ML18038A375

Niagara Mohawk Nine Mile Point Unit 2 Event of 910813

ML18038A375
Person / Time
Site:	Nine Mile Point
Issue date:	09/05/1991
From:	Crenshaw M GENERAL ELECTRIC CO.
To:
Shared Package
ML18038A373	List: ML18038A372 ML18038A374 ML18038A375
References
CON-IIT07-751-91, CON-IIT7-751-91 NUREG-1455, NUDOCS 9109110283
Download: ML18038A375 (134)

Text

Niagara Mohawk Nine Mi>e point Unit 2 Event of 13 August ]991 Report by:

Melvin L. Crensbam Consulting Engineer Po~er Systems Engineeriag Department General Electric Company Schenectady, NY 8 September 1991

4 0

5 iagara Mohawk 5 ine Mile Point Unit 2 E vent of 13 August 1991 05:4S Qn August 13, 1991, at 5:48 Abf the Unit 2 phase B generator step-up transtormer failed. Oscillographic records of the event are available from a digital data recorder at the Scriba Substation. They show various 345 kV and 11 V

5 k system voltages and currents.

Figure A with notations is attached.

The four cycles preceding the fault show no signs of a gradual degradation or a developing disturbance.

The oscillographic traces and station protective relay targets reported, indicate a ground fault occurred on the high voltage winding.

Depression of the 345 kV phase 8 bus voltage to about 39% of the prior value was observed from the oscillographic trace.

This suggests the involvement of only a portion of the entire winding. The 345 kV line currents and voltages show rapid development of the ground fault beginning at point 1 with the ground current reaching a constant value of 1,300 amperes in 1 1/2 cycles at point 4, The flashover in the faulted transformer occurs just preceding a maximum in phase 2 to neutral voltage (as would have been expected) at point 2. The 345 kV line current in an unfaulted phase increases in step function manner to 350/o of the prefault value at point 3, No high speed recordings of voltages or currents within the plant were available.

No sequence of event recordings were available to correlate relay operation times.

Due to the large amount of magnetic energy coupling the generator rotor and stator, and known electrical parameters, thc decay of fault current contributed by the generator to the solidly connected transformer would have spanned a number of seconds as thc field decayed.

Relay operation targets reported were:

l. Transformer Differential Relay (Type BDD) on Transformer 2MTX-XM1B.

2. Transformer Neutral Current Relay (Type IAC).

3.

Overall Unit Differential Relays (Type BDD) in phases 2 and 3.

4.

Generator Phase Overcurrent Relays (Type PJC) in phases 2 and 3.

r

Fv Following isolation of the generator and failed transformer from the power grid, marked 5 on Figure A. only a single 345 kV phase to ground voltage record is available.

The magnitude of this voltage on an unfaulted phase is 74fo of the pre-fault value.

Since generator neutral current is limited to less than 8

amperes. it is known that the faulted transformer appears as a line to line fault with some impedance to the generator.

By trial and error calculation, generator.

line currents are found to be 0, 1.9 and 1.9, multiples of the rated value of 31,140 amperes.

The line-to-line voltages have magnitudes 74 lo 74 lo, and 25%

of the rated value of 25,000 volts. The decay of this voltage for 0.2S seconds of the recording has a measured time constant of 2.7 seconds.

The calculated value of the impedance of the faulted transformer as seen by the generator is 0.23 per unit.

Conditions prevailing during the six cycle time period following the fault, marked 2 on Figure A, cannot be determined with certainty.

The exact nature of the fault within the transformer is not known and the physical evidence will be strongly affected by the continued flow of energy from the generator due to the inherent time constant.

The flashover of only a portion of the HV winding is evident since the 345 line voltages to neutral remain at 39%o, 867o and 86'1o of the pre-fault values.

The presence of "residual" in the measured 345 kV line currents provides the evidence of transformer neutral to ground current.

This requires that the fault involves a path for current to ground from the high voltage winding.

Recorded voltages and currents show a step change to new values and no dramatic change during the time period of the record, which totals somewhat less than 1/2 second.

It could be said they are "cleaner" and less distorted than commonly seen oscillograph recordings of faults.

Given these observations and since both the generator and the system were supplying fault current into the faulted transformer, generator line-to-line voltages preceding isolation would be expected to be greater than those immediately followingisolation.

It has been speculated that very high frequency energy (mHz region) may have causal malfunction of logic and control circuitry in the UPS equipment.

A broad-range of frequencies would be expected in any arcing phenomenon such as occurred in this fai1ure.

Nothing in the available data or design parameters of the plant equipment would suggest an extraordinary generation or propagation of higher frequency components.

'Ihe failure of a transformer and internal arcing is not a rare occurrence.

Comparison of oscillographic charts

I jj

from similar events in other plants show nothing unexpected or unusual in this particular failure.

It must be borne in mind that the sampling rate of the recorder is listed as 5.814 kHz and frequency components in excess of perhaps 500 Hz would not be accurately portrayed.

GE experience in testing of typical power transformers (such as the Lnit Auxiliaries Transformers) provides an indication of the expected coupling between windings at radio frequencies in the region of 1 megahertz:

The attenuation factors range from 1,000:

1 to 10's of thousands:

1.

Direct measurements could be made in this plant to determine attenuation factors for individual transformers over a range of frequencies.

These tests would be made on non-energized transformers using an RF signal generator and a

sensitive, calibrated detector.

Attached recent articles on electro-magnetic interference.

Reference 1

discusses IEC 801,4 and the characteristics of electrically fast transients.

Reference 2 discusses testing of ground connections.

V i

The possibility of elevation of thc station grounding system as a result of this disturbance was postulated.

The relatively high level of ground fault current, estimated at 1,300 ampercs from the available recording, would not have been conducted into the plant.

This current can only fiow in Rom the 345 kV system for the 6 cycle period required for relay and circuit breaker operation to achieve isolation.

The generator ground current would have been limited to less than 8 ampcres by the neutral grounding equipmcnt.

Elevation or differences in ground potential within the plant would therefore not have been expected during this event.

Reference 1 discusses the problem of achieving a "super" ground and concludes that a stable ground reference for interconnected equipment is of greater significance.

Since normally circulating ground currents are not

expected, testing with very low voltages and currents is recommended.

Note especially the recommendation to test with a frequency non-harmonically related to the power line &equcncy.

The transformers stepping the voltage down to successively lower voltage levels are connected in a manner to minimize coupling of power frequency and higher frequency.

components between thc various busses.

Specific configurations are:

1 I

I V

l.

Normal Station Service Transformer-delta 25 kV to wye 13.8 kV with 400 ampere resistive grounding on the 13.8 kV side.

2.

Load Center Transformers-delta 13.8 kV to wye 4.16 kU with 4R ampere resistive grounding on the 4.16 kV side..

3.

Load Center Transformers-delta 13.8 kV or 4.16 kV to wye 600 volts with neutral solidly grounded on the 600 volt side.

4.

Reserve Station Service Transformers-wye 115 kV, delta 4.16 kV, wye 13.8 kV, The 13.8 kV neutral is 400 ampere resistive grounded.

The 4.16 kV circuit is connected to a zig-zag grounding transformer with a resistor in the neutral connection, presumably for 400 amperes.

These configurations provide "effectively grounded" distribution busses as defined in TEE Standard 142 and will serve to limit transient over voltages.

This is in accordance with design practices deemed prudent and conservative within the power industry.

The industry continues to review the effects of geomagnetic disturbances on power transformers.

While no evidence is seen of voltage distortion in the four cycles preceeding the failure, excessive duty could have occurred if these transformers had been subjected to low level direct current previously.

References 3 and 4 are attached for perusal.

I

~

~

r T

~

v!L A

LI4>E II> 3 ~

La+ ~I I,ed'-. ~ >J 444 Jape.

2 V

I V'g n

~ I'.!i

'VVV

>. pi~~c I

C.IIIC (IVI. -$

LJIC 9

.c pN il (l

t~

I H

0 L ~

I J

~

h'

~

r

>>t.S k AV CI IOO I>>I>>Is cc rrM Feuds A

JtI>latir~g g (. CrCeShr>>J

I 1

~

~

fncfustrfal Equfpment EISCtt'OfllCS Ill lndLIStl'Igf APPIiCatienS A Discussion of Fundamentaf ENlC Princfpfes for Electronic Controllers ln an industrial Environment By 4t'iHiam 0. Kimmel. PE Kimmei Gerke Associates, Ltd EDDIC problerrs i~:h:ndustra! ccnuois are aggravated by harsh eni.ronments.

mixed

ec.'".noiogies and a

laclc of umform E!ifC guidebnes. T:is arLicie '>el concentrate on

he common as;ec

s of electronic controls in an

!ndustnal en>".ronment.

which is generally mucn harsher:ban the ofGce envirOnmenL What is the industrial enviroment and what can be core about it!The environmen includes the entire gamut of '.he basic threats.

power disturbances.

RFI.

and ESD. RFI and power disturbances may be locally generated or not. Mixed technolo-gies compourd:he problem. Digital circuits are used to switch."ne voltages via relays.

Analog sensors are input devices:o digital controls.

Increasingly.

there is a

need for a

ooperauve effort between the designers.

manufacturers and instailers to come up with a

rock-solid system.

A common complaint is that the mstallers or mainte-narce people won't follow the instaUation requirements.

This may be true, but it must change.

smce there are problems which cannot be solved at the board level.

It is also true that manufacturers often specify installation.equirements which are not practical to impleinent, and there are documented cases where the prescribed installauon procedures wil! cause rather titan cure a probietts.

The!adt of umiken guidelines has ham-pered EMC prtiless in the industriat arena. Fortunately, the European Commu-nity is working to adopt the IEC 801.x specilications.

and domestic companies would be wise to adopt them, even ifthere is no intention to export.

The Basic Threats The three basic threats to industrial electronics are power disturbances, radio frequency interference, and ESD.

Power Disturbances.

Power distur-EMC Test 4 Design bances are a weH known irdustral problem.

In iac:. when a problem occurs. ke tirst though': is:o blame the power company.

Often power quality is a problem (especially if grounding issues are inciudedl. but the problem is almost always generated by adjacent equipment.

Tradiuonal problems with power include spikes and transients, sags and surges, and outages.

which threaten the eiectroiucs via the power supply.

These problems are fairly weil documented and are often solved using power conditioners or UPS.

The most common power problems confronting electronics today is the sag which ~icaily occurs during turn on and the spikes which typically occur during turn off of heavy inductive loads.

~<..e sags simply starve the electronics.

The high frequency ttansients barrel right through the supposedly Stered power supply to attack the electronics inside.

Digital circuits are most vuhierable to spikes which cause data ettors or worse.

Analog circuits are most vukierable to continuous RF riding on top of the power.

FIPS PUB 94 provides guidelines on eiectrical power for commercial computers.

This is good infortnation, but beware that factory power is much noisier than commer-cial power.

The guidehces of IEC 801.4 speci5es an electrically fast uansient (EFT) that simu-lates arang and other high speed noise.

Ebs are quite short ranged they diminish rapidly with distance due to induc.

tance in the line. But at short range, they are devastatmg.

Unfortunately, attention is placed on the front end of the electronics, the power supply. With industrial controls, the prob-lem is the controlled elements.

If the electronics is controlling line power, the disturbances sneak in the back end where little or no protection exists.

System ground, while not being specifl-cally a power disturbance prob'.em.:s:

the car..er of residual eiiects of o-.

disturbances.

Any 'ind;sinai or commer:

suucture has sigiu8cant

.'ow

~equen currents circulating:hrough t.".e grou system.

sometimes because the.nergy intentionally dumped onto the ground (s.

as with an arc weider) and someur..

because of unintenuonal coupling or v

an inadvertent connection between neu and ground somewhere in the!aciiity.

Radio Frequency Interference.

R dio frequency interfer~nce affects bo analog and digital circuits.

with ana'c circuits being generally more susceptibl Surprising to many, the pnncipie threat not the TV or FM stauon down the roai but rather it is the hand held L~snut:~

carried around by facilities personnel. A or.

watt radio will result in an electnc ne!d

~

(Ne volts/meter at a one meter distance enough to upset many electromcs systems IEC 801.3 speciGes immunity to elec ".

Gelds of one to ten volts per mete depending on the equipment.

with tive volts per meter being Se level for typic equipment. As can be seen from the abov approximation, three volts per meter is nc an excessive requirement, and even:e volts per meter is fairly modest.

Electrostatic Discharges.

EiecLc static discharge is an intense short durauoi

pulse, having a

riseame of about one nanosecond.

This is equivalent:o a burs.

of 300 MHs interference.

Static buildup.

of 15 kV are not uncommon.

Dry climates, including northern climate'.

Wham Kimmelis a pn'ncipal with Kimme.

Gerome Associates.

Ltd. The firm special.

i@ca in preventing and solving electromag.

netic interference and compaabQity (E.Mh EMC) problems.

Mr.

Kimmei can reached at 3544 N Pascal, St. PauL.~i~

55108, or telephone 612.330-3728.

i

~

~

Vcc line pouer I

t I

i I

TRIAC suitcl.e piouer F:g.'re '.. Amp'.'Ger detnodulanon.

Figu:e 2. Transient!eedback path.

n

~".nter.

offer opportumty

!or ESD.

Industnai environments.

with the:r moving eq ipment. are loaded with potenual ESD sources:

rubber rogers. belts, and produc-uon ou'.put such as plasnc and paper roUs.

all add up!o a real ESD th:eat.

and t."is r."meat is more likely to occur even m

elanvely moist environments. Look

o IEC 801.2.'or ESD standards.

Elec!ronics Design Electronics:s generally the ultimate victim of:nterference.

The hter.'erence Gnds its way through various paths:o the electronics equipment itself. Let's concen-trate on w'rat can nappen to your electronics from

!he back door.

that is.

by direct radiation into the electronics and by con-ducted:nterference through:he signal and cont:oi lines.

Sensors.

Low level sensors.

such as ther...ocouples.

pressure sensors. etc.. are characterized by very low bandwidtbs and

!ow signal levels. A major Meat to these sensors is radio frequency interference.

either from nearby hand held transmitters or more distance land mobile or Gxed transmitters.

But:hese are high l'requency, much above the bandpass of your amph6er. right?

Wrong!

Low frequency amplifiers are plagued by two,ybeaomena:

out of band response and

<<stso rectification.

These combine to provide false information on levels to the system.

All amplifiers have a normal bandpass, typi6ed by a 20 dB/decade roUof or more at the high end. But resonances due to stray inductance and capacitance willgive rise to amplifier response Gve orders of magnitude or more above the nominal bandpass of the amplifier. This means an audio amplifier will respond to signals in the hundreds of MHz.

The second aspect occurs when RF en cour ters a noriinearity such as a semicon-ductor device. All such devices give rise to a DC level shift when confronted with RF.

ln a radio receiver they are called detec-tors. Noniinearities are minimized in linear devices, but:here is always enough to cause problems. The upshot is that the ampli6er demodulates the RF. generates an errone-ous signal. and passes this error on. This effec'. is shown in Figure 1. Output hnes are similarly affected, with capacitive couphng back to the input.

The soluuon is to prevent the RF from getting to the amplifier. either by shielding or filtering. The most common path to the amplifier is via an external signal ine from the sensor.

but if the ekctronics is not shielded. direct radiation to the circuit board may also present a problem.

Assuming filteringis the sekcted method.

use a

high

~frequency Glter, designed to bkick signals up to 1 GHz or even more.

Use femtes and high frequency capacitors, Do not rely on your low frequency Glter to take out RF.

At the op amp. you shouM also decouple your plus and minus power to ground at tbe chip. Ifyour ground is carzying RF, you can anticipate the same probkm mentioned

above, since it will comtpt the reference level.

Data Lines. Digital data lines wiU be upset by the RF problem as in anakig, but tbe levels necessary to upset are higher.

Instead.

digital data lines are tnucb more susceptible to transient ghtches. All signal lines should be 6ltered to pass only the frequencies necessary for operation. Ifthe threat lies in the bandpass of tbe signal, then shielding or optical links will be needed.

Switched Power Lines. This refers specifically to the power being contro9ed by the controller device. Industrial control-lers are commonly tasked to control power to heavy equipment. wtuch.s;itc. -.

hy heavy starting.'oads and inducnve a<<wn of TyptcaUY the e!ectronic con:r switch Une power using relays or;;.-.

This exposes the back end of the cont;=

to substantial line transients.

which coi-back to the circuit power and ground disrupt the digital circuitry as shown Figure 2.

