Text

Safety Evaluation Supporting Amend 48 to License NPF-7
	ML20116N538
Person / Time
Site:	North Anna
Issue date:	04/25/1985
From:	; Office of Nuclear Reactor Regulation
To:	;
Shared Package
ML20116N516	List: ML20116N538;
References
	NUDOCS 8505070246
	Download: ML20116N538 (28)
	v • d • e

___

[/p uog D

o

! ~-

UPHTED STATES

{

NUCLEAR REGULATORY COMMISSION a

k..... p',I wasmucToN. o. c. 2osss SAFETY EVALUATION BY THE OFFICC OF NUCLEAR REACTOR R RELATED TO AMENDMENT NO. 48 TO FACILITY OPERATING LICENSE NO. NPF-7 VIRGINIA ELECTRIC AND POWER COMPANY OLD DOMINION ELECTRIC COOPERATIVE NORTH ANNA POWER STATION, UNIT NO. 2 DOCKET NO. 50-339

1. INTRODUCTION By letter dated February 1,1985, the Virginia Electric and Power Company (

i the licensee for the North Anna Unit 2 nuclear power station, requested a ch i

to the plant Technical Specifications to reduce the required testing of the emergency diesel generators (EDGs).

plant had experienced two failures of each EDG.During December 1984 and Januar both EDGs became inoperable at the same time and a forced plant shutdow necessary with several days of lost electrical power generation.

1985, the total number of EDG failures for Unit 2 reached the number thatOn Januar requires that all Unit 2 EDGs be tested every three days.

EDG manufacturer, the licensee has concluded that the test starts and rapid loading rate required by the Technical Specifications are a "significant contributor" to the EDG failures.

Subsequent to the February 1,1985 submittal, the licensee experienced an engine failure of one of the Unit 1 EDGs, and another major engine failure of one of the Unit 2 EDGs.

completed within the 72-hour limit, Unit 2 had to be shut down.Becaus 1985, the North Anna Station experienced an actual partial loss of offsite pow On March 23, i

event that affected some safety buses on Unit I and some on Unit 2.

event, the EDGs successfully started automatically and powered the safety buses.

In this f

included information gained from a meeting with the lic Maryland on February 8,1985, a visit to the plant on February 4

the licensee's supplementary submittal dated March 13, 1985, and 13, 1985.

~

7. BACKGROUND 2.1.1 General Description The North Anna Nuclear Power Station is a two nuclear unit station 1

total of four EDGs.

one EDG assigned to each of the redundant electrical power divi i

nuclear plant.

There is no sharing of EDGs between the nuclear plants at the 1

8505070246 950425

{

DR ADOCK O

]39

. North Anna Station.

is sufficient to provide either for safe shutdown loads or for accid The EDGs at the North Anna Station are manufactured by the Fairbanks Morse Engine Division of Colt Industries.

12-cylinder, opposed-piston, turbocharged engines.The EDGs are powered by mode These EDGs are rated for continuous duty at 2750 kW, for 2000 hours0.0231 days 0.556 hours 0.00331 weeks 7.61e-4 months / year at 3000 kW, _and for 30 minutes at 3300 kW.

to be less than 3000 kW.The plant design basis accident loads are conservatively calcula by a pair of pistons in each cylinder, operating in opposite directio a common combustion space fomed by the pistons.

heads, and there are no intake valves and no exhaust valves.Therefore there are no cylind when the upper piston is in a position to uncover these ports. a ignition of the compressed air-fuel mixture, the downward motion of the lower Following the The turbocharger, driven by the exhaust gas, provides fres i

air directly into the engine intake manifold.

turbocnarger raises the power output a considerable degree.The extra power provide lube oil booster for the upper crankshaft lube oil line.Each EDG is system, an electric heater and themostat keep the lube oil atIn the keep-wam maintain a desired viscosity.

130-135*F to The keep-warm circulating pump sends the warmed lube oil through the lower crankshaft line, maintaining lubrication of the bearings in the lower portion of the engine, and back to the oil sump Due to the possibility of getting oil into the firing chamber of these opposed-piston engines, the keep-wam system is carefully adjusted so as not to circulate lube oil to the upper crankshaft line.

(for manual EDG starts) and run for two minutes prior t the engine.

The prelube system provides lube oil to both the lower and upper crankshaft lines.

is used prior to the actual diesel start, i.e. for those simulated e safety feature (ESF) actuation / loss-of-power tests where prelube is not activated automatically, the prelube system is started manually prior to actually simulating the need for an emergency auto start.

the diesel starts immediately and is not delayed in order to provide prelub The lube oil booster system includes a two-gallon supply in an accumulator that has a piston operated by engine starting air.

As the engine first begins to rotate, the 2 gallons of lube oil is fed innediately to the bearings along the upper crankshaft line.

This booster oil to the upper crankshaft line supplements the prelube system.

to each test start was discontinued when the booster system wa Anna, prelubing was not discontinued.

experienced at North Anna).other station do not appear to be directly The term " fast, cold start" has developed as jargon in reference to the These ambient conditions are not necessarily " cold" in one internal engine temperature of the combustion process. s are obviously far below the operating conditions For plants in general, if a licensee elects to

-s r=

w\\

w o

T.-**

- MTd A

g

\\

u. A i_

i

9 '

p.': y air To

\\

7 g' -.,,,,.,.,.M y) i^^}CwER _

7)f T wr 1;;;;;..;;;.,

l i

n k;

.*.w.v.

M"9 l_

7-=

f m

p

[

- < a J

w i

i

[

hg s

q._,

,r p

x,.

ij l

IP '\\[h(( -

rw e

w 6

V j l

]k-ST g

/

.t e

~

L.

WJU@

?

<j

- - 11 r.

.y h

. e o

~l i--

i.

v

,$f r

%dD' e _q ox

-- a s

_ L-fi -

l5anOnNTa

,,_i;

'~'-~

hiekt

- l'[-

i p

- fli u

7 j.

__[ a 7

)

IJ

d#

M.

Mn e

a'? vkl'b

~_

.s c g

/

l

~

~n

.J Ol Qt l

ec 0

c c,an. -

Q

,e as t-2' m

'~

r, c yW n

a 3

g te

\\

4

!!l pm MN 3

i

\\

m fjgua L l

TURB0 SUPERCHARGED j-8410 OP ENOiNE 12 t v-

"5 SECT 10N t:-

a_

.Ia 2

m m.

. provide a keep-warm system for either the jacket coolant system or the lube oil system or both, such a system may become part of the " ambient conditions" if the system is also considered vital in determinino the operability status of the EDG.

Most licensees, we understand, have elected to add such modifications to the EDGs.

If however, the licensee is not willing to accept the loss of and to perform necessary repair actions, such keep-war allowed in determining what the ambient conditions should be for the required test starts.

For North Anna Unit 2 specifically, the keep-warm system was part of the original licensing basis and is required for EDG operability.

i The concept of " fast loading" during EDG testing may also deserve some explanation.

If the total loads of the safety bus were applied immediately as a single block load onto the EDG, it would surely trip.

loads are divided into a number of smaller loads and, during an accidentTherefore situation, these smaller loads are applied in a pre-deter At After a total of 20-60 seconds, the sequencing of the loads would be complete. During monthly testing the EDG is paralleled to the grid and the load is increased more-or-less con,tinuously by the control room operator.

the load is increased to full load in 60 seconds, this ramp increase would If approximate the sequenced loading that would occur in a similar time during an accident situation.

Therefore, the present Technical Specifications require that the monthly test of the EDG include an increase to full load in 60 seconds.

The test loading has been referred to as " fast loading."

fast loading could be contrasted with a " staircase" loading wherein the load The is increased slowly over a few minutes to the first plateau; after several minutes, the load is again slowly increased to the next plateau.

plateaus, the EDG would have been "staircased" to full load, After 3-5 i

2.2 North Anna EDG Failure Experience At North Anna, the two EDGs for Unit 2 are designated "2H" and "2J", based on the designation of the electrical bus that the EDG provides power to.

