control valve. The foreign substance was believed to be introduced due to inadequate control during in-stallation of the turbine.

(11-2,RCIC)

Brunswick-2 EG Turbine quick-start test failed on 10 events overspeed trip. Control oil had drained down after extended periods between Oct.

and Dec. 1975 of inoperation. The turbine trips RCIC occurred before sufficient oil '

pressure could be developed to actuate the control valve.

(11-3,HPCI) The orifice was cleaned.

Vermont Yankee EG The HPCI turbine failed to start while 2/9/77 testing the system. Excessive oil ,., ,

HPCI pressure within the overspeed trip mechanism, caused the trip to oscillate from a rest to a trip condition. The 4

bleedoff orifice controlling the oil pressure was partially obstructed by scale buildup.

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (11-4,AFW)

Prairie Island 1 PG Turbine tripped twice on overspeed (1) Oil was added.77-015 after starting. The cause was improper governor response due to (2) Revised procedures to check oil 5/1/77 level after each pump shutdown.

AFW low governor oil level.

(11-5,RCIC) Replaced the oil line.

Fitzpatrick EG A broken oil line to the speed 77-028 governor resulted in loss of speed 5/21/77 control.

RCIC (11-6,RCIC) Replaced EGR.

Browns Ferry-3 EG The RCIC turbine tripped on overspeed

/8-032 during performance of flow test. Oil 11/24-78 was observed leaking from the EGR RCIC due to a hydraulic seal failure.

(11-7,HPCI) Installed a new diaphragm.

Peach Bottom-3 EG The turbine tripped during a test. The 81-003 cause of the trip was an oil supply leak.

1/13/81 The oil leak was determined to be a HPCI ruptured diaphragm on the Robertshaw oil pressure control valve. ., .

a 1

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (11-8,HPCI)

Peach Bottom-3 EG An oil leak of three to five gpm was81-009 discovered on the oil supply line to 2-16-81 the overspeed trip mechanism. The cause HPCI was a ruptured diaphragm on the Robertshaw oil pressure control valve.

Tripped on overspeed while performing The oil leak was repaired.

Arkansas-2 EG 81-038 a monthly surveillance test. The trip 10/9/81 was believed to be caused by an oil AFW leak in the turbine hydraulic system.

(11-10,AFW)

San Onofre 2 EG Tripped on overspeed during surveillance (1) Oil leak was repaired.

11/1/82 testing. Subsequent investigation by the 82-143 vendor revealed that the cause was slow (2) Procedure was revised to require AFW governor valve response due to oil leak periodic verification of hydraulic in the electronic hydraulic actuator. Oil level and removal of any accumulated condensate and trapped air.

(11-11,RCIC)

Palisades Elliott Tripped on overspeed while performing (1) Replaced the damaged components.84-009 Turbine surveillance testing. Investigation 9/8/84 revealed that the governor had a (2) The lub,rJcation system for the AFW failed bearing assembly and a bent bearing assembly restored.

spindle. A wick was missing, allowing the oil to quickly drain from the bearing assembly. Thus, lack of lubrication caused the bearing assem-bly failure. The bent spindle was a result of the bearing failure.

-o i

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions j (11-12,RCIC)

LaSalle-1 EG The turbine tripped during testing. The oil system was drained, flushed,84-054 The trip was caused by dirty and/or and new duplex filters were installed contaminated lubrication oil. and refilled with Mobile Vaprotex 9/21/84 light oil.

RCIC (11-13,AFW)

North Anna-1 PG Turbine overspeed tripped upon starting The governor was purged of air and

]

80-087 during a periodic test. Subsequent reset.

2 10/15/80 examination of the governor determined AFW that air was trapped in the governor oil passages.

(11-14,AFW) 4 North Anna-2 PG An overspeed trip occurred during post Air was purged from the control oil

80-099 maintenance testing. The trip was of the governor.

12/10/80 attributed to a governor malfunction AFW as a result of air entrapment in the control oil.

l sv e

G Table 12 Events of Governor Electrical Component Failures (Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence (12-1,RCIC) Replaced transistors.

Brunswick-2 EG The turbine tripped during quick-start 75-140 test. Restarts were attempted twice without success. The trip was a result 11/18/75 of no electrical signal to governor valve RCIC due to two burned out output transistors in the controller.

(12-2,HPCI) (1) The ground was cleaned.

browns Ferry-3 EG During a HPCI startup test. The turbine 9/17/76 cycled between its overspeed trip and the trip reset. The EGM failed due to (2) Installed a new EGM box.

HPCI a grounded connector on the governor.

(12-3,AFW) A new circuit card was installed.

Trojan EG The AFW turbine tripped on overspeed 76-060 while testing after maintenance. The cause was a failure of the speed sensor 10/5/76 AFW circuit card (5430-229A) on the Woodward governor. The cause of the failure is unknown.