It is mandatory that the transient:

rents be diverted or blocked.

since

'igital system cannot withstand t.".e.-.~g tudes likely to occur with an inducuve k:i unless special steps are taken.

Self jamniing can be Unuted by contrcli when you switch

!he Une.

using ze crossing devices. Of partic~ importan is the mrn off. since dtat is when inductive kick occurs.

lfa6 power switching used zero crossu devices. the transient levels in the facto wouki be dramaticaily reduced.

Unfor.

nately, that goal is well off in the futu:

Until then, expect that high voltage pow transients wlloccur, and they must be de wldl.

Optical couplers and relays do not provi sufficient isolation by themselves.

Tht high capacitance provides an excegent hii frequency path, and if they are stacked t in an array, tbe capacitance wi6 add up pass surprisingly low frequencies.

The:

capocitances an't be elim'mated, but yc can design yow control circuits to minimiz couphng paths and to maximize low impe<

ance alternate paths.

Transient suppressors should be installe at tbe kiad, which is the source of the spih but they can be installed at the controUt as weal.

An interestmg effect occurs when con bining zero crossing SCR regulators wu low level Mnsors which use line frequerc noise canceHng techniques.

Very seftsitiv sensors sometimes are sampled. for t JQIL"August a~!

i 4

I I'

'f80 VAC High Current DC burro u Pouer I

wlc Carina PS Electroni>>:s Fir're 3. Common industnai powe. supply.

Figure 4. awful:rple ground paths.

enure power yc!e to cancei the ine freq'ncy conipohent. If t..e samp!e occurs co.","u;.ently wi:h ire power sw'.tching on or oif.:he average

o rhe sensor will be

.pse'.. and an e.-.or ~4 "e recorded.

System Design and Installation Once

>e elect:onics is designed.

it becomes a p.oblem of!he system integratar and installer:o ersure that thc electranics is cravided wuh the environment far which it was des;gned.

Most of thc umc.

this work is performed by power experts and clectricians. and they are not always aware of:he intcrfererce probkm. Often, on site.

rhe power quality is blamed for thc equip.

rrent anomalies.

But the problem can often be avoided by following a (ew basic prina-ples.

The industrial conuol device is either integrated into a system at!he factory or instaL'ed separately on site.

Controllers handle a varicty of devices such as motor speed controls. positioning devices.

weld-

ers, etc.

fnterferencc presented to the electronics can be signiGcandy reduced by appropriate measures outside of the elec-uonics box.

There is no way to accurately assess die rhreat without test data. But rcgardlcss of the br(oration avai!ab!c.

much can be accomplished by correct instagation, and it doesn't cost much if done at the start.

RetroGts become cosdy.

especially if ac-companied with factory down time.

Let's consider dsc same prob!cms from a system standpat.

Your goal is to limit the interference wfskh must be handled by the electronics.

Direct radiation to the electronics is not often a problem in an industrial environ-ment, but itdoes occur, and most often with a

plasuc enclosure.

The NEiMA type enclosures provide enough shielding for most mdusuial needs. Ifyou don't want to use a

metal enclosure, be sure

!o get elecuonics which will withstand the RF which willoccur.

l(ore or'.eh

'.he problem is conducted.

eirher 'aa power or graurd. The problem occurs due:o power and g.ound distur-bances caused by the equipment. It is an all too common pracdce to draw controiier power f;om the same source as feeds the power eqtupmcnt. T."is power may provide

he recessary energy to drive the equip-ment. but it is not suitable to power the electroiucs lFigurc 3).

Hopefully, all indusuial equipment wiU have electronics powered from a separate low power 120 volt circuit. It solves several problems. First. it separates the electron-ics power from the probably very noisy indusu@

grade

power, prevendng the switching transients and startup sags from genug to the electronics.

Second, if it is necessary to condition the electronics po~er from an extet".A problem, it is tar cheaper to condiuon the watts necdcd (or electronics power than it is to condition thc kilowans required by the system.

If power cannot be separated.

then it is necessary to provide a bu!letproaf power supply.

preferably incbrding an isolation transformer, to separate the entire power supply from the electrical equipment.

Ground Noise. Ground noise, inevita-ble in industrial environments.

must be diverted from the electronics modu!c.

Multiple grounds in a system wiG often result in ground curt ents circu!sting through the cquiprnent. and ground noise circuhting through thc electronics path will cause rnalfuncuon.

Figure 4 shows some typical ground loop situations.

A common approach is to demand a super

~arth ground. This is good, but it is not a cure all. and often a super ground cannot be achieved, no matter how you uy. How do you get a super ground from the third 4oor? The real need is to gct'a stable ground reference to all interconnected equipments.

If this equipment is closely located, then a very low impedance interconnect is feasi-ble.

Power conditioners are o(ten tasked to eliminate RF of g.ound noise...at work, but:hese problems can be rc.

with an isolat:on rant(armer:o

=".-..-

neutral to ground noise and ~~rh "~IIpo; linc Glters.

So you may wan:;o try inexpensive approach nrst.

Data I.inks. Data links are su.~ng over the er'.tire!acirty. exposing ther.:

two principle e((ects. ground raise are pickup. Ground nois>>; willcause data er.-

unless the electronics has been designe" accornmodatc potential differences of ->>

eral volts or more. This is accampbs; with dif(crential drivers and receivers:f =.

must be direct coupled.

Optical 'hnks-eventually take over these links.

The other aspect is RF pickup. Inexpe sive shielded cable is suitable (ar:.'urpose.

Ground borh ends! Do not app single point ground techniques to RF. i!

Iaw frequency ground loop problem:s threat.

then onc end can be capaciuve Summary Industrial electronics are subjected rc harsh environment. Good design and inst.

!ation techniques willminimiae problems thc Geld.

Adhcrencc to the Europe; standards, IEC 801.x is a good start, ev>>

ifyou are only markeung in thc USA.

Bibliography FIPS PUB 94, Guideline on Elec'ac:

Power for ADP Installauons.

Scptcmber

1983, IEC 801-2. Electromagnetic compatibilit.

(or industrial. process measurement an<

control equipmcnt, Elcctrostauc dischargr requirements, 1984.

(EC 801-3. Electromagnetic compatibi!it!

(ar industrial. process measurement anI control equipment, Radiated electromag cetic Geld requirements, 1984.

IEC 8014, Electromagnetic compatrbilit:

(or industrial-process measuremcnt ar.:

control cquipmcnt. Electrical fast:ransreht burst requirements.

1984.

EMC Test 8c De.<n

e

industrial Fquipment Equipment Ground Bondlng-Designing for Performance and Life A Discussion of Ground Connection Fundamentals to Control EMl By D.B.L. Durham Dytecna Ltd, UK The problem of achieving satisfactory earth bonds or ground connecuons has plagued EMC engineers for many years, not only because the bonds are often vital for the achievement of satisfactory equipment pei-formance but because they affect the long term performance of equipment after it has been introduced into service.

Recommendations on bonding have ex-isted in the form of military speciBcations, such as Mil Std 1310.

Mil 188-124A and Mil-B.5081 (ASG) for some years and these have generally proved satisfactory for most new builds. However. these speci6cations have certain limitations in that they gener-ally do not spectfy consistently low levels of bond impedance.

nor a

suitable test method.

The introduction of ncw EMC speciGcations in Europe with the EEC Directive on EMC and the requirements for iong tenn stability in EMC characteristics has directed the UK nuTitary to review cidsutlg spcciEcaQons and Introduce a ncw Defence Standard to tighten up perfonn-ance requirements for nurttary equipment.

Dcf Stan ~ (Part 1)/1 has been intro-ducal to address thia area as far as mob9e and transpottablc canmunications installa-tions are conccnteia. but the requirements shoukl have implications in industrial apph-cations and over the wlxHe ekctronics market iflong term product performance is to be guaranteed.

Bond Degradation Earth or ground bonds are generaHy considered essenual not only for safety reasons.

but as a mean of diverting EM

currents, "locking" circuit boards and 38 eqtdpment to a stable ground point. achiev.

ing adequate levels ofcabk shiekhng and for many other reasons.

Many designers un-derstand the requirement forshort. fat bond leads to minimiac ground inductance, but few appreciate that a critical aspect is the connection resistance with which the bond strap is attached to the equipment ground point. Thc basic requirement of any bond is that it should have as low an impedance as possible (uille55 it is a dchbcratc induc-tive bond to limit ground currents).

Tbe impedance is a combination of the resistive and the inductive components.

Thc resis-tive element is a funcnon of thc bond strap resistivity, cross sectional area and length.

see Equation 1. whiht the inductive compo-nent is a more complex function ofthc bond strap characteristics as shown in Equation 2.

R~-

qf 0

A L

~

ln

+ 05+ 02235 j

iZ,j'f b+ c" 2ic L b~c 2f J (2) where R resistance, g ~ resistivity, f>

length. A ~ area, p, pcrmeabiEty of free

space, L ~

inductance.

p

~

feLttivc permcabiTity, b strap width, and c stttp thickness.

The frequency at which the inductive ekment dominates the impedance expres-sion when calcu!ating thc total inductance is, from Equation 3, typically 1 kHs. lt wgl be seen therefore that to al intents and

" purposes thc bond except at DC and power frequencies, may be assumed to be an inductance.

At very high ~equencies:r stray capacitance across the strap donunate.

This means that the volt "-:

across a bond is generaily a funcuon inductance and frequency. Based on Ohr.".

Law this volt drop is shown in Equador.

~

For transients the voltage drop is giver..

Equation 5.

Z Rt + cutLS V

IZ ~ jauU d1 dt (5

where Z ~ strap impedance.

cu ~

tacoma frequency, V ~ voltage, and 1 ~ current.

From this. the higher thc inductance th more isolated the circuit or box become from ground. Ths can have sigru6can effects oa equipment.

inchding enhance ment of noise injecbon onto arcuits, reduc=

non of Ster performance, and loss oi coaunlmication range. From a TEMPEST standpoint it may result in more radiation from emanent. lt would seem from this that the cntcria tor any bond is the inchlctancc and hence thc choice of short fat David Dugan served for 21 yean itt the Brrtiab Anny, where hc gained his Chgree ia dectricaf cnginceting. After service in a variety of appointments hc retired to join the Recalls'ES company as the Techmca/

Mtnsgcr rcspottsiblc /or the cfes/gn and devclopnsettt of commctttication systems.

fn 198$ he jainef Dytecna as the iManager of the &ginecnttg Division. and now ts currcntfy Tchtaica/ Marftcting Manage..