On December 7,1984, the "2H" EDG was taken out-o a

The "2H" troubleshooting and maintenance in order to prevent an actual failure.

one EDG being out-of-service, the plant Technical Specifications Action Statem Due to requires that the other EDG be test started within I hour and every 8 hnurt thereafter.

and was declared to be inoperable.During one of these Action Statement test st The maintenance of "2H" was inmediately halted; the "2H" EDG was test started (successfully) and placed on the 8-hour test cycle, since the "2J" EDG was inoperable. During one of these Action Statement test starts on December 9,1984, the "2H" EDG tripped and was de to be inoperable.

T3ecifications required the plant to be shut down.Because both EDGs wer a !isted by the EDG manufacturer, the second of the EDGs was returned to on 63cember 11, 1984 The details of the repairs may be found in Reference 6 (LER 34-11 Revision 1).

The plant was brought back on line on December 16, 1984 m-

,,-.--n.

--.-.----.----a.,.-e,

- -,.. - - - - ~ - -

,.,e n..

. On January 13, 1985, the "2J" EDG suffered another test failure, due to high crankcase pressure.

piston and associated rings were replaced. Excessive wear was found on the ri 4

At this point, the total number of EDG failures for the nuclear unit had reached the point that the Technical Specifications require both EDGs to be tested every 3 days. At this time, one EDG is being fast cold tested every day and a half.

On February 4,1984, the North Anna Unit 1 EDG "1H" was dec~lirea' inoperable because it had suffered an engine failure.

liner were leaking cooling and engine damage resulted.The seals of the number 3 cyl On March 15, 1985, the North Anna Unit 2 EDG "2J" suffered an engine failure 4

in conjunction with a high crankcase pressure trip during testing. A large number of parts were replaced (cylinders, pistons, insert bushings, and lower rod bearings). Because the repair actions and subsequent testing could not be completed within the 72-hour Technical Specification requirement, Unit 2 was shutdown on March 18, 1985. The plant was restarted the next day.

2.3 Regulatory Background On a generic basis, the NRC has been and continues to be interested in achieving and maintaining a high reliability level for EDGs.

In 1975, an NRC technical report (then as part of the U.S. Atomic Energy Commission, Reference 10) evaluated EDG failure experiences reported during the period of 1959-1973. During this period, EDG testing was conducted on a monthly basis and during refuelings.

This report concluded that one specific problem dominated the failure experiences and identified it as the starting of the engine.

This experience spawned the general perception that if the EDG could start it would likely continue on to accept load and operate in a reliable manner.,Other contributing failures included the engine governor.

In 1977, Revision I to Regulatory Guide 1.108 " Periodic Testing of Diesel 1

Generator Units Used as Onsite Electric Power Systems at Nuclear Power Plants" was issued.

This revision of Regulatory Guide 1.108 was an effort to improve the routine testing schedule by relating the testing frequency to the number of failures being experienced.

believed to be between 31 days and 3 days.At that time, the optimal test interval wa With the emphasis on EDG starts,.

Regulatory Guide 1.108 indicates that EDGs be " fast, cold" started on test intervals that ranged from 31-3 days in 4 steps.

if the total number of failures for 100 tests of all EDGs for the plant wasUnder R greater than 1, the frequency of start tests was increased for all the EDGs for the plant.

An objective of this testing schedule was to encourage utilities with high failure rate EDGs to take major repair actions to avoid the costs of very frecuent EDG tests.

Subsequent reviews of EDG failure experience (for the 3-year period of 1976-1978) indicated that engine starting was no longer the dominant failure mechanism.

Apparently industry efforts in this area were becoming effective.

s.

. -- -, - +

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

n

. being caused by testing.In July 1982, the first EDG failure occurred th Nuclear Station had failed.One of the Nordberg 3500 kW EDGs at the Brunswick examination determined the failure cause to be fatigue due to e starts.

indications of similar fatigue. Inspection of the other EDGs revealed that concerns expressed earlier that excessive tests would be detWmen In September 1982, a staff sumary of EDG experience presented to the AC suggested that: (a) routine test starts on a 3-day frequency should be eliminated; (b) testing should be focused on identifying unreliable EDGs, then major repair action should be pursued, rather than just more testino (c) when an EDG failure has occurred, an initial test of redundant EDG should be conducted with a followup test about every 3 days to provide increased assurance that a new failure has not occurred.

NUREG/CR-2989. " Reliability of Emergency Onsite Power Plants." This study, based upon data fo 0.8 x 10,gverage EDG fg)flure rate is 2.5 x 10~g 1976-1980,per dem industry concluded that the to 10 x 10"

, i.e.

contributors to this failure ra,te are:an average reliability of 0.975. Major testing or maintenance) at about 0.6 x 10-), (2) human-eEDG being out-of-serv (1

hardware-related comon mode failures at about 0.2 x 10~gror-related and water system (EDG cooling) unavailabilities at about 0.4 x 10'2(3) service

, and found that, while the average time to repair an EDG is 20 hours2.314815e-4 days 0.00556 hours 3.306878e-5 weeks 7.61e-6 months , it varie The study widely. No one cause of EDG failures was dominant.

led by) control and logic problems (14.7%), and followed closely by govern (12.3%, cooling water (11.9%), output breaker and sequencer (10.3%), a start problems (9.1%), and several other causes.

This same report shows that between 1976 apparent fail-to-start rate of 1.05 x 10~gnd 1980, North Anna Unit I had an in 191 tests). During this period, the (based upon a reported two failures and maintenance) at a rate of 1.3 x 10'ge EDGs were unavailable (due to test NUREG/CR-2989 also recomended several actions that could be taken to EDG reliabilities.

These varied from having the governor overhauled on a periodic basis, review of repeat failures to determine root causes, use of corrosion inhibitors in jacket water coolants and upgrading EDG-related procedures, to the installation of an additional EDG.

In December,1983, the NRC issued Generic Letter 83-41, which requested licensees to provide information related to the potential detrimental effects of " fast, cold" test starts.

stated that:

The licensee's response daced January 18, 1984 no detrimental effects due to fast cold starts.""A review of design changes had been implemented which alleviated problems associat fast cold starts.

. staff recuirements to improve EDG reliability.In July 1984, th This NRC action requested that them to propose Technical Specification changes. licensee The generic letter also described a reliability-goal-oriented EDG program developed by the staff and requested licensees to provide coments on this program.

provided a partial response to the generic letter in their letter of September 1984 This response did not request a Technical Specification change related fast cold starts, but instead described the EDG lube oil keep-wam system system, and lube oil booster.

this generic letter on FebruaryNorth Anna provided a supplemental response to 28, 1985 which provided the requested data on EDG test failures in the last 100 starts, and provided the licensee's coments on the staff's proposed EDG reliability perfonnance specification.

1985, the "2H" EDG experienced three test / engine fa The data experienced two test / engine failures.

The comments provided basically the subject of this evaluation). reflected the information included in the 3

EVALUATION With assistance from the EDG manufacturer (Fairbanks Morse Engine Divisio Colt Industries), the licensee launched an investigation of EDG failures that have occurred recently.

Based upon the preliminary findings of this investi-gation (as reported in the February 1,1985 submittal), the licensee identified i

several factors of the plant Technical Specifications as "significant contributor to the engine failures. The factors identified are: (1) the fast start (10-second), (2) the fast loading (60-second), and (3) the frequency of suc tests.

The licensee believes that such stringent tests induce severe thermal transients within the engine and that, when repeated very frequently lead to the types of failures that have been experienced.

i The licensee has stated that-the EDG manufacturer concurs.

Accordingly, the licensee has proposed a set of changes that would accompl the following general objectives.

Statement tests, fast starts First, for routine testing and Action be deleted, except for once pe(r 6 months and during the 18 tests.

Second, the accelerated frequency for routine tests (based upon 4

accumulated number of test failures) would be changed to reduce the amount of i

accelerated testing by basing such testing on a reliab EDG.

overhauls.

Fourth, for Action Statement tests, the number of EDG tests would be reduced.

The March 13, 1985 engine failures have not yet been identified. supplemental submittal states that is identified as a potential cause.

Further, electrical overloading Specification language such as " greater than or equal toThe submittal indicates kw" forces the operators to operate the EDGs slightly above the specified value to ensure compliance.

Further, the licensee has recognized that instrumentation inadvertent overload conditions. inaccuracies and meter reading errors

-- 4

. The licensee has therefore proposed also to reduce the specified load value to ensure that operations at the upper limit of the uncertainty value do not result in exceeding the engine ratings and to specify a non-open-ended ope band for such tests.

actions that are necessary to minimize EDG degrada failures.