(12-4,AFW) Replaced a new diode.

PG-PL The AFW turbine failed to start during D. C. Cook-2 11/6/78 a surveillance test. A shorted zener

!. AFW diode in the electronic overspeed monitor resulted in a constant overspeed trip signal.

(Item)

Plant, LER Number Type of Corrective Actions Event Date, System Covernor Description of Occurrence i

(12-5,RCIC) The turbine tripped prematurely during Failed actuator and transistor were Brunswick-2 EG replaced; unit was calibrated.79-055 RCIC component testing. Moisture in 7/14/79 EGR actuator head caused insulation RCIC failure of coil in the actuator, causing excessive currents through a transistor in the EGM unit.

(12-6,AFW) Adjusted the limit switch.

Farley-1 EG The turbine tripped on overspeed upon 80-017 starting. The overspeed was caused by a limit switch on the steam supply valve 2/22/80 being out of adjustment, allowing the AFW ramp signal in the governor control circuit to demand full opening of the governor valve prior to opening the steam supply valve.

(12-7,RCIC) The turbine tripped on overspeed due The EGM box was replaced and Browns Ferry-3 EG calibrated.

i 80-014 to loss of speed control during unit's normal operation while performing

5/12/80 surveillance test. A diode (VR9) in

' RCIC the EGM box failed due to end of life <- -

(Motorola diode, type IN 28168).

(12-8,RICI) Replaced the failed resistor.

Brunswick-2 EG Duri.ng a flow rate test, the turbine

'l 81-039 tripped on overspeed. Failure of a resistor in the 48 Vdc power supply 4/10/81 caused a loss of turbine speed control.

RCIC i

(Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (12-9,RCIC)

During operability test, the turbine Cleared the ground.

Quad Cities-1 EG 82-005 tripped on an electrical overspeed signal. In fact, the turbine never 3/26/82 PCIC reached the 5150 rpm setpoint (3000 rpm maximum). A ground in the RCIC logic circuit caused low voltage at the electrical overspeed motor relay.

It appeared that the bulb had reached its end of life.

(12-10,AFW)

San Onofre 2 EG The turbine tripped on overspeed during The resistor was replaced.

12/8/82 a test and was declared inoperable. The 82-163 turbine trip was due to a failed resistor AFW in the speed sensor circuitry. It was revealed that overheating caused the resistor failure.

(12-11,RCIC)

The RCIC turbine tripped at a speed The burnt out light bulb was replaced.

Quad Cities-1 EG 85-010 below the setpoint. The cause of this 7/28/85 trip was a failed light bulb in the RCIC RCIC overspeed motor relay. It appeared the bulb had reached its end ,, ,

of life.

(12-12,RCIC)

Quad Cities-1 EG The RCIC turbine tripped prematurely. (1) The leads were cleaned.85-015 The cause was corrosion on the leads 8/30/85 to the overspeed relay lamp in the (2) The recurring problems of this RCIC electronic overspeed trip mechansim. electronic overspeed trip relay prompted a recommendation by GE to remove the trip relay.

l (Item)

Plant, LER Number Type of Event Date, System Governor De:cription of Occurrence Corrective Actions I

(12-13,RICI)

Cooper EG During surveillance testing, to prove Repaired the connections.

, 8b-008 auto initiation operability, the

8/24/85 turbine tripped on overspeed. It was HPCI determined that electrical connections between the governor control unit EGM and EGR were reversed, causing the governor valve to fail full open.

J l

.I l,

i l

i Table 13 Events of Unknown Failure Mode (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions (13-1,HPCI) Following the initiation of STI-31 at Browns Ferry 1 EG 25% reactor power, the RCIC system 11/15/73 failed to operate when manually HPCI initiated. The steam supply valve would not open. The HPCI system was then manually initiated to restore vessel water level to normal. Upon initial actuation, the HPCI tripped. No apparent cause of the HPCI system trip has been determined.

(13-2,HPCI) The unit was operating at 92% power when Browns Ferry 1 EG the HPCI system was initiated due to a low reactor water level following a scram.

l 5/5/74 HPCI The system attempted to start but tripped on overspeed. It was placed on manual i

control and was started. The cause of

overspeed trip could not be determined.

(13-3,AFW) While performing a surveillance test, Procedures were modified to include f the 1A TDAFW pump failed to start. the checking of valve position in the Zion 1 PG-PL regular surveillance test.

5/14/74 The IB AFW pump was started and was AFW proved operable. Subsequent investiga-tico revealed that the T&T valve of the . ,

1A AFW turbine was tripped closed. This valve was immediately latched open and the turbine restarted successfully. The cause of trip was unknown.