JulytAugust !991

5 P

sc';o rc(c gq(r

~ rri.

~~~cwc "-av

~l-l 'sis~fN f

$ Qsk gRrrfNT 'N,ff 'Srn ryort:4i csi r:i.rrenl

',:!0 iQ A. os JVA5Hf5 rrlltfR scarc 5 err VOUAGE &ifASI 'REhlENT Figure!. Bond resisurce.

Figure 2. Four wire bridge method.

bond straps.

However.

an anaiys's of:he bondinductancf shows Jiat I'or a bond strap of 100 mm!org, 15 mm ~~de and 2 mm Ihck the impedance at 1 MHa will be 3.8 Ohms.

It sounds extremely sin:pie.

but work performed in '.he USA'nd I:K shows that if an etrot is made 'in the way the strap is ternunated then a progressive increase in

he resisunce of the bond strap to box juncnon can occur as the equipment ages.

Eventuaily the resisunce will begin to exceed hundreds of ohms and may eventu-ally go open c cuit. This can negate thc effect of the bond strap completely as part of the E!rIIprotecdon.

tVhat happens with bonds to cause this change! Essentially a ground connecuon is a

series of irrpedances from thc strap through to the grourd material, as shown in Figure 1. Each point of contact contrib-utes to the total bond per'.onnancc;.

As a result. a change in any contact condition can result in a

change in the total bond resistance.

As is weil appreciated, the" contact resistance between two metal sur-faces is a hnction of the pressure.

The pressure exerted by the tip of a drawing pin is vastly greater than that from the thumb pressing by itself. Thus the contact from a sharp poult gives a tsoch higher prcssure t.'un a Oat point and 4gtiforc lower contact resisuncc. Measutetata have shown that sharp points enable Contact rcsisuncc of a few nicroohm to be achieved whilst similar pressures on Oat surfaces result in mil-liohms of conuct resistance.

It might be felt that there is little or no difference between these values, but in reality there is. An essennai aspect of a good bond is that it should remain so after the equipment has entered usc. High pressures also have the effect olsqueesing out corrosive materi-als and insulating 6lms. The former causes EMC Test & Design progressive degradation nt bonds, whilst the latter can reduce the ef'ciency of the bond f;om the moment it is installed. It is particuiariy important in communications

systems, where 6lters are insuiied and shielded cable ternunations are made Jiat the bords are of ',ow resistance and reuin their perfonnance.

Bond Performance and Mtaalaremcnt Experience has shown over a number of years that for long tenn consistent bond performance a low value of resistance must be achieved. This is typically 1-5 milhohms.

In Def Stan ~ (Part 1)/1 the vahie has been set at a maximum ol 2 miUiohms. This level is measured through the individual bonds.

Thc logic behind this level is twoloid. Firstly. experienc has shown that with communicaions equipment in particu-lar this value ol bond resisunce is required ifconsistent performance is to be achieved in terms of reception ef6ciency and trans-mission characteristic.

This is particularty so for TEMPEST protected equipments.

Tbe second point is that if the bond has a higher resistance then there is a signi8cant likelihood that progressive degradation will occur and the bond resistance wig increase in value. There will then be a progressive loss in perfonnance.

The main problem with measuring bond resistances is that it should bc measured using a

low voltage/curtent tcchtilquc, Moat techniques to date for assessing safety'nvolves driving a large current through the bond.

This checks the bond's abiRty to carry current but does not necessarily check its EMI protection perfonnance.

The rea-son is that many bonds may when in normal use have a high resistance due to oxide and greasy 6lms. but when subjected to a high current thc layers heat up and are vapo-rised. After the current is removed t.".e Sm can return. Thus high currert techmques are not recommended ror:esang EMI bonds. The new Defence Standard in the UK speci6es a maximum probe voitage oi 100 mictovolts. This reprcsenu typically a probe current ol 50 milliamps under shor citciit (<

1 mQ) condiuons.

This -:s insuf6cient to destroy surface 5lms. The chssic method for measuring low resisunce has been to use a four:erminal bndge as shown in Figure 2. In this case'the current is driven between two points and Ae VOltage aCtOSS the Sample iS meaSured witN a high resisunce probe. This removes Jie cfccts ot thc ptobe contact resistance and lead resistance.

This is generally consid-ered to be a laboratory method as the use offour contacts can be awkward. Ifthe lead resistance can be removed by a calibrauon teclniquc then thc four terminals may be replaced with a two terminal system.

h further possible re6ncment to the tecluique is to use a frequency that is not DC or 50/60/iOOHx. In this case 10.4 Ha haa been chosen. If an active 6lter is used to Ster out a9 other electrica noise, then it is posgblc to use the bond resistance meter on powered up systems. It is worth noting that at this frequency the impedance ia stol htgcly represented by resistance rather than inductance.

The two termmal method is shown in Figure 3.

The introduction of new EMC/EMI spcci6cations in Europe has made it more important that once made the bonds have coilslatent long 'tcrlil performance.

Ths means measurin on periodic inspecnon ard aRct maintenance. It is an essential aspect of insutmg consistent perfonnance.

It has been shown that within months apparen Jy good bonds can deteriorate to.-igh resis.

AAeqT

~

ic'oK ue45uI>eNr eiieo otsis ascg 0

FIXED RKSlSTANCE LEAOS Figure 3. Two termmal bridge method.

40 4

y 4

I c<~~ 4c c>~

~~c

~ ( olnpi4lllA Q~+ c~

~

(4 eeoc 4

((i+

yo

~befit%

CUSTOMQAG Ti 4.++++"

9t ~o'4+

a ~~>

t iQN CQMNC Rf AC.

0 odgot tc~

'e+

b~+.=nuators, Coaxial Term.

%/@ < y<c

<4~

connectors,

~4. ~b,~

Cata

~pc+ q4

.attest lE 11?0-1? bnCOln Avenue. FOlbrOOk.;4Y 11741 (S16) ~85.)4tX1 FAX i16.~BR >l14 INFO/CARO 29 tance. Ther.io.e t.'i vces ii '-. ~

suo)ect to testing and.xa.-..;.-.a:.cn-as a ma:ntenance tas~.

LK ~military Experience There have oeen

'.wo ~ lcr -.::

caused by poor bonus expenence degradauon

.n ericrmance ~ready-uoned in:tus a.-.icle. The:oss i;cir, cation range.

poor EMl pe.",'ormance other eifects aL'onrnbu'.e to a cons:dert reduction m equi".-..ent Nc:ency ann a.

ability. The secord ei;ect wnich is difficultto denufy:s:hat ai No Fault F.

(NFF) problems.

An analysis of reN'."

failures from military reiiaLiity data shown that NFF inc:dents can ae extre..

high, particulariy:n hunud;timates.

has been partially confirtned by reports.':

the Gulf Sar when aH forces repor-.e.

increase in availability of equipmen:.;.

'rier clunate.

Many fau!ts are due to "

electrica contacts in connectors. bu: a lar number have been idenufied as excess:

EMI induced through poor ground onnc This may be caused by either a loose groi strap or connector terminauon to:he ':c A significant improvement In equipme availability and perfonnance is expec:e when more recent statisucs are analysed.

The introduction:nto the Bnush Ar.-.

service of Ae Dyteaa Bond Resistanc Test Set DT l09 has enabled the L":

mihtary to measure bond resistances installed equipment and reduce;he curances of NFF errors. The UK mi:ar measurement procedure uses a two:er-...:

nal bridge method and an accurate miiliohm calibration star. dard.

This meas urement procedure and equipment is als in use by other NATO nations and e!se where by iniTitaryand naval forces wro hav recognized the same problem.

Cottclueiotta The problems with ground bonds have become significant with the development o:

sensitive and secure communications equi"-

ment. This coupled with an increasing reed to achieve higher and higher leveh of KMf protectiwt has lead to an increased emphasis beng placed on the effectiveness ofall types of systettt grounds.

These.

further com-bined with a requirement to ensure the long life of systems once in service.

have resulted in the assessment that bonds and terminations are one of the primary causes of EM faBures in systems.

The require-ment to test these is clear. however the means to do so have not always been available to engineers.

July.Aug.st '.9o'.

Panel SessIon PES Summer Meeting, July lQ, 1988 Long Beach, CaHfornia John G. Kappenman, Chairman Power System Susceptibility To Geomagnetic Disturbances:

Present And Future Concerns John C. Kappenman, Minnesota Power The effects of Solar. Geomagnetic Olaturbancea have been observed for decades. on power systems. However, the pro-found impact of the March 13, 1989 geomagnetic distur-bance has created a much greater level of concern about the phenomena in the power industry.

Several man.made systems have suffered d)eruptions to their

- normal operation d)aa to the occurrence of geomagnetic phe-nomena. Moat of the man~e systems, such aa commu-nications, have brett made less susceptible to the phenom-ena through technological evolution (microwave and fiber-optlc have replaced metall)c wire systems).

However, the bulk transmission system, If anything, is more susceptibl ~

today than ever before to geomagnetic disturbance events.

And lf the present trends continue, it la likely the bulk trans-mission network willbecome more susceptible In the future.

Some of the most concerning trends are: 1) Th>> transmission systems of today span greater distances of earth-surface-potential which result In the flow of larger geomagneticaily-IEEE Power Ealineeciag Review, October 1989 Induced. currents in the system,

2) the interconnected sys-tems tend to be more stressed by large region-to.region transfers. combined with GIC which willsimultaneously turn every transformer in the bulk system into a large reactive power consumer and harmonic current generator and 3) in general, large FHV transformers, static var compensstors and relay systems are more susceptible to adverse influence and microperation due to QIC.

TRhNSFORMER OPERhTION The primary concern with Geomagnetlcally-Induced Cur-rents la the effect that they have upon the operation of large power transformers. The three major effects produced by GIC in transformers Ia 1) the Increased var consumption ot the effected transformer, 2l the increased even and odd harmon-Ica generated by the half.cycle saturation, and 3) the possi-bilities of equipment damaging stray f)ux heating. As is weil documented, the presence of even a small amount of GIC I20 empa or less) wN cause a large power transformer to half-cycle saturate. The haif~c)e saturation distorted excit-ing current ls rich ln even and odd harmonica which become introduced to the power system. The distortion of the excit-ing current also determ)nes the real and reactive power re-quirernents of the transformer. The saturation of the core steel, under haif~c) ~ saturation, can cause stray flux to en-ter structural tank members or currant wlndlngs which has the potential to produce severe transformer heat)no.

15

t I

tr:rscs'rr ei

'- ry es.

-e 'io:est 'esul!S trct ~

ste '.na; Stogie CnaSe sr 5.'"r-..erS;ai'CVCle Saturate muCn mere easilv ano!0 4 -"cn greater cegree;han ccrricarable

.;nree onase units.."ese:.ansformers produce higher mag.

nnuces of harn onics and "or sume!arger amounts of rese-

!rve power when comPareO with three phase deslgnS.

ItELhy HAND PROTECTIVE SYSTEMS

>here are three ba5ic faitire modes of relay and protective Svs!ems:nat can oe attr:buteo to ggeomaghetic distur-bances:

r

~

u ~ >>

~ne ear!:i d,

<<3g. 4!'

4 C 4

dra geonag,"et.'c storms when!hey are -'

'- 4,

~ de~~" ~

SUNSPOT CYCLES hND GEOMAGNETIC DISTLRBhNCE CYCLES On the average, solar activity. as measured by t."e nur oer monthly sunspots.

follows an 11 year cvcl'e.

he "esen!

sunsoot cycle 22 had its minimum tn Seoten.oer 1986. a."c is exoected to oeak in 1990-1991. Geomagnetic 'ela o 5 ~

tvrbance cycles do not have the same shaoe as:ne sunscot number cycles. even thOugh they are cyclical. F Sure 1 snows the nature of the sunspot numbers and geomagnetic 3C!i"".v

~

. alse Operation of:."e protection system. such as hav-ing OCCurreO rOr SVC. CabaCitar and line relay Opera-tions where:t e!tow of harmonic currents are misin-

erpreter2 OV the reiaV aS a rault Or OVerlOad COnditicn.

This is the most common failure mode.

~

Failure to Operate when an operation is desirable, this has shown to be a problem for transformer differential protection schemes and for situations in which;he output of the current transformer is distorted.

~

Slower than Desired Operation. the presence of GlC can easily build up high levels of offset or remanent ttux in a current transformer. The high GIC induced off-set can significantly reduce the CT time.to.saturation for offset fault currents.

4uit'ocr or C(rtworts Cctrs rrrr apr25 SurtsIre4 i431

~ 1444 Cycle i7 Cycle la Cycle ia Cyote 20 Cyci~ 21 i

I r

~ )40

120 l

irumoer el i Olslureetd OdyVyear 150 t I

Suitspoi Humber

~ 00 I

IiiIi i

)

Ir I

I l00 j 50 j,

~

~ d0 I

i de i

40 I

~ 20

~

~

. ~ 0 1

45 40 uiI I

\\

!400 35 ee cS Most of the relay and protective system misoperations that are attributed to GIC are directly caused by some malfunc-tion oue to the harsh harmonic environment resulting from large power transformer half-cycle saturation. Current trans-former response errors are more difficultto directly associate with the GIC event. For exampfe in the case ot CT remen-encc. the CT response error may not occur until several days after the GlC event that produced the remanence.

Therefore.

these types of faitures are more difficultto substantiate.

50 55 40 45 70 75 40 Figure 1. Vaitstfons of the Yearty-Averaeed Sunspot Number enid Qetsmaenettealty Olsturbed Oays from

'1 932-1SBB.

cycles from 1932 to 1988 i2, 3l. Note that the geomagnetic dleturbanCe CyClee Can haVe a dOuble peak, One Of WhiCh Can lag the sunspot cycle peak. While geomagnetic activity in the present cycle is expected to maximize in approximately 1993-1994, severe geomagnetic storms can occur at any time during the cycle; the K-9 storm of March 13, 1989 was a striking example.

CONCLUSIONS As evident by the March 13th blackout in the Hydro Quebec system and transformer heating failures in the eastern US, the power industry is facing an immediate and serious chal-lenge. The power industry is more susceptible than ever to the influence of geomagnetic disturbances. And the industry will continue to become more susceptible to this phenome-non untess concened efforts are made to develop mitigation techniques.

EhRTHQURFhCE.POTENTIhL hND GEOMhGNETIChLLY.INDUCEDWURREVTS The auroral electrojete produce transient fluctuations in the eanh'5 magnettc field during magnetic storms. The earth is a conducting sphere and portions of ft experience this time-varying magnetic field, resulting in an induced earth-surface-potentlal lfSP) that can have vetuee of 1.2 to 8 volta/km t2 to 10 volte/mile) during severe geomagnetfc storms in re-gions of low earth conductivity l4),

Geomagnetic Disturbance Causes And Power System EEects flectric power systems become exposed to the 8SP through the grounded neutrals of wye-connected transformers at the opposite ends of long transmission lines, ae shown in Figure

2. The fSP acta ae an tdeal voltage source impressed be.

tween the grounded neutrate end has a frequency of one to a few mittfhene. The gsomagnetlcally-fnduced currents lGIC}

are then determined by dividing the ESP by the equivalent dc resistance of the paralleled transformer windings and line conductors. The GIC ls s ques%tract current, and values in excsee of 100 emperse have been rrNaeured in transformer

neutrals, Vernon D. Albettsw2 University of iiIJnaaota POWER SYSTEM EFFECTS OF GIC The psr.phses GlC in power transformer windings can be IEEE Power Engineering Review. October 1989 SOLhR ORIGINS OP GEOMhGNETIC STORMS The solar wind ie s rsrffied plasma of protons and electrons emitted from the sun. The solar wind fs affected by solar flares, coronal holes, and disappearing filaments, and the so-lar wind paniclee interact with the eenh'e magnetic field to produce auroral currents, or auroral etectrojste, that foltow generally circular paths around the geomagnetic pate! at al-titudee of 100 kilometers or more l1). The aurora borealis ie visual evidence of the auroral electrojets in the northern l6

1

~

44

~ Ina ia.sat~tat.ah n richI:.tars gr~ers ause clay misoperat on tet

~ 44 Al

=.sr~ ti.RFACS

~I~

9 T

SARvii-SuRFACS <T5hti41. ~

t'igui

~ 2. Induced 5aich. Surtact-Poitntial ISSP) Producing Qtamtg.

htticaliy Induced Cunehis IQICI in Power Sytttmt.

io.ai many times larger than:he RMS ac magnetizing current, re-sulting in a dc bias af transformer core flux, as in Figure 3.

I I

lo,pl I~

REF EHENCES

1. Akasatu.

S. I"The OyhtmiC Aurcra, 'ciehut C Ame":~

Mageziht. May 1989. pp. 90 97

2. Jactlyh. J. A.. "Reel-Time Ptedicuch ot Gfcbal Goon agr ti

Activny, Saltr Wind Magnetosphere Cauphi'g, "p.

141. Tarrt Sciehuflc Publishing Company.

oxvo. '986.

3 Thompsah.

R. J., "The Amplitude at Saiar C.cie 22, Riidia ehd Space Sewiciis TtchhiCal Report TR 87 03, ctc bti 1987.

4 V. O. Albertgan ehd J. A. Vth Batten, "Utcuic phd hlaghe.;i Fields at tht Stnh's Surface dut io Aurartl Cunenls." i55i TcahaaCIiohs ah Power Apparatus and Systems. Val. PAS 39 hia. 2. April 1970. pp. 578-584.

5.

J.

Q. Kappehmtn, V. O. Albtrtaon. 9. Mohan.

"Curser Trahatarmtr ahd Relay PtrtolmtnCt ih the Presence ot Gtc magnetically Induced Currents." ISEE Triiheactians ah Paw e~

Apparatus ahd Systems, Vai. PAS.100.

No. 3, pp. 1078-1088. March 1981.

The Hydro-Quebec System Blackout Of March 31, l989

<<ICI<<r~

Rgurt 3.

OC Btt! af Trtntfarmtr Cart Rulc Out to QIC.