The licensee has committed to developing a reliability improvement a direct interchange with the Japan Atomic Power Company.

include review of the Technical Specifications, test procedures, and otherThis inte maintenance information associated with the exemplary EDG reliability experien by the Japanese.

After sufficient information and experience is acquired, the licensee intends to propose additional Technical Specification changes to in the reliability of the EDGs for both nuclear plants at the North Anna Station The licensee has proposed a large number of specific changes to the Technical Specifications to accomplish his obiective of reducing EDG testing.

Technical Specifications.submittals, the licensee discusses the changes in the In the into an order that facilitates ease of understanding and review.For co this evaluation report is structured as follows:

Similarly, Section Page 3.1 General Considerations Manufacturer's r 9

3.1.1 Failure causes (ecommendation 3.1.2 9

Relation to Generic Letter 84-15 experience at Unit 2 vs Unit 1) 9 3.1.3 3.1.4 12 Reliability Improvement Program 12 3.2 Routine EDG Tests 13 3.2.1 How each test is conducted 3.2.1.1 13 Test starts and loading 3.2.1.2 13 Full-load tests vs Accident loads 3.2.1.3 Observation of " slow start" 15 18 3.2.2 How often routine tests are conducted 3.2.2.1 19 Reliability goals and accelerated testing 19 3.2.2.2 Incentive for engine overhaul 20 3.3 Special EDG Tests (Action Statements)

Initial test 21 3.3.1 3.3.2 Followup tests 22 3.3.3 Specific Action Statements 23 24 changes and its basis, our evaluation thereof, and our co specific item.

4

,. ~. _, _

-9 3.1 General Considerations The licensee's technical argument for the case that the required testing is a "significant contributor" to EDG failures seems to be based upon two general points: first, the advice of the EDG manefacturer; second between the engine damaging failures at Unit 2 and the expe,r_thjt contrast ience at Unit 1, with the only identified difference being the testing requirements.

3.1.1 Manufacturer's Recospendation In recent years, we have had technical discussions specifically with Colt Industries on how testing relates to long-term EDG reliability and nuclear plant safety.

We agree that any fast start, fast loading, and operating an EDG -

at or above its 100% rated value may have an adverse affect on EDG reliability in the long term.

Such operations may cause normal maintenance to be needed may lead to premature failures. sooner than originally scheduled or, if the eng However, testing of this nature on some frequency is necessary to demonstrate that the EDG continues to be able to perform as originally specified and as assumed in the accident analysis of the plant's FSAR.

Furthermore, we believe that Colt Industries has come to agree with the staff that, when the plant offs 4te or onsite electrical system is in a degraded state, the immediate concern for nuclear safety must take priority over longer term equipment reliability considerations and therefore special testing of the EDGs is appropriate.

Specifically for North Anna, the licensee states that the manufacturer has stated that the types of wear patterns found on the engine components is the result of either engine overload or excessive fast cold starts.

There is, however, some uncertainty as to specifically which engine component are most adversely affected by fast, cold test starts and fast loadings.

is also some uncertainty re There experienced at North Anna. garding the manufacturer's review of the wear patterns We conclude, therefore, that the manufacturer's applicability to the specific situation at North Anna. advice appea 3.1.2 Failure Causes (Experience at Unit 2 vs. Unit 1)

The second part of the licensee's basis for the request for relief from the Technical Specifications testing of EDGs is the contrast between the failures at North Anna Unit 2 and those at Unit 1.

North Anna Unit I was licensed to operate in November 1977, with Technical Specifications that were standard at that time.

testing requirements of Regulatory Guide 1.108.The Technical Specificatio The requirements for Unit I basis and loaded to >60% of rated output. routine tests are basically The Technical Specifications do not include a rapid loading rate and do not require an accelerated test frequency based upon test failures experienced.

(It is interesting to note that the Unit 1 EDGs have been tested nonetheless at the full-load ra of 2750 kW in virtually every test).

l l

l-

-,---.c-

. The licensee states in the February 1,1985 submittal that the Unit 1 EDGs have "not experienced similar failures." In another part of the submittal, the licensee states the Unit 1 EDGs have " experienced no failures to date."

However, information provided previously by the licensee and reported in Reference 10 indicates that between 1976 and 1980 there were two failure the EDGs at Unit I to start during required tests.

On February _4, 1985, the due to cooling water getting into the number 3 firing cylinder.

During a meeting with the staff in Bethesda on February 8,1985, the licensee provided clarification regarding the history for the Unit 1 EDGs.

During the period of November 1977 to April 1980, the "1H" EJG experienced 10 trips.

Three of these occurred in 1979 and seven during February and March of 1980.

Based upon the 1980 test trips, the licensee increased the setpoint for the high crankcase pressure trip.

In August 1984 the "1H" EDG was inadvertently overloaded to about 3600 kW for about a half hour (i.e., 31% over the 2750 kW rating and 9% over the 30 minute rating).

was inspected with assistance from the manufacturer; however, on Feb 1985, the same EDG ("1H") suffered significant engine damage.

Between November 1977 and February 8,1985, the other Unit 1 EDG ("1J") had only one test failure. This failure (November 14,1984) the EDG output breaker and not a trip of the EDG itself. involved the spurious tripping the licensee also clarified the language of the Februhry 1,1985 submittal.During thi The intent was that up to that time, while there had been " test failures,"

there had been no " engine failures " as had become the case for the Unit 2 EDGs, where internal engine compone,nts were replaced.

During our February 8,1985 meeting, the licensee stated that the EDG manufacturer had informed him that high crankcase pressure trips (on the Unit 2 EDGs) could be caused by less than optimal performance of the crankcase eductor system, which uses air from the turbocharger blower.

The licensee has '.

not fully investigated this yet.

The licensee's February 1,1985 submittal indicates that the only difference between the Unit 1 EDG experience and the Unit 2 experience is the Technical Specifications testing requirements.

Therefore, by deduction the licensee claims that the testing is a "significant contributor" to the December 1984 -

January 1985 EDG failures.

This argt e is attractive by virtue of its simplicity and relationship to ongoing. v ic NRC activities to reduce the severity of test starts.

However, as dist. ssc

%1ow, we are not convinced that the testing requirements are a primary cor.~

hutor to the Unit 2 EDG failures.

In an attempt ti issess the impact that the required testing may have had, we have considered m nature of the corrective actions taken after the October, November, and Du er 1984 failures. On October 19, 1984, "2H" EDG tripped on indicated b'f jacket coolant temperature; the temperature switch was re-calibra M. On October 20, 1984, the "2H" EDG tripped on high crankcase pressure; a leaking air start valve gasket was replaced.

On October 22, 1984, the "2H" EDG tripped after 11 hours1.273148e-4 days 0.00306 hours 1.818783e-5 weeks 4.1855e-6 months , again on high crankcase pressure; the lube oil strainer was cleaned, and "2H" successfully completed a 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months full-load test run.

On November 2,1984, the "2H" tripped again on high crankcase

- pressure; after the crankcase air ejector was cleaned, the EDG tested satisfactorily.

The licensee reported that none of these four failures were considered to be " valid" test failures (because both high Jacket coolant temperature and high crankcase pressure trips are bypassed for any actual emergency start) and that all were due to spurious instrumentation problems.

On December 3, 1984, the "2H" failed to start within the required time due to a problem with the air start distributor.

was found to trip randomly at various points and was replace _d.Also, the cra The licensee then reported that this malfunctioning pressure switch could explain the earlier EDG trips.