(13-4,RCIC) During surveillance testing with the Pilgrim 1 EG reactor operating at 70% power, an 4

O 4

(Item)

Plant, LER Number Type of Corrective Actions Event Date, System Governor Description of Occurrence overspeed trip of the RCIC turbine 6/8/75 occurred during startup. Preparations RCIC were initiated far a unit shutdown. An hour later the system was satisfactorily tested. The preparations for a shutdown were stopped. No apparent cause of the trip was evident.

(13-5,RCIC) During a operability test with plant at Monticello EG normal operation, the RCIC turbine 9/1/76 tripped on overspeed. It operated successfully on three subsequent tests.

RCIC Four weeks later a second overspeed trip occurred again. The exact cause could not be determined.

(13-6,RCIC) Following a reactor scram, the RCIC EG turbine tripped on overspeed and the Brunswick 1 2/13/78 T&T valve could not be reset. Since the HPCI system was operable, the RCIC turbine RCIC trip has no safety significance. Subse-quent licensee investigation could not determine the cause of trip. The investigation included checking of the turbine speed control circuitry and the mechanical component linkage. ,., ,

(13-7,RCIC) While performing a maintenance request EG functional test on the RCIC turbine, the Hatch 1 turbine tripped on overspeed. At this75-074 time, the HPCI was operable. A subsequent l 9/6/75 operability test on the RCIC turbine was RCIC l

l

o (Item)

Plant, LER Number Type of Event Date, System Governor Description of Occurrence Corrective Actions performed and the turbine operated success-fully. The licensee could not find the cause of failure.

(13-8,RCIC) With the unit in Hot Standby after a low Hatch 1 EG reactor water level alarm and scram, the 80-050 RCIC turbine tripped during a manual 5/14/80 start attempt. The RCIC system was RCIC declared inoperable. A 7-day LC0 was initiated as per Technical Specification requirements.

(13-9,AFW) With unit operating at 75% power the Zion 1 PG-PL overspeed trip valve on'the 1A AFW 6/6/74 turbine was found in the tripped AFW position by the operator during a tour of the plant. The trip valve was reset and an attempt was made to run the turbine, which operated normally. The reason for the valve trip was not known.

(13-10,AFW) During a reactor trip, the TDAFW pump l

Arkansas 1 EG automatically started per design but l

80-021 tripped on overspeed. Immediate action

! was to start the motor-driven AFW pump,

! 6/24/80 which also functioned as per design. The . ,

AFW cause of the overspeed trip was postu-4 lated to be spurious, since during the transient it successfully started twice l as needed. Later, the pump was proved l

operable per monthly surveillance i

. c.

b

(Item) Type of Plant, LER Number Corrective Actions

Event Date, System Governor Description of Occurrence 1

1 i

procedure before returning to power operation. The cause of trip could not be determined.

4 Following manual reactor trip at 40%

(13-11,AFW) power, the AFW turbine trip valve North Anna 1 PG-D tripped closed. This made the TDAFW 79-042 pump inoperable. The trip valve i 3/30/79 tripping closed was due to overspeed

! AFW trip signal from the turbine. A subsequent performance test was conducted and the TDAFW pump was started successfully. The cause of the malfunction could not be determined.

! The TDAFW pump (2P7A) tripped on over-

! (13-12,AFW) speed following a manual start signal.

Arkansas 2 EG

The motor-driven AFW pump (2P7B) was-80-030 verified to be operable. Following the

! 5/20/80 trip, the AFW turbine was reset and AFW successfully started two times. Subse-quent investigations did not , reveal the i exact cause of the turbine trip.

'r Immediately following a plant trip, the (13-13,AFW) turbine for AFW pump 2P7A tripped on Arkansas 2 EG overspeed immediately after starting 79-72 upon receipt of a safety demand signal.

8/18/79 The motor-driven pump, 2P7B, was then l

AFW started manually to provide feedwater.

The turbine was tested successfully after the overspeed trip. No cause could be found. -The licensee assumed that the i

4 e -

n 0

(Item) Type of Corrective Actions _

Plant, LER Number Description of Occurrence Event Date, System Governor trip probably was caused by a sticky linkage on the turbine governor which restricted the response of the governor valve.

The governor was replaced with a new With plant at 100% power,.an overspeed one and the TDAFW pump was returned (13-14,AFW) trip of the TDAFW pump occurred during to operable status.

North Anna 2 PG-D testing.

The turbine governor failedThe to 81-010 control the turbine speed properly.

1/13/81 cause of the governor failure could not AFW be determined.

i While performing TDAFW pump operability (13-15,AFW) tests, the TDAFW pump was declared EG inoperable when it tripped on overspeed.

Farley 1 81-039 The cause of the overspeed trip was not AFW determined but the immediately subsequent Two test was successfully completed.

follow-up tests were also performed with-out trips.

l i