The half.cycle saturation of transformers on e power system is the source of nearly all operating and equipment problems caused by GIC's during magnetic storms. The direct conse-quences af the half-cycle transformer saturation ere:

~

The transformer becomes a rich source of even and odd harmonics

~

A great increase in inducttve vers drawn by the trans-former

~

Possible drastic stray leakage fiux effects in the trans-former with resulting excessive localized heating.

There are a number of effects duo ta tho generation of high levels of harmonics by syatim power transformers.

includ-

ing, Overloading ot capacitor banda

~

Possible rntsaperottan of relays

~

Sustained overvoltogos on lang.line energizattan Higher socondory arc currents during single. pole switching

~

Higher cfratffCMaker recovery voltage

~

Ovorlaadtno of harmonic fttfyrs of HVOC converter ter-minals, and dtotantan in tfio ac voltage wave shape that may result in loss ot dc power transmission.

The increased tnductfvo vora drawn by system transformers during halfwycl~ soturat)an are sufficient to cause intoler-abto system voltage depression, unusual swings in MW and MVARflow on transmission linea. and problems with gener-ator var limits in some instances.

In addition to the halt-cyclo saturation ot power trans-formers, high levels of GIC can produce a dlstarted response IEEE Power Engiaeerintf Review, October f989 Daniel Saulier, Hydro-Quebec On March 13. 1989. an exceptionally intense magnetic storm caused seven Static Var Carnpensators ISYC) on the 735-kY network to trip or shut down. These compensetors are es-sential for voltage control and system stability. With their loss. voltage dropped and frequency increased.

This ted to systartt instability and the tripping of all the La Grande trans-mission lines thereby depriving the HQ system of 9500 MW of generation. Tho remaining power system callapsed within seconds of tho loss of the La Grande network. The system blackout affected

~it but a few substations isolated anto lo-cal gene~sting stations.

Pawer was gradually restored over a nine hours period. Oe-leya in restoring power wore encounterea because of dam-aged equipment on tho La Grande n>>twark and problems with caid load pickup.

SYSTEM CONDITION PRIOR TO THE EVENTS Tatal system goneratfon prior to the events was 21500 MW.

mast of it coming from remote power-generating stations at La Grande, Mantcouagon and Churchf8 Felts.

Exports to neighboring Systems totalled 1848 MW of which 1352 MW were on OC interconnections. The 735-kV transmission net-wark was laded at 90% of tts stability limft.

SEQUENCE OF EVENTS At 2:45 o,m. on March 13, a very intense magnetic storm ted to the conaequortttal trtp or shut down of seven SVC's, Contafnlng tho impact of tho event through oporatar inter-yention was impassible att SVC'a having tripped at caaaod to function within o ano minute period.

A fow seconds l8-8 s.) after tho loss af tho last SVC, ~ II five 735.kV lines of the Lo Grande transmission netwark tripped duo to an out of step condition. These Itne trips deprived the system ot 9500 MWot generation and subsequently ted to a campteto system cat tapao.

17

1

\\

i O ~

ecuon ance re~air ir 9 'cur SVC's snot sown ov caoac tor

~oitage uncaiance prctec::o."

>>aivsis ot vootage ano cur.

rent osc:itograms taxen at:ne Ct ioougamau site before tne SVC tnps snowed tne '.oitowing narmonic contents.

.wC

,hC Current ar l6 ky Harnaiiic Vcliage Order at.35 ky TCA Bracche TSC Brasche 100'"o 1OO 3'io a

~

I ifo 3~o

'Ã"

9 of I Oof 5 Oof I

~

3~o I00~o 36 ~o 24

~

}6 ~o 5%

16 ifo a ff Quasi DC currents generated by:he magnetic disturbance, saturating in tne SVC coupling transformers are thought to be the cause for such a targe second harmonic component of currant in the TSC branch.

DIsturoances On Poiyer Transformers Hooert J Hingjce James B. Stewart Power Techttoloip'es Inc.

This discussion addresses the effects of geomagnet:c cistii" oances on power transformers.

The primarv effect:s cue tc core saturation resulting!rom geomagneticaltv incucea c

r ~

rents. GICs. Core saturation can imoose severe temoerature problems in windings. ieads, tank plate ana structurai mer.-

bars of transformers and place heavy var and harmonic oi.r ~

dens on the power system and voltage support equiprriant.

GIC's of 10 to 100 amperes are more:hen mere nuisances in tha operation of power transformers, the rnanr.er of !Iow can result in saturation of the core and consequent changes in system var requirements.

increases in harmonic curren.

magnituctes.

increased transformer stray ana eady tosses.

and problems with system voltage control.

GENEAhL OBSEAVhTIONS ON THE SYSTEM BEHhVIOA The system blackout was caused by loss of all SVC on I.a Grande Network. Seven SVC tripped or stopped functioning.

Prior to and during the event all the OC interconnections be-haved properly. No relay false trips or misoperation of special protection systems were observed.

Telecommunications ware not affected. No equipment damage was directly attrib-utable to GIC but once the system split, some equipment waa damaged due to load rejection overvoltagea.

RF'VIEDIhL hCTIONS ThKEVi Since the event. the following actions were implemented:

~

SVC protection circuits have been readjusted on four SVC's so as to render their operation reliable during magnetic storms similar work is being performed on the four remaining SVC's,

~

Energy, Mines and Resource Canada now provides Hy-dro.Quebec with updated forecasts on the probability of magnetic disturbances.

Thaao forecasts are used by the System Control Center dispatcher to position the transmission system within secure limits.

~

A.C. voltage asymmetry ia monitored at four koy lo-cations on the system (Bouchorvitto, Amaud, LG2, Chhtgeaguay).

Upon detection of o 3% voltage aaym-rnetry at any ona location, the ayotom control center dispatcher ia alarmed and willimmediately ta'ko action to position system tranafor levels within secure limits if this haan't already been dane because of forecasted magnetic activity.

OPERhTING LIMNIDURING MhGNETIC DISTURShNCES (hND hLERT SITUhTIONS)

The fallowing operating limits are now being appliedt

~

10% safety margin shall be applied on maximum trans-fer limits.

Maximum transfer limits shell not take into account tho availability of static componaators deemed unreliable.

~

Adjust the loading on HVOC circuits to be within tho 40% to 90%, or loaa. of tho normal full load rating.

l8 CIC EFFECTS VERSUS CORE hND WIIOoDINC COiVFIGURhTIONS Principal concerns in this discussion are for EHV systems with grounded Y transformer banks providing conducting paths for GIC and zero sequence currents. Cora and winding configurations respond differently to zero sequence open.cir-cuit currents end to GICa. Note: aa used here. the term "open circuit"refers to tests performed with all delta connections opened or "broken." For example, the three. phase three leg core form transformers aro less prone to GIC induced satu-ration than three-phase shall form transformers.

But. both core form and shell form single phase transformers are sus.

ceptibl ~ to GIC induced saturation.

Winding and lead arrangemanta respond differently to GIC induced core saturation aa well. For example, the current dis-tribution within pareil ~ I winding paths and within low voltage loads depends upon the leakage flux paths and mutual cou-pling. Loaaea within windinga and leads may change signifi-cantly under GIC induced saturation owing to the change in magnetic field intensity. H, and the resultant changes in the boundary conditions for the leakage field path.

EDDY LOSSES IN STEEL MEMBERS The changes in the magnetic intensity. H, and the magnetic boundary conditiona resulting from the GIC excttation bias can increase tho loaaoa in steel plate, the losses for fields parallel to the plane of the plato increase nearly aa the square of H. Note also that the level of losses increase approxi ~

mately aa the square root of the frequency ot H. owing to the effect of depth of penetration. Tho magnetic field along yoke clamps and leg plates in core form transformers and in Tee beams and tank plate In ahoN form transformers closely matches tho magnetic gradient ln tho core. Areas of the tank and core clamps are subjected to tho winding leakage field.

If the coro saturates, the magnetic field impressed upon the steel members may rise ton to one hundred times normal duo to the saturation and the offocto of the leakage field. The loaaoo in the stool momboro willriao hundreds of times nor-mal, even under half-cycle saturation. On tho steel surfaces.

eddy lose density moy rise ton to thirtywatts par square inch, approaching the thermal flux density ot an ~lactric range ele.

ment.

Surface temperatures riao rapidly with this thermal flux and can result in degradation of insulation touching tne steel IEEE Power Eatpaeeriag Review. October l989

r

Destgn Ocf tcfency Ocl lrlent vendor aafma1 UPS haS no battery test CI I'CUIt Vendor naINIa1 na intenanc e sect loA docs AOI QCAt I OA batteries.

Design Deficiency Battefies have not been replaced in 6 ycafs Design Deficiency AC input to logic poucr salty is naihtchahcc preferred Back up batteries degraded or dead k.S relay charactertst-ICS PfCveAtS transfer to inverter output.

Breaker ffcIction per design Ground fault occurs on B phase of aain transfofncr Vol tagc traflicnt oA stat lofl AC pouer steeply AC pouer to logic nodule for UPStA.D,G cspcrIchccs the transient out pu'I voltage goes tou logic trips OA poucr SISIPty failure.

2VSS UPStA,S, C,O,G trip Breakers CS-I,2 3

open; Ch-C does not close ups loads do hot auto transfer to mint.

supply loss of all loads on UPS1A D,ri Faul t 1$ clcafcd in 6 cycles; transfer cocptetcd in 12 cycles Pernisstves prohibIt CS C

breaker froze clo>Ing CS'4 ncsvh<

to Ifunsfer na Int sullpl y to lnvcl tcf out put

r

1. Static DC testing was performed on individual chips from the effected circuits in order to characterize Latch-up susceptibility.

Static testing is performed with fixed voltage settings.

The testing was performed on the

4049, 4011, 4044 and 4068 devices.

The test scheme included:

A. Output voltage below Vss (Vss is the ground or negative power supply).

B. Output voltage above Vdd (Vdd in the positive voltage power supply).

C. Input voltage above Vdd.

D. Power supply overvoltage.

Vss>>Vdd The testing revealed that permanent physical damage was induced by tests C

and D.

Test B did not induce latch-up but test A

consistently induced latch-up on the 4049 and at higher voltage differentials on the other devices-The 4049 was quite sensitive to this test.

2. Voltage Dropout testing, where the Vdd was cut out and restored during time periods ranging from 20 mS to 2 seconds did not result in latch-up behavior on the individual components.

A "breadboard" test circuit was fabricated to simulate a typical circuit path, i.e.

a 4044 latch driving a 4049 invertor driving a 4068 gate driving a 4011 gate driving an MC1615 lamp driver.

Slow rate dropout testing of the test circuit did not result in latch-up or other anomalous behavior.

3. Board level testing has been performed on the A13A21 cards f'rom UPS units A, B and G.

A stock card has also been tested.

A test fixture was fabricated to hold the cards and supply paver and input selection and output monitoring.

An MC 1615 lamp driver chip with LED's (light emitting diodes) was connected to the card outputs.

Static boaef testing revealed no anomalies on the A card, a damaged U10 4049 oN the B card, No anomalies on the C card and a failure to set, any of the 4044 latches on the G card.

(Note that latching of the 4044 chips is normal while latch-up is an abnormal condition.)

The. functional failure of the G board has been traced to the Kl relay and/or the SW 1 reset svitch on the circuit board.

Failure simulation testing on the functional A board has revealed that the lighting pattern reported during the incident, whero the lamps on the card were extinguished and "downstream" lamps remained e

illuminated, has been duplicated by setting the PSF-not

latch, lowering the DC voltage to approximately 4

VDC and lifting (floating) the DC ground for several seconds.

If the DC voltage remains low after the ground fault the lamps on the board will reset and the external lamps (UPS fail, Logic fail and the SSTR lines) will remain illuminated.

The simulation conditions are unlikely to be those of the failure event, however, it has been demonstrated that the board can operate in an this illogical state.

A more complicated failure simulation (approximating the actual event) may produce the same results.

4.

High Speed transient. testing on the power lines is planned and delayed pending laboratory analysis of the degraded samples discussed in the following section.

M C

0 A.

NEGATIVE VOLTAGE ON THE OUTPUT OF THE 4049 MILL INVARIABLY CAUSE LATCH-UP-INJECTING NEGATIVE VOLTAGE INTO THE OUTPUT IS IDENTICAL TO RAISING THE GROUND Vss ABOVE THE OUTPUT.

B.

THE INITIALFAILURE CONDITION IN TERMS OF THE LAMP SETTINGS CAN BE DUPLICATED UNDER UNLIKELY CONDITIONS AND MAY BE POSSIBLE UNDER MORE PLAUSIBLE CIRCUIT CONDITIONS.

The following samples have been submitted for laboratory analysis:

1.

One battery pack.

Battery pack 1 from UPS 1C.

n'.

Two (2) integrated circuits from a failed A20 card.

A 4049 and a 4011.

3.

A failed U10 4049 from the A13121, UPS B card.

4.

The U10 4049 from the A13A21 cards from UPS's A and G

5. The UPS 6 Kl relay and SWl switch.

The following results have been obtained.

1.

Two (2) of the Three (3) batteries from UPS 1C were dried out.

Water was added and charging did not result in recovery of the cell. It is concluded that the cell failed due old age wearout.

Analysis is continuing on the other two (2) cells.

2.

The 4011 from the A20 card is electrically good.

Internal

I J

~

~

I

~ ~

~

lj

~

. inspection of the Die revealed no anomalies.

The 4049 is

'electrically bad.

A catastrophic failure.

Internal inspection of the die revealed severe damage centered on the Vss and Vdd power lines and several input/outputs.

The initiating overstress was introduced on the Vdd or Vss lines as indicated by arc-over damage acoss the oxide between the two power lines.

3. Electrical testing of the U10 4049 from A13A21, UPS B revealed that it was not functional.

Internal inspection of the die revealed a

fused aluminum metallization line to one part of the internal circuitry.

Probing revealed no )unction damage on either side of the fuse site.

This damage is characteristic of classic SCR latch-up.

4.

Electrical testing of the two (2) additional 4049's was performed during circuit board test.

The devices were functional.

Internal die examination revealed no damage on either device.

5.

Electrical testing of the A13A21 circuit board from UPS G

revealed that SW1 was intermittently open circuited in the normally closed position.

This condition would provide continuous reset signals to the 4044 latches through the K1 relay.

The switch was isolated and a

good switch was placed across the Kl relay.

The circuit board inputs still would no latch the lamps.

The Kl relay was removed and testing revealed that it was electrically good. The findings reveal that either the Kl relay is either intermittently bad or there are other problem components on the circuit board.

Analysis is continuing.

R 0

U 0

A.

THE DAMAGE NOTED ON THE 4049 FAILED INTEGRATED CIRCUIT FROM UPS B

WAS INDUCED ON THE Vss (GROUND)

SIDE AND SUGGESTS A

GROUND TRANSIENT MAY HAVE OCCURRED.

THE DAMAGE ON THE 4049 FROM THE A20

. BOARD INDICATES THAT THE DAMAGE WAS INITIATED BY A TRANSIENT ON EITHER THE Vdd OR Vss LINE, BUT IT MAY ALSO HAVE BEEN INITIATED BY LATCH"UP.

B.

AT LEAST ONE OF THE BATTERIES FAILED DUE TO OLD AGE WEAROUT.

C.

THE CAQNE OF THE FAILURE OF THE UPS G

TO SET THE LATCHES ZS UNKNOWN AT THIS TIME.

THE SW SWITCH HAS BEEN FOUND TO BE DEFECTIVE AND ANALYST'S WILL DETERMINE IF THE COMPONENT IS RELIABILITYRISK.

I

-UPS lA 1B 1G TEST

SUMMARY

page 1

Purpose:

To prove that the DC logic power for the Exide UPS is powered from the B-phase maintenance supply.

The K-5 pickup and drop out voltages and the DC trip-point o=

the DC logic will be recorded for UPSlA, not for UPS13 and UPS1G.

The internal batteries will be tested and replaced.

Results Summary:

1.)

On UPS1A, UPS1B and UPS1G, it was verified that the DC logic power supplies are fed from the B-phase maintenance supply.

2.)

The K-5 relay drop out and pick up voltages were recorded for UPS1A and they were found to be below the trip point of the DC logic power.

3.)

On UPS1A,

UPS1B, UPS1G, the maintenance supply was opened with the UPS feeding the loads and no UPS trips occuzredo 4.)

On UPSlA, UPS1B and UPS1G, the batteries were replaced.

CONC U

This test proves that the DC logic power is fed by the B yhaae maintenance power.

It proves that the internal batCOries were effectively dead.

For UPS1A it proves that on a slow transient that. the DC logic power will drop out before the K-5 relay will transfer to UPS power.

Numerical Results:

page 2

1.)

The UPS1A DC logic trips at <16.7 VDC. (with.75.6 VAC on input).

2. )

UPS1A:

K-5 relay drop out 47 VDC K-5 relay pick up

- 52 VDC 4.)

The internal battery voltage was measured:

UPS1A:

UPS1B:

UPS 16:

Positive-Negative-Positive-Negative-Positive-Negative-0.54 6.2 18.3 0.69

)

J

2VBB-UPS1C TEST

SUMMARY

page 1

Purpose:

To prove that the DC logic power for the Exide UPS is powered from the B-phase maintenance supply and that if a transient occurs on the maintenance supply it can effect the DC logic such that it will trip the unit.