On December the Number 10 lower piston rings were found to be shattered.9,1984, "

Shattered rings suggest an original problem with the rings whi thus generating an abnormal crankcase press,ure.ch allowed some blow by and Even though it did occur on the sixth Action Statement test within 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months , shattered rings does not seem to suggest fast start or fast loading rates as likely cau 9, 1984, the "2J" EDG tripped on high crankcase pressure.ses. Also on December-The number 2 and number 3 upper pistons were found to be leaking and the number 11 liner seal was leaking.

as the cause. Such failures do not suggest " excessive fast cold test starts" During our visit to the North Anna Station on February 13, 1985, we examined 1985 engine failure and photographs were made.the parts that had manufacturer's analysis of the failure had not been completed.The licensee stated In our be related to fast cold test starts, but rather appears to have inadequate or degraded lubrication.

jacket seal failed, causing dilution of the lube oil.It appears that the cylinder water Consequently, the pistons and cylinder liner became overheated and scored; there was excessive i

wear on the piston pin bushings and the main bearings.

thermal stresses, the cylinder liner cracked in several places. Finally, due to 26, 1985, In a telephone conference on March orally that the recent failure of EDG "1H" is most likely related to theth inadvertent overload of this engine in August 1984 In the supplemental submittal of March 13, 1985, conclusions of the previous submittal remain unchanged, and that the rootthe li cause(s) of the engine failure have not yet been identified and that electrical overloading is now considered to be a potential cause.

We remain unconvinced that the primary contributor to these type failures is test starts.

It seems to us that the primary cause may be yet to be identified.

The number and types of different weaknesses found on the Unit 2 EDGs (i.e., temperature switch out of calibration, dirty lube oil strainer, dirty crankcase air ejector, air start distributor " problem") may suggest the need for more stringent maintenance. The repeated failures at North Anna wherein cooling water has entered the firing cylinger suggests that the cylinder liner to the water jacket 0-rings may be failing.

This EDG vendor Buna-N to a high temperature Viton material, which has be reducing seal damage due to overheating and overloads.

We understand that w -

---n

. of the North Anna EDG 0-rings are still Buna-N.the same change shown in the licensee's supplemental submittal of MarchAlso we believe that, as 13, 1985, when potential significant inadvertent engine overloads may have occurred.

believe that the primary cause of the engine failures could be any of the three Therefore, we items above, or a combination, or some items yet to be identified.

3.1.3 Relation to Generic Letter 84-15 NRC Generic Letter 84-15 needs to be made clear.The relation three distinct parts, the first and third of whir,h are relevant. Generic Letter 84-15 h The first part encouraged licensees to reduce fast cold test starts; the second asked for data regarding the present reliability of each EDG; the third asked for connents on staff-proposed new testing requirements.

Letter 84-15, the changes proposed by the licensee can be grouped into thre categories.

essentially identical to the first part of the generic letter.The firs The second category of the licensee's proposed changes is some of the features from the staff's proposed new requirements described in the third part of the generic letter.

The third category of the licensee's proposed changes are not related to the generic letter.

The NRC has an ongoing generic action that relates to improving EDG re (GI8-56).

In fact, Generic Letter 84-15 is one of the elements of this action.

The staff has not yet completed its review and determined the optimal changes that should be made to the Technical Specifications on a generic bas Since this matter is still under review, the staff response to a request for Technical Specification changes at this time typically would be to defer the specific situation is unique and warrants prompt treatment action can be taken.

We believe this to be the case for North Anna.

decisions on the generic action are made and are being implemented on all When the plants, any final optimization changes can be made for North Anna at that time.

3.1.4 Reliability Improvement Program

~

In our review, we considered the question of the potential adverse effect of reducing the test frequency in the situation where failures are occurring.

If the tests are conducted less frequently, latent failures may not be detected for a longer time.

actions, the root causes may not be identified.Also, considering the nature of typic the licensee has proposed that a reliability improvement program will beAlthou the following major elements: implemented upon approval of the requested 1.

Performance Monitoring Surveillance 2.

Discrete Frequency Spectra Analysis 3.

Evaluation of Past and Present Practices 4

Slow Start Testing Training 5.

Japanese Experience 6

Nuclear Guidelines for EDG Operation 7.

Maintenance Training c-.

...--.m.e w.---,.

. We note particularly that these elements include trend analyses of key performance parameters, that lube oil will be analyzed quarterly, and that the services of an independent consultant have been acquired to detennine the root causes of past failures.

We have concluded that any potential adverse effect of reduced testing will be offset by the licensee's reliability improvement program.

3.2 Routine Test Starts

.~

Routine surveillance tests can be considered to be those tests conducted periodically when all redundant systems or equipment (in this case, the EDGs) are believed to be operable.

The purposes of routine tests are:

(1) to confirm that the equipment remains operable or to detect previously unknown failures, and (2) to gather additional statistical data to improve our confidence in the.

value we believe to be the intrinsic reliability of the equipment, or to refine this value.

because of failures that have occurred previously. Routine tests wou to special tests perfonned because the plant is in a degraded state. Routine tests ar The proposed changes to the routine tests may be divided into two groups.

first group deals with how (the method by which) each test start of an EDG is The to be conducted.

to be conducted.

The second group deals with how often such 3 test start is The relevant portions of the present North Anna Unit 2 Technical Specifications that are significant are Sections 4.8.1.1.2.a (wherein the test frequency Table 4.8.2 is specified); 4.8.1.1.2.a.4 (wherein a 10-second engine acceleration is specified); 4.8.1.1.2.a.5 (wherein a 60-second loading to full rated load is specified along with a 60-minute full-load run); and Table 4.8.2 and associated notes (wherein an accelerated test frequency is specified, based upon.the total number of EDG test failures in the previous 100 tests, on a per-nuclear-unit basis).

_3.2.1 How Each Test Is To Be Conducted 3.2.1.1 Test Starts and Loading The licensee has proposed that these specifications be revised to provide that each test start be preceded by "prelubrication and any other starting warmup procedure reconsnended by the EDG manufacturer." The licensee has stated that the EDG prior to every planned test start.the current practice at North A The current Technical Specifications do not prohibit such prelubrication, etc. However, the licensee has proposed to state the practice as an explicit statement in Section 4.8.1.1.2.a.4 with this approach.

We concur However, we believe that the statement is applicable to all planned test starts, including those that are part of the 18-month surveillance l

tests (Section 4.8.1.1.2.d).

Some may be concerned that, in order to simulate accident conditions more closely, at least the 18-month tests should be conducted without any precondi-tioning.

we find that a number of actual demands do occur that are no either due to actual loss-of-power situations or.are due to ESF actuations.

,,--,--m-.

,.-,e-~

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

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

- ~

.. It has been suggested that the rate at which a complete loss of offsite power occurs.at a nuclear plant averages out to be on the order of 0.1 per year.

Losses of power to a single bus and ESF actuations that cause EDG starts seems to be even more frequent.

EDGs were fast cold started on two occasions due to ESF actuations.D i'

We believe that " test credit" should be given for these events as demonstrating that the EDGs can. start without any preconditioning.

Therefore, we believe that the preconditioning statement should be a footnote that is made applicable to all planned test starts at North Anna, i

The licensee has proposed to delete the requirement in Section 4.8.1.1.2.a.4 that every routine test start be a fast (10-second) start to full engine speed, and generator voltage and frequency).(i.e., acceleration The licensee is developing procedures, in consultation with the manufacturer, that would have the engine start and accelerate to an initial setting of 450 rpm followed by manual increase of the speed setting to the full 900 rpm over a period of 1 to 2 minutes. As a practical matter, the actual conduct of the test is expected to require the EDG to continue at full speed but not loaded for several minutes (e.g.,15 minutes).

temperature begins to equilibriate. During this time the internal engine The licensee

has also proposed to delete the requirement that every test loading be at a rate such that full load is reached within 60 seconds after synchronization to the bus.

The licensee proposes to replace the " fast load" requirement with a test loading rate of about 10 minutes, in accordance with the manufacturer's recommendation for routine tests.

To supplement these more gradual accelerations and loadings, the licensee has proposed a new section (4.8.1.1.2.a.6) which requires that a fast (10-second) start to full speed and loading to full load in 60 seconds be performed every 6 months.

The requirement to stagger the fast starts amongst the various typesi of automatic EDG start signals has been put'with the fast start tests every 6 months.

The design is such that a simulated auto start signal and a " slow" start are incompatible.

The original intent of the fast cold start specification was to simulate the fJst start of loss-of-power / accident conditions. The manufacturer has recommended slowing the engine acceleration time anytime.the EDG is routinely operated, in order to reduce unnecessary wear and to minimize the thermal transients within the engine.

The NRC Generic Letter 84-15 included as a proposed new requirement that " fast cold" starts be conducted only every 6 months rather than a monthly test.

We agree that the primary purpose of routine (monthly) testing should be to implement the manufacturer's recommended testing and to verify starting and load handlin simulate the design basis accident conditions. g capability, rather than to l

We believe that it is not necessary to define the specifics of this test, but it is sufficient to regulate at the level of requiring only that the manufacturer's recommendations I

be followed in this case.

In view of these considerations, we find this portion of the licensee's proposal to be acceptable.

I

, _...,__...~_...____ _ _ _. _.,_ _.,. _ _., _ _._,..,..,.___ __ _. _

- ~.

The licensee has also proposed to combine Section 4.8.1.1.2.a.5 (the 1-hour load run) into Section 4.8.1.1.2.a.4 (the start test). The purpose of this change is to acknowledge that EDG operation at no-load or light-loads (less than 25%) has been shown to be detrimental.

The manufacturer has recommended that every EDG start be followed by at least 60 minutes at at least 60% load.

The effect of this change is to disallow surveillance test starts without a subsequent power run.

acceptable.

We find this portion of the licensee's proposal to be Although not an explicit part of the proposed Technical Specifications, the licensee has stated that, as part of this periodic power run, engine performance data will be collected after temperature stabilization and the data will be trended across similar tests. The licensee states that the manufacturer has stated that this method is the best type of trend analysis to detect problems.

Further, routine testing without fast starts will still be adequate to establish trends in starting times which would be indicative of the need to initiate preventative maintenance.

critical parameters as a tool to indicate the need for maintenance.We c s

3.2.1.2 Full-Load Tests and Accident Loads 4

The licensee has proposed to change the required full load values for the

" monthly" and 18-month tests.

the EDG load for the monthly test to be " greater than or equal to 2750 Every 18 months, a 24-hour load run is required, the first 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months of which i

i are at the 110% value and the next 22 hours2.546296e-4 days 0.00611 hours 3.637566e-5 weeks 8.371e-6 months are at the 100% (continuous duty) load value than or equ;al to 2750 kW."specifically, at " greater than or equal to 3025 kW" and " grea the monthly test be "2500-2600The licensee proposes that the required load for 18-month load run be "2800-2900 kW" and "2500-2600kW" and that the required kW," respectively.

The licensee believes that the open-ended language " greater than or equal to" has the potential for routine overloading of the EDGs.

Specifying an u limit would eliminate this potential. The licensee believes, moreover,pper instrumentation inaccuracies and meter reading errors introduce an uncertainty tnat of +200 kW.

The licensee states that if the tests are conducted at levels abo've 2750 kW and the uncertainty is adverse, the time until the next routine maintenance is needed would actually be reduced but this change would not be realized by operations or maintenance personnel. On this basis, the licensee proposed a non-open-ended operating band that is reduced by an amount approxi-mating the potential adverse uncertainty but rounded off to the nearest 100kW.

The rounding off is intended to be consistent with the markings on the kW output t

meter, the snellest divisions of which are 100 kW increments.

" greater than or equal to 2750 kW" would be replaced by "2500-2600 kW" andTherefore the " greater than or equal to 3025 kW" would be replaced by "2800-2900 kW."

~ ~ ~ ~ ~ ~ ~ ~ ~ ~

i l

16 -

of the magnitude of the electrical loads that co r

basis accident as compared to the electrical loads specified for perio tests.

During the accident situation, the initial electrical loads may be large but typically are required only for a short time - a feu seconds o few minutes. Further, in a t cantly after the first hour. ypical load profile, the loads decrease signifi-During plant licensing, we have found an EDG sizing to be acceptable if the short-term 2-hour rating of the EDG than the initial peak accident loads and if the long-term 2000-ho the EDG is greater than the long-term accident loads, of From a hardware viewpoint, EDGs in general have multiple load ra i

on an anticipated annual maintenance schedule., The 100% " nameplate,"

, based continuous duty load rating ind value continuously for one y, ear (icates that the EDG can be operated at this The 2000-hour rating generally indicates that the ED maintenance.

one quarter of the time) without having to perform r the end of the year.

be operated at this even higher load for any 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months out of 2 without having to perform maintenance on a special schedule.

accident load is calculated to be 2938 kW.Specifically for Nort ce The EDGs are each rated as follows:

8000 hours0.0926 days 2.222 hours 0.0132 weeks 0.00304 months 2750 kW 2000 hours0.0231 days 0.556 hours 0.00331 weeks 7.61e-4 months 3000 kW 168 hours0.00194 days 0.0467 hours 2.777778e-4 weeks 6.3924e-5 months 3100 kW 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months 3150 kW i hour 3300 kW

~.

duty (8000 hours0.0926 days 2.222 hours 0.0132 weeks 0.00304 months ) rating, they are less than both ous the 2000-hour rating.

ng and by Regulatory Guide 1.108 and the Standard Techn a

s ed specify that, during the monthly test, the EDG is t These duty rating.

the EDG is to be loaded to the 2-hour ratin by 22 hours2.546296e-4 days 0.00611 hours 3.637566e-5 weeks 8.371e-6 months at the continuous duty rating. g for the first 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months , followed to avoid exceeding the continuous duty ratinFor the monthly test, the intent is performance degradation prior to a failure. g on a frequent basis but to detect failure before a second EDG failure is likely to occur.shou testing, the test loads envelope the calculated accident loadsDuring the 18-month accident loads, which might occur once in 10,000 y It is our s

repeated 12 times each year.

the accident loads every 18 months is sufficient.We have determined that simu

,,,_._______.__.-,.-_,.,...__m_..

.r

. It is interesting to note in passing that in Japan the design basis accident loads on the EDGs are simulated similarly on a refueling basis (i.e.,12-18 months).

In view of the general considerations discussed above, we conclude that the monthly EDG tests should avoid exceeding the continuous duty rating.

We concur with the licensee that, in view of past enforcement actions, tWe open-ended

" greater than or equal to" load requirement creates a situation wherein operators will tend to operate the EDGs at a load for which the meter indicates a value greater than the specified value.

non-open-ended load would be more appropriate.Therefore, we agree with the licen Further, for human factors engineering reasons, we agree that an operating band equal to the smallest meter graduation (100 kW) is appropriate for North Anna.

Since the exact value of the load during the monthly test is not critical and since overloading has been identified as a possible cause of the engine failure experienced recently at North Anna, we agree with the licensee that the specified load should be the continuous duty value less an amount approximating the uncertainties.

The licensee has presented an analysis of the instrument inaccuracies and potential meter reading errors.

The analysis includes consider-ation of a 1.5% allocation for potential inaccuracy in the turns ratio of each of the voltage and current sensors, 0.3% inaccuracy in the kW transducer, 1.5%

(kw) meter inaccuracy,1.0% allocation for to calibration drift, and 0.6%

( +25 kW) for errors in reading the meter due to parallax affects.

factors are accumulated, the overall uncertainty is 5.23% or +220 kW.When these licensee has rounded this value to +200 kW. We note that the 1.5% allo The possible deviation in the voltage sensor and another 1.5% for the current sensor are estimated values because the turns ratios have not been measured.

that, once this determination is made and this offset is incorporated into the It seems calibration process for the instrument channel, this inaccuracy is in effect nulled out and no longer a factor.

is an obvious source of potential overload. Inaccuracy in the power level measurement ;

Therefore, it appears that, while the output meter may have been calibrated, the licensee has not taken much initiative to avoid this source of overload.

the present uncertainties at North Anna are not trivial.We conclude that, unfortunately, Corrective actions to reduce this uncertainty are desirable, but have not been proposed and are beyond the scope of this review.

ties, an operating band of an indicatedConsidering the magnitude of th'ese uncerta'n-2500-2600 kW is acceptable for the monthly test and the longer portion of the 18-month load run.

However, for the first 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months of the 18-month load run, we cannot accept the proposed band of 2800-2900 kW. The 18-month test is not a frequent test, and special calibrated equipment could be used if the licensee believes it is vital not to exceed 3025 kW.

The primary consideration is that the speciffeo load must envelope the design basis accident loads of 2938 kW. We would find an operating band of an indicated 2900-3000 kW to be acceptable. At the worse case, if the indicated load were to be 3000 kW and if the licensee's value of uncertainty of 200 kW were to be actually adverse, the true load would be 3200 kW.