This test is done with the old internal logic batteries and then repeated with new ones.

Each of the inverter trips will be tested to verify that each circuit is still intact except DCOV., An AC input transient to UPS will be simulated to verify that the unit can "ride out" a normal AC input transient without tripping.

The K-5 relay pick up and drop out voltages and the DC trip-point of the DC logic will be recorded.

Results Summary:

1.)

Zt was verified that the DC logic power supplies are fed from the B-phase maintenance supply.

2.)

A rapid open and closing of the upstream normal AC input breaker to the UPS was done and the unit did not trip or go on battery.

No noticeable effect was seen on the UPS output.

3.)

Each inverter trip circuit except DCOV was tested and each functioned as designed.

4.)

Fast transient tests:

With the old batteries still installed a voltage interruption of 100 - 150 msec duration was given to the AC input to the DC logic of UPS1C.

The DC logic was initially at 19.86 VDC.

The unit tripped 3 out of 4 times.

This was done first with the loads on

. xaintenance supply and then also with the loads on UPS power ~

With the new batteries installed there was no trip when the fast transient test was performed 25 successive times.

There were no trips but a repeated SCR short alarm occurred which is indicative of noise spikes within the unit.

page 2

5.)

The K-5 relay drop out was recorded and was found to be below the trip point of the DC logic power.

6.)

Normal transfers were done, UPS to maintenance and maintenance to UPS, with dead batteries and there were no trips of the UPS.

The maintenance supply was opened with the UPS feeding the loads and no UPS trips occurred.

CONCLUSZON This test proves that the DC logic power is fed by the B phase maintenance power and that it is susceptible to voltage transients on the maintenance supply.

Zt may be susceptible to other transients as well because it is directly tied to maintenance supply.

The test DOES NOT prove the level of susceptibility, that is, it does not prove that the transient was of any set voltage or duration.

The test implies that the batteries may have mitigated the trip but is not conclusive.

Each trip circuit was tested successfully so no failure to any of these occurred that caused the trip.

The fast open/close of the normal AC input breaker proves that the unit would withstand an AC input transient without failure or without going on battery power.

I

'5

~

Numerical Results:

1.)

Fast Transient Tests a.)

W't existin batteries page 3

1.)

With loads on maintenance:

At 19.86 VDC (90.0 VAC) - ~tri (150 msec.)

At 19.86 VDC (120 VAC) - ~t '150 msec.)

2.)

With loads on UPS power:

2 Tries, 1

~t i (200 msec.)

b.)

W't new batter'es-1.)

Approx. 20.0 VDC - 25 times,

~no tri (100 msec.

)

2.)

The DC logic trips at

< 16.9 VDC. (with 84.59 VAC on input).

3.)

K-5 relay drop out - 45 VAC K-5 relay pick up

• • not recorded 4.)

The following trips tests were done:

a.)

b.)

c.)

d.)

e.)f.)

g

)

h. ).

OV/UV ACUV ACOV DCUV Frequency fail Logic Failure Power supply failure Clock failure 5.)

The internal battery voltage was measured:

Positive -

+0.6 Negative -

+0.04

l

~

4 5

5.)

Xndividual cell voltages:

page 4

1.)

2.)

3.)

4. )

5.)

6.)

Batte Volta e

1. 19 2.48 2.24 0.17 0.79 1.78 New Batte Volta e 6.10 6.07 6.10 6.09 6.10 6.12

V

-UP D

S S

page

Purpose:

To prove that the DC logic power for the Exide UPS is powered from the B-phase maintenance supply and that if a transient occurs on the maintenance supply it can effect the DC logic such that it will trip the unit.

This test is done with the old internal logic batteries and then repeated with new ones.

The K-5 pickup and drop out voltages and the DC trip-point of the DC logic will be recorded.

Results Summary:

1.) It was verified that the'C logic power supplies are fed from the B-phase maintenance supply.

2.)

Fast transient tests; With the old batteries still installed a voltage interruption of 100 - 150 msec duration was given to the AC input to the DC logic of UPS1D.

The DC logic was at 20.9 VDC.

The unit would not trip.

The AC input voltage to the DC logic was then reduced such that the DC logic was at 20.0 volts.

When the test was performed with the DC logic power at 20.0 VDC the unit tripped.

This was done first with the loads on maintenance supply and then also with the loads on UPS power.

With the new batteries installed there was no trip when the fast transient test was performed though there was significant hits shown on the DC logic power bus as seen by the oscilloscope.

3.)

The K-5 relay drop out and pick up voltages were recorded and they were found to be below the trip point

. of the DC logic power.

4.)

Normal transfers were done, UPS to maintenance and maintenance to UPS, with dead batteries and there were no trips of the UPS.

The maintenance supply was opened with the UPS feeding the loads and no UPS trips occurred.

I a

pl lt 4i V

1i

, ~

CONC page 2

This test proves that the DC logic power is fed by the B phase maintenance power and that it is susceptible to voltage transients on the maintenance supply.

It may be susceptible to other transients as well because it is directly tied to maintenance supply.

The test DOES NOT prove the level of susceptibility, that is, it does

not, prove that the transient was of any set voltage or duration.

The test implies that the batteries may have mitigated the trip but that is not conclusive.

ta 0

Numerical Results:

1.)

Fast Transient Tests a.)

W't e istin batteries page 3

With loads on maintenance:

At 20.9 VDC five tries, no trips.

At 20.7 VDC - one try, one ~tri (150 msec.)

2.)

With loads on UPS power:

At 20.06 VDC - one

~t (100 msec.)

b.)

W't new batteries 1.)

At 20.05 VDC - Five tries,

~so tri s.

noticeable DC hit on each transient.

2.)

The DC logic trips at <17.3 VDC. (with 84.5 VAC on input).

3.)

K-5 relay drop out - 42 VDC K-5 relay pick up

- 55 VDC 4.)

The internal battery voltage was measured:

Positive-Negative-

+0.6

+0.14 (the negative battery set was actually slightly positive).

5.)

individual cell voltages:

t V tae w Batte Vo ta e

2.)

3.)

4 ~ )

5.)

.254

.570 1.03

.07 1 ~ 17 6'0 6'6 6 ~ 10 6'0 6'3

t 0

I t

6.)

1.39 6'9

1

resulted until such time the power to these UPS's was restored.

However, one or more of the other lighting systems

namely, emergency, normal and 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> battery pack were available in these a'reas during this event.

The stairwells are provided with essential lighting only except where 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> battery pack lighting is added for Appendix R compliance.

Illumination to these stairwells was not available due to loss of normal UPS.

The 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> battery pack lighting did not energize because there was no loss of normal power.

Event Evaluation The root cause of loss of power from normal UPS is evaluated separately in another report.

During the

event, some areas of the plant lost partial illuminations provided by essential lighting for sometime.

Areas critical for safe shutdown where this lighting are identified are provided in Attachment 7.

During this event the plant was safely shutdown from the control room.

Because the control room is provided with adequate lighting without the essential lighting, loss of essential lighting did not adversely affect the operator actions needed to bring the plant to safe shutdown.

The access route used by the operators during this event for restoration of the normal UPS power (Attachment

8) supplies was illuminated from normal lighting except for the stairwell where portable handheld lights were used.

(FSAR Sec.

9.5.3.3 allows the use of portable lighting in the form of handheld flashlights for short excursions into the plant).

The normal UPS locations were illuminated by normal lighting.

Therefore, restoration of UPS power was unaffected by loss of essential lighting.

Even though the essential and egress lighting are powered by three normal UPS's, during, this event due to multiple

'ailures of all normal UPS's, essential and egress lighting systems were not available.

The existing essential and egress lighting design is adequate, however, the root cause of the multiple failures of the normal UPS's will be determined/evaluated and appropriate corrective action taken if required to ensure that multiple normal UPS's failure will not reoccur.

The proposed plant modification 89-042 will enhance the reliability of stairwell lighting where 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> battery pack lighting is provided.

NSK2 14

Status of Normal Reactor Buildin Li htin During the event on August 13, 1991, it was reported by the operators in the reactor building that some areas of the reactor building lost lighting momentarily.

Event Evaluation Normal lighting for reactor building general

areas, work area's and electrical equipment areas is provided with low wattage high pressure sodium vapor lights.

In the event of power interruptions or voltage dip lasting for more than one cycle, these fixtures extinguish and do not restart until the lamp cools and pressure decreases.

When a power supply to continuously energized sodium vapor light is interrupted, it has a

cooldown period before a restrike of the lighting can occur.

The cooldown period depends upon the rating of the light bulbs.

During the event, the emergency distribution system experienced a,transient due to the fault on the Phase B main transformer.

During the

event, the Reactors Building normal lighting in certain areas where these high pressure sodium vapor fixtures are provided, was interrupted for approximately 30 seconds.

This momentary loss of lighting was due to the inherent design of low wattage high pressure sodium vapor lighting which requires cooldown period prior to restrike whenever power is interrupted.

The same scenario could occur in the plant wherever power supply to high pressure sodium vapor fixture is interrupted momentarily, however, there is no indication of such a

loss in other areas of the plant.

Therefore, it is consistent with NMP2 lighting design and USAR Section 9.5.3.

h)

Grou 9 Isolation Valves Closure The group 9 primary containment isolation valves are part of containment purge system.

These valves are listed in Technical Specification Table 3.6.3-1, Page 3/4 6-24 and USAR Table 6.2.56, Page ll of 24.

The function of group 9 isolation valves is to limit the potential release of radioactive materials from primary containment.

These isolation valves are opened during power operation only at infrequent intervals to allow injection of nitrogen into primary containment to inert or de-inert the primary containment at a desired pressure.

These valves, if open, receive signal to close if any of the following happen:

NSK2 15

a)

High radiation through standby gas treatment system (SGTS).

The SGTS radiation monitor located in the main stack is designed to continuously monitor offsite release and provide isolation signals to these isolation valves.

b)

High drywell pressure.

c)

Reactor low water level.

d)

Manual isolation of main steam isolation valves.

Event Evaluation During this event, the group 9 primary containment isolation valves closed.

This isolation is the safe mode of operation limiting potential releases of radioactive material from primary containment.

The initiating condition for these valves occurred as a

result of loss of power to radiation monitor 2GTS-RE105 when UPS power to the DRMS computer was interrupted.

The actual isolation occurred when the logic was reenergized upon restoration of the UPS power supply to the monitor's auxiliary relay circuit.

Therefore, group 9 isolation valves closed as designed and is consistent with USAR Section 6.2.5.2.4, Page 6.2-77.

j)

Reactor Manual Control S stem The reactor manual control system (RMCS) provides the operator with means to make changes in nuclear reactivity via the manipulation of control rods so that reactor power level and core power distribution can be controlled.

This system is a power generation system and is not classified as safety related.

The RMCS receives electrical power from the 120 V AC normal UPS.

The RMCS does not include any of the circuitry or devices used to automatically or manually scram the reactor.

The RMCS control and position indication circuitry is not required for any plant safety function nor is it required to operate during any associated DBA or transient occurrence.

The reactor manual control circuitry is required to operate only in the normal plant environment during normal power generation operations.

The discussion of RMCS is consistent with USAR Sections 7.7.1.1, Pages 7.7-1, 2,

14.

NSK2 16

Event Anal sis The RMCS was lost during this event because its power

source, the normal nonsafety related

UPS, was lost.

The loss of RMCS is not a concern during this event.

Since the plant was automatically scrammed during this

event, the RMCS need not perform any function after the scram.

This RMCS is used by operator only during normal plant operations.

Therefore, if the plant had not automatically scrammed during the event, loss of RMCS would not have caused a safety concern based upon the following:

2)

EOP's provide guidance to the operator under situations involving failure to scram, and various ATWS mitigating design aspects of the plant were fully operable throughout the event.

Although this system was lost'during this event, its importance diminished once the automatic scram occurred.

Therefore it is concluded that the RMCS function was consistent with USAR Section 7.7.1.1.

k)

Feedwater Control S stem The feedwater control system controls the flow of feedwater into the reactor vessel to maintain the vessel water level within predetermined limits during all normal plant operating modes.

During normal plant operation, the feedwater control system automatically regulates feedwater flow into the reactor vessel.

The system can be manually operated.

The feedwater flow control instrumentation measures the water level in the reactor vessel, the feedwater flow rate into the reactor vessel and the steam flow rate from the reactor vessel.

During automatic operation, these three measurements are used for controlling feedwater flow.

The feedwater control system receives its normal power supply from the normal UPS.

The feedwater control system is designed to lock in its last position upon a loss of power to its control electronics.

The feedwater control system is discussed in USAR Section 7.7.-1.3, Page 7.7-23.

Event Anal sis During this event, upon loss of the normal UPS's, the feedwater control system performed as designed and failed in its last position.

Therefore, it is concluded that the feedwater control system function was consistent with USAR Section 7.7.1.3.

NSK2 17

h

'1

l)

Feedwater Pum Tri Feedwater is provided to the Reactor Pressure Vessel (RPV) via the Condensate Pumps',

Condensate Booster Pumps and the Reactor Feed Pumps shown in Attachment 9.

The Condensate Pump draws condensate water from the Condenser and provides the necessary Net Positive Suction Head (NPSH) for the Booster Pumps.

The Condensate Booster Pumps provide the necessary NPSH for the Reactor Feed Pumps.

A minimum flow control header is provided off the discharge header of each pump to ensure that the minimum flow is maintained through the associated pump.

The minimum flow control valves and associated instrumentation actuates to maintain this minimum flow.

The main feedwater control valves (LV10), located on the discharge header of the Reactor Feed

Pumps, modulate to control reactor water level.

The feedwater control system is powered by normal UPS power supplies.

The above discussion is consistent with USAR Section 10.4.7.

Event Evaluation It was reported during this event that feedwater pumps tripped.

An evaluation of this condition reveals that reported happenings are consistent with the system as designed and is in consistence with USAR Section 10.4.7.

The instrumentation controlling the minimum flow recirculation valves on the condensate, condensate booster and the feedwater pumps is powered from the normal UPS's.

These instruments are also designed to open the valve upon loss of power in order to protect the pumps.

Upon loss of normal UPS, the feedwater control valves fail locked in their last position.

Following the turbine trip, an ATWS signal would attempt to drive the feedwater control valves closed, however since an ATWS signal was not present, this did not occur.

With the feedwater control valves failed locked and the minimum flow control valves (FV2) driven full open, feedwater flow increases and approaches pump run-out.

The Reactor Feed Pump NPSH decreases to the low-low pressure trip point, tripping the Feedwater Pumps.

The Feedwater pump control circuit does not utilize an auto transfer logic to standby Feedwater Pump; therefore, feedwater flow is lost.

The instrumentation circuits for all other minimum flow control valves are also powered by normal UPS power supplies.

These valves all fail in the open position with the loss of UPS and contribute to the loss of Feedwater pump and condensate booster pump.

This is consistent with USAR Section 10.4.7.

NSK2 18

'I I

m)

Annunciators and Com uters The plant annunciator system provides information to the plant operators by windows located on the main operator panelboards and on back panels within the Power Generation Control Complex (PGCC).