If one interpolates the ratings on these EDGs, 3200 kW is at or

-,-,,----,..,,,,,,,-,,-wv,

-a

~,,,n,,.--,-,,-,,,-,-a Y

. below an equivalent 2-hour rating.

In view of our understanding of the ratings, we find that even at worse case an indicated range of 2900-3000 kW does not appear to be intolerable or unjustified. At the other hand, when the indicated load is 2900 kW, the actual load could be as low as 2700 kW, which is below the accident loads.

However, it is most likely that the uncertainties will not be all in either direction and to some' extent will cancel each other.

In view of the potential hazard of engine overloading, we conclude that the range of 2900-3000 kW is a better choice than any other similar band.

However, we are requiring that the target value of 2950 kW be specified.

Therefore, we conclude that a relaxation from " greater than or equal to 3025 kW" to the indicated band of "2900-3000 kW" is acceptable.

3.2.1.3 Observation of Slow Start" As part of our review of this proposed change, we visited the North Anna Station on February 13, 1985 to observe the performance of a required " fast cold" test start, and to observe the performance of the more gradual test start and load-ing being proposed. During the " slow start" test, we noted a few potential problem areas which we asked the licensee to address and improve as appropriate:

1.

EDG inoperability during the test.

2.

Critical engine speeds during the acceleration phase.

3.

Electrical loading sequence.

There has been considerable effort over the years to assure that the EDGs I

remain operable during testing; i.e. able to respond to automatic or manual emergency start signals.

is switched to " local, manual" which makes the EDG inoperable.To con This action is taken prior to reducing the engine governor setting to 450 rpm.

engine has reached full speed, the control is switched back to the controlAfter the the test procedure will be revised to alert operations personne room.

inoperability. We require that:

is identified in the control room in accordance with Regulatory Gui (b) the test procedure be reviewed and revised as necessary to minimize the duration of the inoperability, and (c) operating procedures be revised to include appropriate steps to promptly regain EDG control in the control room without risk to personnel near the EDG, to be used during a loss of power situation.

l minutes / test and we would expect a typical time of about 10 min Considering that this brief period would occur only 1-4 times per month, the added availability (about 4x10-4) is minimal and is more than offset by the expected benefits of the " slow start."

,..,.,v.,--

r

, - " - - " " ' ~ * ' '

' " " ~ ~

. The licensee has investigated the question of critical engine speeds between 450-900 rpm.

Based upon infomation from the manufacturer, the licensee reports that this is not a significant problem.

The demonstration test performed for the NRC on February 13. _1985 did not agree with the EDG manufacturer's recomendation regarding the sequential increases in EDG loading. The licensee did not address this item in the March 13, 1985 supplemental submittal. We require that the loading sequence be revised to agree with the manufacturer's recomendation.

3.2.2 How Often Routine Test Starts Are Conducted 3.2.2.1 Reliability Goals and Accelerated Testing The licensee's proposals in this area relate to Table 4.8-2 which specifies how routine testing should be accelerated based upon test failure experience.

This present table originated from Regulatory Guide 1.108, Revision 1.

The licensee's proposal is to continue along the same general approach as Regulatory Guide 1.108, but to add certain upgrades based upon recent EDG experience.

The first specific proposal is to change the reliability goal from a per-nuclear-unit basis to a per-EDG basis. The primary purpose is to avoid test starting all EDGs due to the failures experienced on one EDG that has brought the plant totaT Tnto a range which requires significantly accelerated testing.

We concur with this purpose.

Having agreed to shift from a per plant basis to a per EDG basis, an appropriate value must then be selected. The revised reliability goal is proposed at 0.95/EDG. The proposed staff requirements described in Generic Letter 84-15 include a reliability goal of 0.95/EDG.

This portion of the licensee's proposal is acceptable.

Along the line of the 0.95/EDG goal, the accelerated testing frequency table has been revised such that when an EDG has experienced five or more failures i

in the past 100 tests, the testing schedule would be advanced from monthly to weekly.

The intermediate step of bi-weekly testing is omitted and the 3-day test interval is deleted. These changes are acceptable.

Previously the EDG manufacturer had stated that testing on a 31-day interval was too infrequent.

The reason given at that time was the need to turn over the machine periodically to assure that oil film was maintained on critical The licensee states that the manufacturer has now revised his position pa rts.

and concurs in 31-day testing. We understand this revised position is based upon the modification that added the lube oil keep warm and prelube systems.

j l

l

~1

i

?

\\ The licensee has also proposed to expand the accelerated test frequency table to provide that accelerated engine testing (weekly) may be required based also on the number of failures of that engine in the last 20 starts.

have occurred in the last 20 starts, this could be a point indication of aIf two failures failure rate of 0.1 (or 0.90 reliability). To enter accelerated testing at this point would provide a better sensitivity to the possibility of abrupt EDG degradation and provide a timely response.

Increasing the -test-frequency would provide a faster accumulation of test data upon which to judge the reliability of the EDG.

This additional data can be used to distinguish between failures which occur close together simply due to random chance and such failures that are indicative of an abrupt decline in the actual reliability. The weekly test schedule would be continued until two conditions have been satisfied.

First, seven consecutive successful tests have been accumulated. Second the failu in the most recent 20 tests have been reduced to one. Seven succe,ssful tests 4

indicate a reliability of at least 0.90/ demand but at only the 50% confidence

~

level.

Continuing the accelerated testing until the number of failures is 1 out of 20 adds further assurance that the EDG has not degraded below the 0.90/ demand level.

Furthermore, the proposed staff requirements described in Generic Letter 84-15 include accelerating the test frequency based upon two EDG failures in the last 20 tests.

this portion of the licensee's proposal is acceptable.In view of these various cons 3.2.2.2 Incentive for Engine Overhaul The final change proposed by the licensee is in relation to Table 4.8-2 and would provide an explicit direct incentive for the utility to take major corrective action on the EDG.

hensive complete overhaul of the EDG that is approved by the ED the EDG would be rebuilt to like-new conditions.

the EDG would become operable after it successfully passed the appropriate surveillance tests one time.

However, in return for the overhaul, the utility -

would receive the benefit of wiping the slate clean of all previous failures on' that EDG if an acceptable reliability can be demonstrated.

With "no previous failures" in the past 20 or 100 tests, the EDG would re-enter the test schedule at the monthly test frequency.

required until either 2 failures in 20 tests or 5 failures in 100 testsAccel i

the table after having a series of failures (i.e., many month occur.

l years), the magnitude of this incentive becomes more obvious.

An engine over-haul would focus on the internal components and therefore not necessarily address the statistically most prevalent failures. Nonetheless, this overhaul is con-sidered worthwhfie.

excessive failures has long been the goal, this concept is accepta The question that remains is how does one demonstrate that the rebuilt EDG actually has an improved and acceptable reliability.

that the reliability criterion be the successful completion of 14 consecutiveThe li j

tests, at least 4 of which would be " fast cold" starts.

probability that an EDG with an actual reliability of less than 0.90/demandStatistically will satisfy this 14-test criterion is no greater than 20% and decreases rapidly

21 with the actual reliability.

The probability that an EDG with an actual reliability of less than 0.95 will satisfy the 14-test criterion is 42% or less.

Said in the converse, if the rebuilt EDG passes the 14-test criterion, the statistics say that the probability that the actual reliability of the EDG is 0.95 or better is about 50% and that the probability that the actual reliability is 0.90 or. better is about 90%.

criterion.to be acceptable.

We, therefore, find the proposed 14-test attempts to satisfy the 14-test criterion.One should be careful however about repeat In such a case the statistical situation changes.

If an EDG passes the criterion on a second attempt, the probability that the actual reliability is 0.95 would be reduced to only about 25%.

the previous test failures could not be disposed of until some is negotiated with the NRC.

In addition to the post-overhaul test starts and 1-hour load runs (T.S. 4.8.1.1.2.a.4), we would like to see a full-load test run for a significant duration be a part of the required reliability demonstration.

The licensee has stated that both break-in load runs and full-load runs wi be included (via the manufacturer's requirements); however, the duration would be case-dependent and related to the degree to which major parts were replaced.