This system does not include annunciators on local panels throughout the plant and on special panels, e.g., fire protection, within the PGCC.

The plant annunciator system is non-safety related and is connected to the normal power distribution system through normal UPS's.

The plant annunciator system is not discussed in the USAR.

Several computer displays, with inter-active keyboards, are located in the PGCC.

These displays are from the following computer systems.

PMS

Plant Process Computer LWS

Liquid Radwaste

Computer, which has the following subsystems:

LWS Liquid Radwaste Control GENTEMP Generator Temperature Monitoring ERF Emergency Parameter Display System SPDS Safety Parameter Display System DRMS Digital Radiation Monitoring System 3D Monicore A system used primarily for core calculations and monitoring In addition, noble gas information is provided to the plant operators from the GEMS (Gaseous Effluent Monitoring System) computer by chart recorders on a back panel; and the operators have access to the GETARS (General Electric Transient Analysis

& Recording System) computer.

All of the above computer systems are non-safety related and are connected to the normal power distribution systems through normal UPS's.

There are some safety related radiation monitoring skids that provide input to the DRMS computer.

However, these skids also provide safety related indication in the PGCC that is independent of the DRMS computer.

NSK2 19

The Plant Process Computer is discussed in Section 7.7.1.6 of the USAR, where it is mentioned that the computer is non-safety related.

USAR Section 11.2.1.2 covers the Liquid Radwaste System design basis and states that the power supply for all Radwaste System components is provided from non-Class 1E power sources.

This is compatible with the Safety-Parameter Display requirements since NUREG-0737, Supplement 1 states that the SPDS need not be qualified to Class 1E requirements.

The process and effluent radiological monitoring and sampling systems, which include the DRMS and GEMS computers, are discussed in USAR Section 11.5.

This section defines which monitors are safety related and which are non-safety related.

The DRMS and GEMS design complies with this USAR section.

Area radiation and airborne radioactivity monitoring instrumentation, which include the DRMS and GEMS computers, are discussed in USAR Section 12.3.4.1.

This section defines which monitors are safety related and which are non-safety related.

The DRMS and GEMS design complies with this USAR section.

A description of the 3D Monicore computer system was added to USAR Section 7.7.1.6 by LDCN U-1235.

This LDCN states that the 3D Monicore system is non-safety related.

The designation of the computer systems mentioned above as non-safety related is consistent with the explanation of the Uninterruptible Power Supply System in USAR Section 8.3.1.1.2.

In this section it is stated that 2VBB-UPS1A feeds the radwaste computer

hardware, 2VBB-UPS1B feeds local non-safety related radiation monitoring microprocessors, and 2VBB-UPS1G feeds plant computer loads.

Event Evaluation With the loss of the normal UPS's, the plant annunciation system and the computer systems listed above became inoperative due to the loss of power.

This is consistent with the plant design and the description of the plant in the USAR.

These systems are non-safety related

and, hence, are not required to shut down the plant following a design basis event.

NSK2 20

CONCLUSION Based on the above evaluation, it can be concluded that the plant responses during the event on 8/13/91 is consistent with USAR descriptions..

RECOMMENDATIONS Based on the above evaluation, the following long term recommendations are provided.

1)

Plant Oscillograph The in-plant oscillograph should be replaced with a more reliable and functional unit.

If this oscillograph was functional during the event on 8/13/91, adequate data could have been available to accurately evaluate the cause of the disturbance.

2)

Essential Lighting The proposed modification 89-042 should be implemented as soon as possible to enhance the rel'iability of stairwell lighting where 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> battery pack lighting is provided.

3)

Control Power Supplies During the Electrical Distribution System evaluation, it was revealed that most of the systems important to plant operations such as feedwater

system, annunciation

system, etc, receive their control power from either normal UPS 1A, 1B or both. It is recommended that control power supplies for these systems be evaluated and reconfigured to avoid plant transient due to loss of single normal UPS.

4)

Main Generator It is recommended that a thorough visual inspection be performed of the generator stator and winding support system during the next refueling outage (see Attachment 10).

NSK2

345KV TO Sl "A STATION LINE 23)

ATTACHMENT'-1 115KV SOURCE 'A'LINE 5) 345/25KV UNIT TRANSF.

498HVA EACH SPARE 115KV SOURCE '8'LINE 6) 115KV SOURCE 'B'OR

'A'2/56/78HVA RESERVE BANK 'A' WJGKV UNIT 25KV~

~NORMAL STA. TRANSF.

24.')KV/13.8KV 199-59/59HVA 42/56/79MVA RESERVE BANK'8'UXBOILER NO CUB ONLY 2NPS-SWG991 13.BKV NORHAL AUX.TRANSF.

CUB. ONLY 2NPS-SWG993 AUX.TRANSF.

NC

. ~ 13.BKV AUX.BOILER BUS 2NPS-SWG892 NO NORHAL 4J69KV 2NNS-6WG914 2NNS-SWG91 5 2NNS-SWGBII 2NNS-SWG812 2NNS-SWG913 4J6KV 0.16KV NO STUB BUS STUB BUS WJGKV~

CUIL ONLY NC 2ENSiSWGIB) iJ69KV OIV.I >

EHERGENCY BUS EGI OIV.1 4489KW 2ENS~SWGI92 ~ 4.169KV DIV.3 EHERGENCY BUS EG2 DIVE 2688KW 2ENS~SWG193 4.168KV OIV.2 EMERGENCY EG3 BID 4489KW ONSITE A.C. POWER SUPPLY

5

BIIKLE LQK FOR NN-IE tPSIA IILICJOol(LI)L3(L38 2NPS-SWGB8)

ONV) 2NPS-SWG883 03.8KV) 2NPS-SWGBBI a@SKY) 2NPS-SWGM3 (688 V) 2NJS-US4 (688V) 2NJS-US3 BUS 8 2NJS-US3 BUS A 2VBB-TRSI AUTOHATIC TRANSFER SWITCH (688 Vl (688 V)

(6MV) 2LAT-PN.388 2VBB-PNL381 2NHS-NCC996 BUS A N

N N

UPS UPS BAT-IC (688 V)

A (6BSV)

A (688 V) 2NJS-US6 2NJS-USS 2NJS-US6 AS UPS UPS BAT IB IO BAT-IA IC BAT-IA IA BAT-IC 2NJS-USI BUS C 2NJS-US4 BUS 8 UPS IN ~

BAT-IC UPS 3A BAT-18 UPS 3B N

(6MV)

(688V)

(688V) 2NJS-PNL')81 2NJS-PNL5M 2NJS-PNL688 (688V)

(688V)

~ INTERNAL BATTERY NO 2LAT-PNL188 2NJS-PIL482 ALTERNATE 2%S-SWGBIS HJGKV) 2NNS-SWG914 HJ6KV)

(4JSKV) 2NNS-SWG815 (6MV)

(688V) 2NHS-HCCB)6 2NJS-USS BUS B (6MV) 2NJS-US6 03J)KV) 2tfS-SWGM3 2NPS-SWGBBI 03.8KV) 2NPS-SWG883 0XSKV)

(689V)

(4J6KV)

HJ6KV) 2NJ6-US')

2NNS-SWG814 2NNS-SWG915 BUS 8 SINGLE LINE FOR CLASS IE LES - 2(L2B OXSKV) 03J)KV) 2NPS-SWMBI 2NPS-SWG883 HJSKV)

HJQOO 2ENSiSWGIBI 2ENS+SWG)83 (688V) 2EJSiUSI (6MV) 2EJSiUS3 (6MV) 2EJS+PNLIBBA 2EJSiPNL38BA BAT-2A UPS 2A BAT-28 UPS 28 (6MV)

(688V) 2LACiPNLIMA 2LACiPNL3M8 (688 V) 2EJSiUSI (688V) 2EJSiUS3 (4J6KV)

HJGKV) 2ENSiSWGI91 2ENSiSWGI93

3889.5A 2SPUY82 3889-5A 2SPMXBI 2HTX-HIA HAIN XFHR 292KV-24~KV 498/457MVA OA/FOA

&89-SA CORE GAP 2SPHXBI 63-1 63-1 pe65 HGA 25989-5A 39-1 39-1 HAA 3898-SA 3898-SA 2MTX-HIB 63-63-pe65 HGA 258M-SA 39-39-3BM-SA 2MTX-HIC 63-63-pe65 HGA 25889-5A 8-2SPUZBI 1289-SA 2SPMZBI pe66 2HTX-MID SPARE 59 51 IAC 22-2YXCNBI 63-4 63-4 pe65 HGA 87 87 PB65 BDD 39-4 39-c P864 BDO HAA 86-1 P866 HEA I

[

6 I pe66~HEA I

P866 HFA I

86-1 I P864~HEA I P864~HEA PSa< ~MFA I

86-1 IPBSS~HEA I

PBSS~HEA P865 HFA-15889-SA 2SPGZB2 45989-5A CORE GAP 2SPUY92 2GHS-Gl GEN.

13<e.iMvn BPF FIELD 25K'BBRpM 75PSIG H2 3PH 6BNZ I

pe67 psc z2-2YxcNBI I

r 86-1 P867~HEA I

6-P867~HEA IP867~ MFA 45988-5A 2SPUY82 2STX-XNSI AwhcHeem g PACjp g~g ATTACHHENT Q

LIST OF PROTECTIVE RELAY ACTUATED ON AUGUST 13 1991 Unit Protection Alt 1 Protective Rela Lockout Rela Action Ref. Dwg, 87-2SPMX01 Main Transformer Differential Protection Relay 86-1-2SPUX01 86-2-2SPUX02

~ Initiate Turbine Trip ESK-8SPU01

~ Initiate Fast Transfer ESK-8SPU02 to Reserve Station ESK-5NPS13 Transformer ESK-5NPS14 Unit Protection Alt 2 Protective Rela 87-2SPUY02 Unit Differential Protection Relay Lockout Rela 86-1-2SPUY01 86-2-2SPUY01 Action

~ Initiate Turbine Trip

~ Initiate Fast Transfer to Reserve Station Transformer ESK-8SPU01 ESK-8SPU03 ESK-5NPS13 ESK-5NPS14 63-2SPMY01 Fault Pressure Transformer Unit Protection Backu Protective Rela 86-1-2SPUY01 86-2-2SPUY01 Lockout Rela

~ Initiate Turbine Trip ESK-8SPU03

~ Initiate Fast Transfer Sh.

2 to Reserve Station ESK-8SPU03 Transformer Sh.

1 ESK-5NPS13 ESK-5NPS14 Action 50/51N 2SPMZ01 Protection Relay 86-1-2SPUZ01 86-2-2SPUZ01

~ Initiate Turbine Trip ESK-8SPU04

~ Initiate Slow Transfer ESK-5NPS13 After 30 Sec.

ESK-5NPS14 Block Fast Transfer After 6 Cycles Generator Protection Protective Rela Gen.

Phase OC During Startup 50-2SPGZ02 Lockout Rela 86-1-2SPGZ01 86-3-2SPGZ01 Action

~ Initiate Turbine Trip

~ Initiate Slow Transfer After 30 Sec.

Block Fast Transfer

~ This Relay Picks Up Only When Unit is Off Line Ref,D~

ESK-8SPG01 ESK-8SPG04 ESK-5NPS13 ESK-5NPS14 HSKl

II

AXTACEKNT 3 PAGE 3 of 3 Switch ear 2ENS*SWG103 Degraded Voltage Lockout Rela 27BA-2ENSB24 27BB-2ENSB24 27BC-2ENSB24 Action No Action Took Place Degraded Voltage Stays During Fault Conditions Ref.

Dw ESK-5ENS18 ESK-8ENS02 2ENS*SWG101 27BA-2ENSA24 27BB-2ENSA24 27BC-2ENSA24 No Action Took Place Degraded Voltage Stays During Fault Conditions ESK-5ENS14 ESK-8ENS01 2ENS*SWG102 27BA-2ENSC08 27BB-2ENSC08 27BC-2ENSC08 No Action Took Place Degraded Voltage Stays During Fault Conditions 807E183TY Sh 7

NSK1

J'

4

~a Agee F,-+

0 c~ pg Lc~ePi 2~

Le~ z) K

&~a>t F<.lg 1

v X Pr P

V

~ 4 U

P Vv>>'Pi JN; x~,

'NfV,i'l gl/lj />

V L~~~ xs 83 l

Lgve g./4/g

[Q

}

l~e Z

0fl e

~-re S

fC Jt NC,

/'(,<3 gA

Mkiiw:~.

~

~

~

0 0

()

I'&~g r=--

E~-Al E-n kc p

==

3 Q

s Vt le)I P(

trFC b H

D N

~ g> tll~

'F 9 I

tl l

3) 9

'Y 3

C l fJ U

y tgg "l

1

~ r.

Y!

W

~

~

0

~

e 0

U

<)

/

s~t3n 9 Ee.<

ep 4

X I

I 8Fo IgglPAAV Lt~e G Jg g~g 6 Cg.AI

't vV-

)

(1

()

i i Q

(:)

i)

~')

g)

V U

~,

4

A~7ACLW64 '$

VOLTAGE PROF [LE FOR THE NORMAL SUPPLY PPCE4opy TO UPS DURING Fl-"[ULTEO CONDITION 345/

, Qi 23,85K7 345KV 2MTX-XMIA4Bo C 2GMS G1 FOULTEO Col;OITION

[

Qi

[88KV N,

232K V x82.5K V 2

232KV Kv i

[6<V 21.55KY

[6KV 03 i3.sKv 24.9/

2NPS

~ 2STX.XNS I 7,5%

SMG883 03 16KV 21.55K V 16KV 8.9KV[

I I

((.9KV 8.9KV Q3 i3.53/

4.16KV

~ 2ATX-XS3

~ii 5.57.

0'.9KV 11.9KV N

2, 8KY 04 688V 2N JS 12V QROP 688/

~288. 128V US6 A2N JS-X3E 5.75%

~ 2VBB-XQ688 5.54 0>>

3 8.9KY 2.58K V 2

3.46K V 2.58KV 2.58KV 368V 482V 368V REGULATING 05 348V 478V 89V

((4V!

UPS 2VBB-UPS10 348V 65V 114V 45V 154V FROM SCRIBA SUBSTATION OSCILLOGRAPH RECORO 12V QROP ALLO((iIEO IN THE CABLE FROM THE LOAO CENTER TO THE UPS

~

I/

Z C

( C Pg I

'g QQ LJLI I W PiZBTECTION QL~IJA+2 1.

Q,E. HI 3 P 84'4 FiA.( O Iirl: Crrr QAI(r( ~p op XPIRR

. 8~X -XKl5 DiFFERE(vT(AI-p~T EcAQLJ l ~uT

%~LA'(

pg,~~ow I ~~un R<~'(

I/

ZC~- PNL. 8L +

u<n

><oTET(o< ALTa<Nhr~ Z

~.Q. HIP PS 45 LA( 5 I(~( '. (D L(v (T P((ALE 7.

L(( FFcR oLT'IIAI-VV(T 7(r((OE g rLFrFSR'OV(T*L

~9 Uun 7aeTEcaou Lcc.Kb~ ~~E~Y Llu W

'7(OOTS'T( o N LOLLOvF fTELA.Y Z.

@SABA(N SrEPL(p XPA(E..

H'TX-Xg((3 FAO(-T Pr<<<E (LO3)(Z) > (b3'/2.)

ZCE(.-'PNL globo UN IT

%fTKCT(QiU pAL(r gp c.e. His-PRt

'"'r(VlA(L( ~P L'P XPPIR. HFX-XM(

345 KY LIGuT.

('ROOLIQ C(.C. ((~~T) 5 c~ '.C.Ec PdL-. Ffo7

- " &E(vERATc(I2PR.oreczog (@pc(LuP)

'15

- Pg47 ML& lrv: ~i (rr~((ORATOR STAILFVP P((AOE P(LCrfKCWo(v (h((OILS k pic Itl PiAOIS)

CQ (oEoEILATAR DAcJvP P<'-o<ELofoLI LoeKovT gE~ j 8~vPEe~M 0<

W<~v'c R&-4-Y 2 p~P"f~ 5 P&POf 4OF Z.

2 E,k/5 ~ SvJc to~

Z ~ t'3.R l N

X,Fphg,. FOR..

ZCSS-rom (

u~ bMAo~&CC RE.MY F'~ 5 i x o u

')

27 SA ZEU S SZQ

~) zq ipse z.c+5 g) ~g rgb 2.EU &

5M& [G [

xF~<. FaR.

2 EZ'5 X IA

[Q ( l Q U~b~VsaAC.E, RE~A't Pc 4<5 i.) zq i&+ zeal p~<

2) Zj pl@

Z.~gg pZg 2.EACH A>'6 plows~ s~~ i.~

uncs Qs~~i ~c C.u rs ic~~

V~bcR.vo~<Ac a 'Sm9S FLA( 5 la; l.) 21 BA z.) z,z zs

~) z~ ec z a vs s~~

c c z.

p.v y. Se t L.BIZ.

<) ppg -+I+

SuJ&.OQ2-

%i - 2.- ZA6+ 55l (j.EST;) F~+~

p~~p+~~ s pP6K z.iA

lNTHINAl.CONIRSfONDSHCI NON1 1154%~

I$41MO PROg

h. R. -

LAk~

To Distr QT~kCHei~

PRE i~FW gv

>~~~ Nine Mile Point Nuclear station OATL 15 August 91 FfLXCOal Nine Mile Point Fire protection Program Post Event Xnterviews After interviews conducted today with Fire Chief Bernie

Harvey, and Firemen Fat Brennan and Nark Locurcio, and concurrence with Terry Vermilyea, System Expert Pire Detection and John Pavlicko of Caution Equipment Enc.,

X have reached the following conclusions.

1 ~

Of the 20 fire panels at Unit 2, 18 maintained a normal power supply o 2a.

Two Cire panels LFCP113 and 123 transferred to internal battery backup.'b.

These two panels wb'ile on battery willstill function normally as long as the 120 VAC is available in the LPCF, which it was.

There was no interruption or decrease of fire proteotioni detection/suppression at the local fire panels.

Fire Panels 849 and 200/1 being fed from QN did have a power interruption.

This would have left the control switches operable at Panel

849, (as they are fed from LPCP), but control Room with no Cire annunciation.

Any fire suppression/indication could also have been initiated locally.

ABA:dlc T, Tomlinson A. Julka (FAX 7225 - SN)

D. Pringle

QJQ 15 '91 15' ttt tKT NaÃC Ss~ DTS INTSHNALCOhllSNKNDRNCK INN'tMAeSN pg4gy Lo en ~

To FiI SUSJCCT

+~AD-+8'~

4 CLopf g 7

P.3 Nine M5.1e Point Nuclear Station 15 'August Ql RLR COOI Nin<<Mile Point, Unit 2 Fire Protection Program Feet Rvent 8/13/91 Interviews

?ire Dept. Personnel Interviews(

Poet Iveat of august 13( 1901 Bernie Harvey - Chief ln early fox coverage, interviewed for loss cf power in Contxol Building.

Lights blinked, loud noise (louder than ever heard in plant),

was in Fire Dept. office( told shift to get out into plant.

Pat Wilson was in Rx Bldg> switched radios to Channel 10, standard Fire Dept. practice if suspect lose of repeator.

Pat Bxennan was in the Foam Room and proceeded to the Chief(s desk.

Chief Harvey heard fire panel alarming when he got to Control Building.

Went past Fir~ Panel 114 in Turbine buildf.ng passageway(

no audible alarms(

seemed normal.

Mar}c Locurclo went to Panel 12d - 214 elev. while Chief Harvey went to Panel 127>> 244 elev. t these were sounding trouble alarm and D&

was clear.

Went past Panels 120(

121 (

128 f they were normal - no audible.

Pxior to Site Aea Emergency (SAE) message and evacuation being announced

- Pat brennan reported Panels R.B.

normal, called cn Gaitronics had to silence Panels 113 on I,B.

250 and than silenced all Fanele in R.Be ( Panels

101, 103(

104(

105(

10d(

10>

and XQ8.

Chier Harvey wae Cofnp to tri~e etage wet ln R h.

a.nd have aan in R.s.

cnarde Lynn Roon, accocpanled hy Larry cchaner, called hfa supervisor, when they saw transformer blow.

Chief Harvey wou14 have liked tO get to transfcrmer quf.oker for

".ire evaluation.

He feels it was at least one hour be fore,'

OV41QLCiOt1 e ch'f Harvey feels Fire Dept.

should have been part of 'nl investigation/inspection team with Operationse

QJQ f5 '9t 15~41 tt% JKT l%WT & CCttl DTS 1991 1'nterviev (Cont '4)

P.4 5'ff'Acp-m~y-g p~r-gg p g Pat Pat m in foam Room approximately 0550> heard loud noise, vent to Chief"'s office and asked what noise was.

Lighting dimmed, one string of lights off (NOTE:

these feed fram Emergency - UPS should have gone off)

Then he vent on rover -

heard alarms - which vere on vater treatment system panel, then vent to Panel 123.

There were no displays on DAX panel, vas blank no lights vere on.

Power lights vore off.

Trouhle Light blinking.

Rent to T.S, 261 NN, eigned sheet, stairtower dark'no

problem, knew vay around), Turbine Traok Bay 4imly lighted.

Wont to T.S.

306 - OK, sign>>4 sheet T B Svgr 277 OK) signed sheet T.Q.

250 by Feedpumps - noted not running by Panel 113 - no Lights on, no audible or trouble alarm estimat>>s time approximately 0605 Continued raver rounds to Panel 106 South Stairtover R.S.

249 vas alarming display sai4 "on internal clock" had tvo troubles displayed Went to Q.B.

215 tire panel 103 alarming - silenced R. B.

190 Fir~ panel 101 alarming silenced both panels were in trouble - unknovn R.B.

175 Signed sheet R.B.

261 SBCTS - OX Panel 105 - silenced troubles CQ2

Room, about t?)is time, evacuation alarm sounded went to Unit 2 Control Room assembly point Walked around with Pat Brennan on 8-15-91 ta Panel 123 and Panel

113, power on light vas burned out on Panel 123.

"Power on" Light was on, on Panel 113

~

.AJG 15 '91 15:41 N1 le MGKI Sc COPFl DTS P.5 Aff~Alm8Ng g P08

+opp post I+gt kuy. 13, 1991 Xnterview (Cont'4)

C Mark ZcNtok,o (Calle4 at home)

Has located in the Fire Dept. Office when lights flickered and noise was heard.

Radio communication was gone, Hear Here was out.

Chief Harvey directed personnel to cover vital areas.

Pat wilson was in RX Bldg.

Pat Brennan was roving T.B.

Bernie f Hark were tc cover Control Bldg.

Trip to C.B. uneventful Panels Passed in route)

Panel 114 Elect.

Bay Elv.

261'anel 120 C. 5 ~ Elv. 261

'anel 128 C.5. Elv.

261~

Panel 121 C. B. Elv, 261'anel 125 C.B. Zlv.

261'anel 127 C.B. Elv.

2i4'anel 126 C.B. Elv.

214'ormal Normal Normal Normal Normal Trouble, Horn sounding-SU,enaod Trouble Horn sounding-lilenoed, also an amber light was lit on panel Checked valve room on C.b. elv. 2ii'ight was on in room.

No indication of system actuation.

stairwells were dark, Elv, 261'.B. was dark.

s.A.E. announcement and reported to Control Room.

I

A<7~mgur 7 pt)c,p k~

NMP2 LIGHTING SYSTEM POWER SOURCE AND MINIM)St ILLUHINATIONAVAILABLE-CRIT]CAL AREAS MODES OF OPERATION PAGE I

LOCA St SEISMIC WITH LOOP TRANSIENT WITHOUT LOOP ILLIIL SIXRCE6 PROVIDED BY POWER SIXRCE MIN.

AVAIL (FOOT CAN)LE)

X MIN.

POWER AVAIL.

SDUIKES