For example, following the replacement of a couple of pistons in December 1984 (which was not even a complete overhaul), the manufacturer required a 12-hour full-load run of this EDG. Therefore, based upon this information and a desire to provide for an appropriate degree of flexibility, we find the condition that each overhaul, including post-maintenance testing, be specifically approved by the manufacturer to be acceptable.

3.3 Special EDG Tests (Action Statements)

When plant systems are in a degraded mode (i.e., certain equipment is temporarily inoperable), continued operation for a limited period of time may be acceptable, i

provided that certain additional or compensatory measures are performed.

the plant is in such a degraded mode, the routine surveillance program is not When sufficient.

For such situations, the Action Statements of the plant Technical Specifications specify the time limit for continued plant operation and the required special actions.

Often the special actions are intentionally specified to be the additional instances of the same type of actions that are performed also as part of the routine surveillance program.

Because of this close relationship, Action Statement special actions are sometimes confused with routine surveillance actions.

they are selected and perfonned (maybe on a special frequency) for q different reasons.

It is important to retain the distinction between routine tests and special tests conducted as part of an Action Statement.

Further, as mentioned earlier, when some plant system is in a degraded mode, the immediate and more important concern for nuclear safety takes priority over any consider-ation of equipment reliability decrements that may occur in the longer term.

. The present Technical Specifications provide Action Statements for the electric power systems (T.S. 3/4.8.1) to cover the following degraded conditions:

(a)

(b) loss of both an offsite circuit and an onsite source. (c) los of the two required offsite circuits, and general the specified special actions are:(d) loss of both onsite sources.

In breaker alignment and indication of power ava(ilability, and (2) for the on sources (EDGs), performing a " fast, cold" test start of every operable EDG.

These actions are specified by reference to the corresponding routine surveil-l lance test requirement. Both of these special actions are required to be conducted initially within the first hour and then every 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months thereafter.

The allowed time to restore the equipment to operable status (i.e., recover from the degradation condition) ranges from 72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months to 2 hourt depending upon the severity of the ::ituation.

If either the special actions are not performed or '

the degraded condition is not recovered within the allowed time, the Action Statements required that the plant be shutdown.

In general, the technical changes proposed by the licensee can be characterized (1) The tests of the EDG would become pre-conditioned starts instead of as:

fast cold starts and would include a 1-hour full-load run. This change is a consequence of specifying the test via a reference to the routine test, which is proposed to be revised as discussed earlier; (2) The time for the initial test of the EDGs would be increased to either 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months or 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months depending upon the severity of the degraded condition; (3) There would be no repeat testing of the EDGs.

The licensee proposes no change regarding the special 4

actions related to the offsite power circuits and no change to the period allowed for recovery from the degraded condition.

3.3.1 Initial EDG Test In evaluating the proposed changes, the purposes of the original actions must i

be reviewed.

The purpose of the initial test of the EDGs is twofold:

determine if the cause of the initial EDG failure to affected redundant equipment, possibly via some com(degradation) has also and to provide additional assurance that they remain operable and hence wouldon m be available if needed due to a further degradation of plant conditions.

These test start requirements were developed at that point when it was load and operate satisfactorily. perceived that if the EDG could start, it wou The test was envisioned as operating the manual start switch on the main control board and verifying 10 seconds later that the EDG has come up to proper speed, frequency, and voltage -- operations that could be accompf fshed easily in far less than I hour.

Since then, greater emphasis has been placed upon evaluating reactor operating experiences, increasing attention on human factors considerations regarding operator actions, and improving EDG reliabilities.

As a result our beliefs regarding the optimal EDG test requirements during Action Statements have become refined.

Especially since the accident at Three Mile Island, greater attention has been given to the more likely plant abnormalities.

There is a heightened awareness that, for a large plant, an abnomal event tends to develop or unfold over several minutes or even a period of hours.

i O

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

_.,,,,,.--...,,--._...n.,,,_,,_-,,-~,.m,,,.n,..,.,,n,.,,,-.----.,,-,,e

,,, n,,

. There is an increased consideration for the operator actions during plant abnomalities.

Operator capabilities (instrumentation) to diagnose and track plant conditions have been increased and priority of operator actions has been given further attention. Actions to assure adequate reactor core cooling may continue over hours and have priority over the testing of redundant equipment.

Operator actions that could distract the operator, or require his time to be spent on lower priority actions, or have the potenttET to unnecessarily generate doubts or operator confusion need to be minimized.

More specifically related to the EDGs, the potential adverse effects of no-load (or only lightly loaded) conditions have become2more fully cppreciated.

With this comes the point that EDG test starts should be followed by a period of heavily-loaded operation (i.e., loaded to greater than 50-60% of full load).

For the situation of one EDG failure, the redundant EDG should be inspected prior to a test start to detect an" external conditions that would indicate that starting the EDG might cause further degradation or damage. Further, we are convinced that in general, fast cold starts have been detrimental and, in the worst case, could cause loss of the EDG when it may be needed most.

When these considerations are integrated, we find that the original purposes of the special EDG testing during Action Statements can be fulfilled without fast, cold starts on such a rapid basis.

Therefore, as a general matter, the type of changes proposed for Action Statements are acceptable (the specific changes are discussed below).

The practice at nuclear plants is that, when the equipment that had been temporarily inoperable becomes restored to operable status, the degraded plant condition is terminated.

Uncompleted and subsequent Action Statement require-ments become non-appifcable when the Action Statement is exited.

In this practice, when an EDG is restored to operable status prior to conducting the initial test of the redundant EDG (i.e., in less than 8 or 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ), this test would not be required.

If the EDG became inoperable for reasons other than preplanned preventative maintenance or testing, the detemination of whether the failure affected the redundant EDG remains important regardless of how soon the first EDG is restored.

to conduct this initial test.

The licensee agrees and says it intends l

indicate this action explicitly.Therefore, appropriate text has be.en added to.

3.3.2 Followup EDG Tests The Ifcensee's proposal does not include any followup EDG tests subsequent to the initial tests.

In a 72-hour period, an initial test (within 1 hour1.157407e-5 days 2.777778e-4 hours 1.653439e-6 weeks 3.805e-7 months ) and a followup test every 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months thereafter would lead to a total of nine EDG tests. The licensee believes this number of tests is excessive.

We tend to agree with the licensee that some relief in this area is appropriate.

The real cuestion becomes: how often should an EDG test be conducted to have reasonable assurance that a new failure does not render the EDG inoperable, when a plant systen is already degraded.

Our belief is that the optimal value might be every 3 days.

If every 3 days were used, no repeat tests i

would be required in a 72-hour period after the initial test at the 24-hour mark.

Followup EDG tests on a 3-day basis would become applicable only if the Action Statement recovery period were to be extended.

n-e.---

3 1

('o 3.3.3 Specific Action Statements Regarding the specific changes to the Action Statements of the Technical Specifications, the licensee has separated the first Action Statement into two statements (this change forces the re-numbering of the Act4en Statements).

The first Action Statement deals with the loss of either an offsite power-circuit or an onsite EDG.

The present Technical Specifications require that all EDGs be tested initially within the first hour and every 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months thereafter, regardless of how recently they may have been previously tested. In that the loss of an offsite source does not directly imply a failure of the EDGs, the Action Statement tests for these two cases should be different.The proposal is that, when an offsite circuit is lost, the EDG would be tested on a special basis only if the most recent routine test of the EDG had not been within the

~

previous 7 days.

If such a test had been successful this recently, the i

little additional assurances to be gained by another test may not be worthwhile.

Further, since the test would involve operating the EDG in parallel with a degraded system, the possibility of losing the EDG during such testino is increased. The EDGs are designed and intended to be standby power sources.

Therefore, the negative implications of such tests are not trivial.

On balance, we believe that, when a part of the offsite power system is lost, the EDGs do become more important and they should be tested if they have not been tested in the previous 24-hours.

A period of 24-hours after the offsite power loss would be provided in which to conduct these EDG tests.

If the offsite power is restored before the 24 hour2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months period expires, the EDG test would not be necessary.

i When an EDG is inoperable, the redundant EDG would be test started and loaded in accordance with the manufacturer's recomendations within 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months .

In view of the general considerations discussed above, this portion of the licensee's proposal is acceptable.

When both an offsite power circuit and an onsite EDG are inoperable, the proposal is to initially test the other EDG within 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months . Since the allowable outage period for this Action Statement is only 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months , followup EDG tests are not applicable.