~~~

<FOOT BY POWER CANDLE)

SIXRCE POWER SIXRCES X

HIN.

AVAII POWER BY POWER CANDLE)

PROVIDED

)FOOT ~ES SOURCE X OF PROVIDED CANDLE)

(FOOT SOURCE X OF POWER SOURCES PROVIDED SOURCE MIN.

ESSENTIAL AVAIL.

LIGHTING (FOOT UPS PAtEL CANDLE)

IL CIXITROL ROOH IDPERATING AREA 5 RELAY PANEL AREA)

CONTROL BLDL EL.

336'E-65E N)RMAL 58 ESSENTIAL

)9 EHEROE 49 8 NXR BAT.PACK ~

ESSENTIAL 18 8 NXR BAT.PACK 159 NORMAL ESSENTIAL EMERGE 8 NXR BAT+PACK ESSENTIAL 8 NXR BATo PACK NORHAL 58 ESSENTIAL 18 EHERGENC 48 S NXR BAT.PACK NONE S NXR BATT.PACK HANXACTIRER-EXIDI (TYPICAL)

CONTROL ROON MRTM-SOUTH CDRMDORS)

CONTROL BLDL EL 3P&'E-&5E CONTROL ROOM (SHIFT SLPERVISOR OFFICE)

CINTROL BLDL ELo 3P&'E-65E RELAY AND CO)4'UTER ROOM RELAY PAMPAS)

CONTROL BLDL EL 28F-6'E~

RELAY AND COtPUZER CQ4%6EA IKXNO CONTROL BLDG.

EL 288'-6'E-650 HORHAL 189 BAT.PAtx NONE 8 NXR NRHAL 58 BAT.PA)X NONE S NXR S NXR BAT+PACK S NXR BAT PACK 8 NXR BAT.PACK NORHAL 58 ESSENTIAL 18 S NXR BAT. PACK NORHAL 58 ESSENTIAL 18 EMERGENC 49 S NXR BAT. PACK NORHAL

'Q ESSENTIAL 18 8 NXR BAT. PACK S NXR BAT.PACK NORHAL ESSE NTI EMERGENC S NXR BAT+PACK NORHAL ESSENTIAL S NXR BAT.PACK ESSENTIA.

S NXR BATe PACK YES 16 S NXR BAT+PACK NORHAL EHERGE S NXR BAT.PACK ESSENTIAL EHERGE S NXR BAT.PACK S NXR BAT+PACK YES 16 S NXR BAT.PA)X NORMAL 58 ESSENTIAL 18 EMERGENC i8 S NXR BAT.PACK ESSENTIAL 18 EMERGENC

<9 S NXR BAT.PACK ESSENTIAL 18 8'NXR BAT+PACK 2VBB-UPSID 2VBB-UPSID 2VBB-UPSID N THE LIGHTING WINGS, CIRCUITS TARTING WITH AN 'N ICATE NRMAL POWI WITH A %'H)ICATE ESSENTIAL POWER, WIT)

AN 'E'INDICATE EMERGEN:Y POWER.

AtrocmC~r 7

PAcEX,~<<

'MP2 LIGHTING SYSTEM POWER SOURCE AND MINIMUM ILLUMINATIONAVAILABLE HOOES OF OPERATIDN PAGE 2

LOCA tt SEIS)GC WITH LOOP TRANSIENT WITHOUT LOOP It OF

~R ILLLSL SOURCES PRDVIOED BY POWER SOLSCE MIN.

AVAIL.

LFOOT CQOLE)

ILLLSL 2 DF PRLMDED AVAIL BY POWER C~E)

LFOOT SOURCE POWER SDLRCES 2 OF AVAIL.

POWER BY POWER CAN)LE)

LFOOT SOURCES SOURCE Hl)L

/ OF ILLLSL AVAIL POWER ILLUM PROVIDED LFDOT SOURCES PROVIDED BY POWER CANDLE)

SOURCE BY POWER SOURCE MIN.

AVAIL.

LFOOT CANDLE)

ESSENTIAL LIGHTING UPS PANEL IIL RELAY AN)

CQ%'UTER ROOM LCLXLRIDLRS)

CONTROL BLDG.

EL 288'-6'E-65D DIESEL GEtKRATOR BUILDING tWORKING AREA)

EL 261'E-6BC DIESEL GEtKMOR BLBLDING LELECTRICAL EQ)IPMENT AREA)

EL.

261'E-6BC DIESEL GENERATOR BUILDING LGENERAL AREA)

EL 261'E-68C ESSENTIAL 18 S NXR BAT.PAIX N)RMAL 78 S NXR BAT.PA)X ~

NORHAL 78 ESSENTIAL 18 EMERGE 28 S NOLR BAT.PAtx N NORHAL 98 ESSENTIAL 18 S NXR BAT.PALX ~

BAT.PACK S NOLR NORHAL 78 ESSENTIAL 18 S NXR BAT.PACK NORMAL 78 ESSENTIAL 18 EMERGENC 28 S NXR BAT.PACK NORMAL 98 ESSENT)AL 18 S NXR BAT.PACK NORMAL ESSENTIAL S NXR BAT.PACK NORHAL ESSENTIAL EHERGE S NOLR BAT.PACK NORHAL ESSENTIAL S NOIR BAT.PACK NORHAL ESSENTIAL S NOLR BAT.PACK YES YES S NXR BAT.PACK NORHAL S NXR BAT.PACK NORHAL ESSENTIAL S HOLR BAT,PACK NORMAL ESSENTIAL S NXR BAT.PACK YES YES NORHAL 98 ESSENTIAL 18 BAT.PACK S HOLR NORMAL 78 ESSENTlAL 18 EMEROE NC 28 S HOLR BAT.PACK NORMAL 78 ESSENTIAL 18 EHERGENC 28 S HOLR BAT.PACK NORHAL 98 ESSENTIAL 18 S NOtR BAT.PaX 38 2VBB-ASS 2VBB-UPSID 2VBB-UPSID 2VBB"UPSIO

11

HHP2 LIGHTING SYSTEH POWER SOURCE AND HINIHJH ILLUHINATIONAVAILABLE

. HXKS OF OPERATION PAGE 3

LOCA 8 SEISHIC WITH LOOP TRANSIENT WITHOUT LOOP POWER SO(SCES 2 OF PROVIDED BY POWER CA)b)LE)

(FOOT SOURCE POWER SOURCES Hl)L 2

I ROY)GEO AVAIL.

BY POWER (FOOT SOURCE POWER SOURCES 2 OF ILLUH.

PROVIDED BY POWER SOURCE HIlL 2

AVAIL+

POWER PROVIDED CANDLE)

BY POWER SO(SCE HIN.

I OF AVAIL.

POWER (FOOT S~ES PROVIDED CANDLE)

BY POWER SOURCE HIN.

ESSENTIAL AVAIL.

LIGHTIW (FOOT UPS PAtKL CANDLE)

IIL REHARKS REHOTE SHUT DOWN ROOH CONTROL BLDL EL 261'E-&5C EE-165C EHERGE 8 8XR BAT.PACK YES 16.5 8 HOLR BAT.PACK YES EHERGE 8 HOLA BAT.PACK YES IL5 EHERGENC YES BAT.PACK 8 HOLR NORHAL YES EHERGENC YES BAT.Pox N

E 8 dna STANDBY SWITCHGEAR ROOH a) SWG. PANELS (2) HCC FRONTS CONTROL BLDG.

EL 261'E-65C EE-165C EPICS SWITCtSEAR RQOH CONTROL BLDG.

EL 261'E-65C EE-165C STANDBY SWITCH%EAR ROOH (CORRIDORS)

CONTROL BLDL EL+

26)'E-65C EE-16SC N(RHAL ESSENTIAL EHERGE 8 BXR BAT.PACK 8 BXR BAT+PACK 8 NXR BAT.PACK ESSENTIAL 8 HOLR BAT.PACK NORHAL ESSENTIAL EHERGENC a exa BAT.PACK NORHAL ESSENTIAL 8 NQLR BAT.PACK NORHAL ESSENTIAL EHERGE 8 N(XR BAT.PACK NORHAL ESSENTIAL EHERGE 8 (XXR BAT.PACK ESSENTIAL 8 HRR BAT.PACK YES 35 15 ESSENTIAL YES EHERGE YES BAT.PACK 8 HXR NORHAL NONE ESSENTIAL YES EHERGENC YES BAT,PACK 8 IRLR ESSENTI(V.

YES BAT.PA(X 8 NXR 15 NORHAL YES ESSENTIAL YES EHERGENC YES 8 HOLR BAT.PACK NORHAL YES ESSENTIAL YES EHERGENC YES 8 (K)LH BAT.PA(X NORHAL YES ESSENTIAL YES BAT PACK NONE 8 N(na 2VBB-UPSID 2VBB-UPSID 2VBB-UPSID

NHP2 LIGHTING STSTEH PO)fER SOURCE AND MINIHUH ILL'NfINATIONAVAILABLE HODES OF OPERATION PAGE LOCA 8 SEISMIC

))ITH LOOP TRANSIENT NITHOUT LOOP STAHDBT S)fITCHGEAR RDOH IEAST CABLE CHASE AREA)

COHTRCL BUXL EL Ãl'E~

EE-)at)c PDVER AVAIL HIM.

SOURCES fFDDT BY POVER C~E)

SRSCE NORHAL 199 ESSENTIAL YES POVER AVAIL HIN.

SOURCES fFOOT BY POVER CAtOLO SOLACE NORHAL IBB ESSENTIAL YES Pom

'"LU)L PROVIDED BY POVER Sm)RCE ESSENTIAL NONE HI)L AVAIL.

POVER fFOOT SOURCES CANDLE)

NORHAL ESSENTIAL ILLlH HIN K

PROVIDED AVAIL BY PDVER C~E) fFDOT SOURCE YES POVER SOURCES ILLUH PROV)DED BT POVER SOURCE HI)L ESSENTIAL AVAIL.

LIGHT1tO lFOOT UPS PANEL CANDLE)

ID.

2VBB-UPS)0 REMARKS COMHON INSIDE AREAS ESSENT IIL I99 ISTAIR)fATS) fALL IRI)VDOS)

ESSENTIAL 198 ESSENTIAL IBB 199 2VBB-UPSIC 2VBB-UPS10 NOTE a HOUR BATTERY PACK PROVIDED It4.T OH SAFE SHUTOOVH PATHS.

COHMCN INSIDE AREAS fEGRESS PAT)a fALL NhtfltCS) 8 tKXA BAT.PACK NORHAL YES ESSE)fl)AL YES 8 HOLA BAT+ PACK ESSENTIAL YES 8 )NUR BAT+pAcK 8 HDtA BAT+ PACK NORHAL ESSENTIAL YES 8 )K)UR BAT+ PACK 2YBB-UPSIC 2VBB-UPSID HDTE 8 HOUR BATTERY PACK PROV)DEO CN.T OH SAFE SHUIDOVN PATHS.

BAT.PACK 8 talA COtafDN ESSENTIAL 199 EXIT SIGNS OV.L DRIVI)OS)

BAT. PACK 8 H)XA ESSENTIAL 199 BAT PACK YES a

tK)UR ESSENTIAL tONE 8 HOtA BAT. PACK ESSENTIAL YES IBB a

tK)UR BAT. PACK ESSENTIAL IBB 2VBB-UPSIC 2VBB-UPSID

Af'fnareevV 7 IAAF g NMP2 LIGMTING SYSTEM POWER SOURCE AND MINIH)M ILLUMINATIONAVAILABLE M(K)ES OF OPERATION PAGE LOCA h SEISMIC WITH LOOP TRANSIENT WITHOUT LOOP STA)4)BY SWITCNGEAR RMM (GENERAL AREAS)

(XNTROL BLD(L E(

2GI'E-65C EE-)65C CtÃNN IHSIDE AREAS (STAIRWAYS)

(ALL DRAWINGS)

POWER AVAIL MIN.

SOURCES (FOOT BY POWER CA)NLE)

SOURCE ESSEHTIAL YES ESSENTIAL IBQ POWER SOURCES NORMAL ESSEHTIAL ESSENTIAL

/

MIH.

AVAIL.

POWER PROVIDED (FOOT Q)URCES BY POWER C~E)

SDLRCE NORMAL ESSEHTIAL ESSENTIAL X

ILLUM.

PROVIDED BY POWER SOURCE AVAIL.

POWER Ml)L (FOOT'ES PROVIDED CA)K)LE)

BY POWER SOURCE ESSEHTIAL YES ESSENTIA.

188 MIN.

AVAIL.

POWER (FOOT SOURCES CANDLE)

NORMAL ESSENTIAL ESSENTIAL 2

PROVIDED AVA'L-BY POWER C NXE)

(FOOT SOURCE YES IBB ESSENTIAL LIGHTING UPS PANEL ID.

2VBB-lFSlO 2VBB-lPSIC 2VBB-UPSID CROCK INSIDE AREAS (EGRESS PATH)

(ALL DRA)O(GS) b NOIR BAT+PACK ESSENTIAL YES S )K)LR BAT.PACK NORMAL ESSEHTIAL YES YES S )K)LR BAT.PACK NORMAL ESSENTIAL S HOLA BAT PACK ESSEHTIAL YES S HOLA BAT PACK HORMAL ESSENTIAL YES 2VBB-UPSIC )NTE S HOLR 2VBB-UPSID BATTERY PACK PROVIDED ONLY ON SAFE SWTOOWN PATHS.

b HSLR BAT, PACK S HOLR BAT.PACK S HOLR BAT.PACK YES S HOLR BATs PACK YES S HGLR BAT.PACK C&tQN IHSIDE AREAS EKIT SIGNS (ALL Df(AWINGS)

ESSENTIAL IBB ESSENTIAL ESSEHTIAL ESSENTIAL IBB ESSENTIAL IBB 2VBB-UPSIC 2VBB-IfSIO

l'll

ISIP2 LIGHTING SYSTEM POWER SPLRCE AND MINIINM ILLUMINATIONAVAILABLE MOPES OF OPERATION ATf4~6nlT' P~~ gong PAGE 5

TRANSIENT LOCA dc SEISMIC WITH LOOP WITHOUT LOOP CLXITRPL RXN LEAST-WEST CORRIDON CONTROL BLDG EL. 386 EE-65E Q)RTfKAST STAIRS CONTROL BLDG.

EE-65E EE-650 EE-165C EE~

ILLLH.

SDLRCES PROVIDED BY POWER SXRCE ESSENTIAL 33@

b HXR BAT.PACK ESSENTIAL 198 S WXR BAToPACK MIN.