The most likely actual situation is not that these would both be lost simultaneously but rather sequentially. As part of the equipment is restored, the plant returns to an earlier Action Statement. The licensee has added explicit language to highlight this transfer back to an earlier Action Statement (here, and as part of later Action Statements).

The transfer statement also prevents having to retest an EDG just because a new Action Statement is entered if the EDG has already been tested as part of this Action Statement.

The transfer statement also clariffes the starting time for l

the requirements of the Action Statement being entered. This language provides desirable clarifications. These specific portions of the licensee's proposal are acceptable.

l.

25 -

When two offsite power circuits are inoperable, the proposal is that, if the EDGs are not already operating as a result of the power loss, they would be tested within 8 hours9.259259e-5 days 0.00222 hours 1.322751e-5 weeks 3.044e-6 months .

In view of the considerations discussed above, this specific portion of the licensee's proposal is acceptable.

The licensee has proposed no changes to the last Action StatemEtit'which deals with the loss of both EDGs.

4 SUMMApY The licensee has proposed general reductions in the testing requirements for the onsite emergency diesel generators (EDGs) in the plant Technical Specifications. These changes involve both routine surveillance testing and special testing due to degraded plant conditions. These changes would reduce the severity of each EDG test start and reduce the number of test starts.

These changes continue along the same general direction of existing regul6 tory guidance (e.g., Regulatory Guide 1.108 and Generic Letter 84-15) plus a few new proposals, and are viewed as a further optimization in the process of improving and maintaining EDG reliability.

These changes are plant-specific actions being taken at this time rather than waiting for inclusion in the NRC generic actions currently ongoing. The primary basis for this action is that, in addition to the belief that " fast, cold" test starts are generally detrimental, significant engine failures have occurred at this plant and testing requirements may have aggravated the

~

causes of these failures.

In our evaluation of this matter as stated above, we have concluded that there is an adequate basis for reducing the testing requirements at this time.

We believe that any potential adverse affect of reducing testing (that might reveal EDG weaknesses) at a time when significant failures are occurring will be offset by the reliability improvement program being implemented by the licensee. We find acceptable the licensee's proposal that EDG monthly tests be conducted with the more gradual engine acceleration and loading rate recommended by the EDG manufacturer, and at a lower indicated full-load value.

Howem r, as discussed earlier, improvements in the conduct of these " slow starts" are required. Further, we find acceptable the licensee's proposal to schedule accelerated routine EDG tests based upon the reliability goal of 0.95/EDG.

The incentive for a major engine overhaul is also acceptable. During the 18-month EDG load run, we require that the first two hours be at an indicated load of 2900-3000 kW (vs. the 2800-2900 kW proposed by the licensee) to provide reasonable assurance that the FSAR accident loads of 2938 kW are enveloped.

We find acceptable the proposed reductions in EDG testing during Action Statements with one exception.

In the case that one of the offsite power circuits is lost and an EDG that has not been tested within the previous 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months , (vs. the 7 days proposed by the licensee) it must be tested in the next 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months .

As discussed earlier, certain minor clarifications to better convey the intent of the proposed Technical Specifications are necessary. The exact changes are therefore shown in Enclosure 1.

We have discussed these clarifications with the licensee and he agrees with them. Based upon the information provided by the licensee and in view of the considerations discussed above, we conclude that the enclosed changes are acceptable.

vn..,-

,,-.-n

.,~,,-,,.

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

.e. a

-. ENVIRONMENTAL CONSIDERATION This amendment involves a change in the installation or use of a facility component located within the restricted area as defined in 10 CFR Part 20.

The staff has determined that the amendment involves no significant increase in the amounts, and no significant change in the types, of any effluents that may be released offsite, and that there is no significant_ increase in individual or cumulative occupational radiation exposure. The Commission has previously published a proposed finding that the amendment involves no significant hazards consideration and there has been no public comment on such finding. Accordingly, the amendment meets the eligibility criteria for categorical exclusion set forth in 10 CFR 951.22(c)(9).

Pursuant to 10 CFR H51.22(b), no environmental impact statement or environmental assessment need be prepared in connection with the issuance of the amendment.

CONCLUSION We have concluded, based on the considerations discussed above, that (1) this amendment will not, (a) significantly increase the probability or consequences of accidents previously evaluated, (b) create the possibility of a new or different accident from any previously evaluated or (c) significantly reduce a margin of safety, and, therefore, this amendment does not involve significant hazards considerations; (2) there is reasonable assurance that the health and safety of the public will not be endangered by operation in the proposed manner, and (3) such activities will be conducted in compliance with the Commission's regulations, and the issuance of this i

amendment will not be inimical to the common defense and security or to the health and safety of the public.

Date:

April 25,1985

(

Principal Contributor:

l J. T. Beard h

.s 5.

REFERENCES 1.

Letter, W. L. Stewart (VEPCO) to Harold R. Denton (NRC), February 1,1985 (application for exigent Technical Specification change).

2.

Letter, W.- L. Stewart (VEPCO) to Harold R. Denton (NRC), September 28, 1984 (response to NRC Generic Letter 84-15).

3.

Engineering Report, Fairbanks Morse Engine Division (of Colt Industries),

j File Number VTS-985-082881-OlR, dated August 28, 1981, " Engine Prelube/

Keepwann System Requirements."

4 Letter, E. Wayne Harrell (VEPCO) to NRC, November 14, 1984 (LER 84-11)

"2H EDG Trips" on October 19, 22, and November 2,1984.

5..

Letter, E. Wayne Harrell (VEPCO) to NRC, January 5, 1985 (LER 84-013)

" Forced Shutdown Required by Technical Specifications Due to Two Inoperable t-l, Emergency Diesel Generators" on December 9,1984 6.

Letter, E. Wayne Harrell (VEPCO) to NRC, December 28, 1984 (LER 84-11, Revision 1).

7.

"Recent Operating Experience With Emergency Diesel Generators," a l

presentation to the NRC/ACRS Subcomittee on AC/DC Power Systems; J. T. Beard, September 8,1982.

8.

NRC Memorandum for R. Fraley (ACRS) to William Dircks (EDO), " Testing

~

l Requirements for Diesel Generators," September 16, 1982.

l 9.

NRC letter from H. W. Lewis (ACRS) to William Dircks (EDO), " Diesel Generator Failures at North Anna Power Station, Unit 2," January 15, l

1985.

s l.

l _

10. AEC Technical Report, " Diesel Generator Operating Experience at Nuclear Power Plants," 00E-ES,002, June 1975.

11. NRC Technical Report, " Enhancement of Onsite Emergency Diesel Generator Reliability," NUREG/CR-0660, January 1979,

12. NRC Technical Report, " Reliability of Emergency Onsite AC Power Systems" NUREG/CR-2989, July 1983.

13. NRC Generic Letter 83-41, " Fast Cold Starts of Diesel Generators" December 16, 1983.

I

14. Letter, W.L. Stewart (VEPCO) to Harold R. Denton, (NRC), January 18, 1984 (Response to Generic Letter 83-41)

15. NRC Generic Letter 84-15, " Proposed Staff Actions to Improve and Maintain Diesel Generator Reliability," July 2,1984.

16. NRC Regulatory Guide 1.108, " Periodic Testing of Diesel Generator l

Units Used as Onsite Electric Power Systems at Nuclear Power l

Plants," Revision 1, August 1977.

i l

l L

g i '

17 NRC Memorandum from Steven Hanauer, Director DST, to D. Eisenhut, et al. " Diesel Ganerator Reliability of Operating Plants,"

May 6, 1982.

18. NRC Memordanum from J. T. Beard to Gary Holahan, " Testing of Emergency Diesel Generators in Japan," February 5,1985. ~~

19. Letter, W.L. Stewart (VEPCO) to Harold R. Denton, (NRC).

1 February 28, 1985.

(Supplemental response to Generic Letter 84-15)

20. Licensee Event Report 50-320/84-018, Rev.1, February 27, 1985.

21. Letter, W.L. Stewart (VEPCO) to Harold R. Denton, (NRC),

March 13, 1985 (S';pplemental submittal to reference #1 above) i s

i

. -. -

ML20116N538

Text

Navigation menu

Search