AVAIL POWER LFppT SOURCES PROVIDED CA)6)LE)

BY POWER SOURCE S HXR BAT.PACK ESSENTIAL 189 S BXR BAT.PACK MIN.

AVAIL.

LFOOT CANDLE)

POWER AVAIL.

MIN.

SDLRCE$

LFOOT BY POWER CAINLE)

SOURCE ESSENTIAL NONE

$ HXR BAT.PACK ESSENTIAL NONE

$ HXR BAT.PACK POWER SDLRCES M)R)LAL ESSENTIAL S HXR BAT PACK b HOLR BAT.PACK ILLLSL

/

pROVIDED AVAILe BY POWER CANDLE)

(FOOT SOURCE POWER SOURCES ESSENTIAL S NXR BAT.PACK ESSENTIAL S HNR BAT.PACK MIN.

ESSENTIAL (FOOT UPS PAHEL AVAIL.

LIGHTING BY ~R CANDLE) 10 SOURCE 2VBB-L$%$

333 VBB-iX%10 SOUTHWEST STAIRS CMfRPL BLDG EE-66B EE-650 EE-66F EE-65C EE-165C INTER BAY RA)4 EL 258 Tp EL 261 EE-7%

ESSDITIAL 189 S IXXR BAT.PACK YES ESSE NTIN.

YES S IKXR BAToPACK ESSENTIAL 188 S HOLR BAT. PACK NORHAL YES ESSENTIAL YES S HXR BAT+PACK ESSENTIAL NONE S HOLA BAT.PACK

'YES ESSENTIAL NONE S INLR SAT PACK S HOLR BAT.PACK NORMAL ESSENTIAL b NXR BAT+PACK ESSENTIAL S HOLR BAT.PACK ESSENTIAL S HOLR BATs PACK 189 2VBB-UPS10 2VBB-UPS10

1

%%>2 I.IGHTING SYStEH POwER SOURCE ANO HINI)K)H ILLUHINAtlONAVAILABLE HOOES OF OPERATION PAGE 6

LQC4 LOCA 5 SEISHIC WITH LQIP TRANSIENT WITHOUT LOOP X IF

~R ILL)M.

PROVIDED BY POWER SOURCE Ht)L 4VAIL

)FOOT CANOLE)

ILLU)L X OF PRQVIOEO AV41 BY POWER CANOLE)

<FOOT SQRCE POWER SOURCES X

HIN.

4V4IL POWER

<FOOT SQRCES BY POWER SOURCE tu~

BY POWER ~E)

PROV IOEO FOOT'V4IL, PQWER SQRCE

% OF ILLIH.

PROV IOEO BY POWER SOURCE HIN.

ESSENtIAL 4VAIL L[GHTINO (FOOT UPS PANE)

CANIXE) 10.

REACTIR IXOO.

EL 353'-N'EETJ NQRHAI.

YES SSENTIAL YES

'ftAL YES N)'IAL NQRHAL ESSENTIAL YES NQRH41.

SSEN TIAL YES YES ZVBB~IO REACTOR KOG.

AUX 84YS NQ)TH EL215'~

EE&TL 8 HOIR BAT PACK NORHAL YES SSENTIAL YES 8 NXR BAT.PACK NTIAL YES 8 HOUR BAT PACK NORHAL SSENTIAL 8 HQR BAT PACK NQRHAL ESSENTIAL YES YES 8 HQR BAT PACK NORHAL SSENTIAL YES YES ZYBB~IC REACfOR M)O.

AUX BAYS SQ))H EL 2)5'448'-

EE&7L 8 NXR BAT. PACK NQRHAL YES TIAL YES 8 HOUR BAT.PACK

'IIAL YES 8 HOUR BAT. PACK NORHAI ENTIAL 8 HQR BAT. PACK NORMAL SEN'flAL YES YES 8 HQR BAT. PACK NORHAI SSENTIAL YES YES ZVBB~IO AUX SERVICE BLOG. SOUTH EL 261'E&7P 8 HQ)R BAT.PACK NQRHAL YES SSENTIAL YES 8 NOIR BAT.PAOC NQRHAL YES TIAL YES 8 HQ)R BAT. PACK NORMAL YES 8 HQR BA'f.PACK NQRHAL SENrtAL YES YES 8 NQR BAT.PACK NORHAL SSEN)'IAL YES YES ZVBB~IO SCREENWEIL SLOG.

EL 261'E-728 8 NXR ltAT PACK HORHAL YES SSENTW.

YES 8 NXR BAT. PACK TIAL YES 8 HQR BAT. PACK fIAL 8 HQR 84T. PACK SENT IAL YES YES 8 HQR BAT. PACK NQRHAL SSEN'ftA).

YES YES 2VBB-UPSIC 8 )KXR BAT.PACK 8 HQR BAT. PACK 8 HOUR BAY.)AC

'YES 8 NXR 84T. PACK YES 8 HQR BAT PACK

~ LGHTIl4 IN REACTOR BUILOING IS POWEREO FR(H PLANT EHERGENCT POWER OISTRIBUTIQI STSTEH.

OURING LOC4, 'fHIS NQI-IE LIGHttNG SYSTEH POwER IS TRIPPEO BY AN ACCIENt SIGNAL

J8%'2 LIGHTING SYSTEM POVER SOURCE ANO MINI%H tLLUHINAI'IONAVAILABLE PAGE 7

HQQES QF 6%JIATION LOCA 8 SEISHIC

'VITH LOOP TRANSIENT VITHIXJT LOOP TIR BLDG.

Et 2I5 EEOC SPENT RKL CQm.DKI AREA 8 OF SIXRCES TLLI84.

BT neCR SCXJRCE HJRHAL YES S%NTIAL YES 8 NXR BAT,PACK HIM.

AVAIL IFOOT CAHJLEJ 8 HXR BAT,PACK YES HIJL 8 OF PRQVIOEQ AVA~

BY POVER IFOOT MIRCE NORHAL Nt&

SSEMTIAL

'IES POVER SmRCES NORHAI.

SSEMTIAL 8 HXR BAT, PACK YES ESSENTIAL 8 NXR BAT, PACK HIIL ILL~

AVAIL POVER BY POVER (FOOT SIXRCES SOURCE 7

OF ILLUH.

PROV!OED BY POVER SOURCE YES HBL AV40 POVER CANm.E)

BY POVER SOURCE NORHAL YFS ESSEM'f IAL YES 8 NOIR BAT. PACK HIIL ESSENTIAL AVAIL.

LIGHTING IFOOT UPS PAWL CANQLEI IL 2VBB~IC REMARKS TOR BLDG.

EL. 2<8 M EE+7D

%7ITIAL YES 8 HXJR

, BAT.PAI7(

NORMAL YES 8 HXIR BAT, PACK YES NORHAL HOKED ESSENTIAL YES 8 NXR BAT, PACK YES SSEMTIM.

8 NXR BAT, PACK YES YES YES SSEMTIAL TES 8 HXR BAT. PACK VBB~SIC NORHAL YES SIJITIAL YES 8 HXJR BATe PACK 8 HXR BAT, PACK YES NORHAL NO%'TIAL

'YES 8 NXR BAT PACK YES HORHAL SSEMTIAL 8 NXR BAT, PACK YES YES HJRHAL YES SSDITIAL YES 8 MmR BAT, PACK 2VBB~SIC ACCESS P4TH t74.Y QRHAL YES SDITIAL YES 8 HXR BAT,PACK 8 HXR BAT.I ACK YES YES 8 NXR BAT, PACK YES MORHAL SSEMTIM.

8 NXR BAT PACK YES YES HJRHAL YES

$%MTIAL YES 8 IKXR BAT, PACK 2VBB-ITIC TOR BLQd.

Et 328'-l8'%7M WeeL YES SEJITIAL YES 8 HXR Bar. I ACX 8 HXR BAT. PACK YES ESSOITI YES NDIHAL SSEMTW.

8 HXR BAT. PACK YES ESSENTIAL 8 NOIR BAT. PACK YES YES NORMAL YES SSBITIAL

'YES 8 NXR BAT. PACK 2VBB-IPSIQ i NIXBIAL LIGHTING IM REACTOR BUILDING IS POVERED FROM PLANT EHERGEMCY PQVER OISTRIIXJTION SYSTEH. OIRIMG LOCA, THIS NON-IE LIGHTII4) SYSTEH PQVER IS TRIPPEO BZ AN ACCIDEMI'IGNAL~

avTmH>~f 7 pres g'+

%(PZ LIGHTING SYSTEH POwER SOURCE ANO HINIHUH ILLUHINATIQNAVAILABLE PAGE 8

HOOES OF OPERATION LOCA 8 SEISMIC WITH LRP TRANSIENT WITHOUT LOQP ACT6I BLOC.

ST4IRS EE-67E EE-67F E'E-67O EE%7H EEWTJ PROV IOED POWER BY PQWER S(X)RCE SDITIAL IBB HIN.

4V4ll (FOOT C(WE)LE)

POWER ILL~

AvAIL.

HIM.

SOURCES (FOOT BY POWER S(mRCE SENT IAL 188 POWER SOURCES I LUH

HIN, PROVIOEO 4"4'"-

8'I POWER ~E)

(FOOT SOURCE POWER SOURCES ILLUH.

AVAIL

~ER HIN.

Y POWER C~E)

PROvtOEO

(~T SOURCE ILLUIL PROvtOEO BY POWER SQLIICE te8 HIN.

ESSENTtAL 4V4IL LIGHTINO (FOOT UPS P~

CAIZILEI

[O.

ZVBB~LB REHARKS T& BLOC.

AUX.BAYS N(XITH ST4IRS EE<7L 8 )CUR BAT,PACK TIAL IBB 8 HOUR BAT, PACK YESi 8 IXXXI BAT,PACK SENTIA(.

8 IKXXI BAT PACK 8 H(XXI BATe PACK SSENTI4L ZVBB~LC ACTUI BLOG.

AUX.BAYS SOUTH ST4IRS EE.67L 8 HER BAT, PACK SENT IAL IBB 8 Ie)R 9AT. PAO(

ESSENTIAL YES0 8 IKNXI BAT,PACK 8 IXXXI BAT PACK IBB 8 H(X)R BAT, PACK ZVBB-UPSLO 8 IOLXI 84T, PACK 8 HOUR BAT+PAO(

YES' IKWXI BAT PACK YES 8 IKX)R BAT, PACK 8 IKXXI BAT+PACK HORHAL LIGHTING IN REACTOR BUILOING IS POWERED FROM PLANT EHERGENCY POWER DISTRIBUTION SYSTEH. IXNING LOCA. THIS NON-IE LIGHTING SYSTEH POVER IS TRIPPEO BY AH ACCIDENT SIGNAL.

NMP2 LIGHTING SYSTEH POWER SOURCE AND MIHINJM ILLUMINATIONAVAILABLE MODES OF DPERATIDN PAGE LOCA 4 SEISHIC WITH LOOP TRANSIENT WITHOUT LOOP TLRB. BLDG.

GRMM) FLOOR CORfODOR ELe 258'E-66B POWER SOURCES ESSEHTIAL X

ILLLÃ.

PROVIDED BY POWER SOLRCE MIN.

I AVAIL.

POWER ILL (FOOT SOURCES PROVIDED CA)K)LE)

BY POWER SDLRCE NORMAL YES ESSEHTIAL YES MIH.

AVAIL.

IFOOT CANDLE)

POWER AVAIL.

SOURCES IFOOT BY POWER C~E)

SOURCE HORMAL NONE ESSENTIAL NONE POWER SOURCES NORMAL ESSENTIAL OF PROVIOEO BY POWER C~E)

(FOOT SOURCE YES POWER SOURCES ESSENTIAL ILLU)L X OF pRovloEo AVAIL.

BY POWER C~E)

(FOOT SOURCE YES YES ESSENTIAL LIGHTING O'S PANEL ID.

2VBB-LPS1D SAFE S)IJTDOWN PATH O)l.Y No ELXJIPPKNT TURLBLDL CLEAN ACCESS AREA EL. 2QY EL 2Q'L.

288'-8'L.

396'E-66H TtSL BLOL CLEAN ACCESS AREA STAIRS EL+

258'L 2Q'L 288'-8'L 386'E-66H 8 NXR BAT+PACK NORMAL ESSENTIAL 8 )KXR BAT.PACK b HOLR BATo PACK 8 HLXR BAT+PACK NORMAL YES ESSENTIAL YES 8 HL)LR BAT.PACK ESSEHTI)V.

YES 8 HOLR BAT.PACK 8 HOLA BATs PACK YES S HXR BAT PACK YES ESSENTIAL NONE 8 HXR BAT PACK NORHAL NONE ESSENTIAL NONE 8 NXR BAT+PACK NORMAL ESSENTIAL S )NLR BAT.PACK S HOLR BATo PACK YES YES 8 HOLR BATe PACK NORMAL ESSENTIAL S HOLR BAT.PACK ESSENTIAL 8 HOLR BAT.PACK YES YES YES 2VBB-UPSIO 2VBB-UPSIC 2VBB-UPSIC SAFE SHUTDOWN PATH D)LY NO EOUIPMENT SAFE SWTDOWN PATH M.Y ND EIXJIPPKNT

IU

ACCESS ROUTE TAKEN BY OPERATOR FROM CONTROL ROOM TO UPS ROOM IN NORMAL SWITCHGEAR BUILDING ON 8-13-91 The following route was taken by operator from control room to go to UPS room in switchgear building to transfer alternate power source to UPS units

,Operator left the control room EL 306 through south door and proceeded to west.

Then he turned north along the corridor on the westside of the control room. Then he exited the control room building through north west door EL 306 to Auxiliary building.

He then took the stairway just south of the elevator to go to EL 261.

Then at EL 261 of Auxiliary building, he proceeded to south and entered the corridor

( Electrical equipment Tunnel). From the corridor he entered the normal switchgear building EL 261 and proceeded to stairway located in the center of the building (West half) down to EL 237 where the UPS units 2VBB-UPS1A, 1B, 1C and 1D are located.

He then transferred the UPS power to maintenance power source.

After restoring power to the above UPS units, the operator proceeded to UPS 2VBB-UPS1G via the door on the east end of the room, went south down the hall, through the door on his left (eastside) and entered the control building.

He then took the stairs down to EL 214 where UPS 2VBB-UPS1G is located and transferred the power to the maintenance power.

CONDENSATE/FEEOMATER SIMPLIFIED SKETCH lCQOCM JEANS FEll4 I

I I

I I

I QlZEMSATE BOOSTER I

I FVRi

)

FOI)MFlO PAWL I

I I

I I

I I

I I

I I

i FEEOvATER COMlROL LOO)C

)

I I

)

I I

LVIB FL FVg PRESS)I)RED REAC)TX)

VESSEL O'RVI C(ÃH6ATE PAP HIM.FLOM IEgKR ICQI+QN COMPENSATE )MISTER PNTP HDLFLOM TEAOER FEEOPUH'lM. FLOV TRACER FOXBORO PANEL TRAIN A - 2CEC-PNL825 TRAIN B - 2CEC-PNL826 TRAIN C - 2CEC-PNL827 FEEDWATER CONTROL PANEL 2CEC-PNL6I2

(

TYPICAL TRAIN A, 8,8, C

CROSSOVER HEADERS BETWEEN TRAINS PROVIDE CROSSOVER FLOW

Oe GE Industrial A,T rAQftttl+IIf'/ Q a power Systems PRO Pomr Generation Services Department General Electric Company 3532 James St.. PO. Bott 484t, Syraorse. Ny t322t NIAGARA MOHAWK POWER CORPORATION NINE MILE POINT NUCLEAR STATION UNIT g2r GENERATOR g180X632 GENERATOR INSPECTION POSSIBLE PHASE-TO-PHASE FAULT cc:

NIAGARA MOHAWK POWER CORP.

R. Abbott N. Kabarwal M. McCormick GENERAL ELECTRIC COMPANY August 28, 1991 L. Jordan (37-3)

W. Judd S. Kolb R. Smith W. Turk Mr. Anil K. Julka NIAGARA MOHAWK POWER CORPORATION 301 Plainfield Road North Syracuse, New York 13212

Dear Mr. Julka:

Due to the August 13, 1991 failure of the phase B step-up transformer on Nine Mile Point Unit g2, General Electric Generator Engineering recommends performing a thorough visual inspection of the generator stator end winding support system at the next convenient opportunity.

The inspection should include all accessible components of the stator end winding support

system, including stator bar end arms, blocks, ties nose rings, and outer axial supports.

This inspection should be accomplished by a

GE Generator Specialist trained to detect the potentially subtle indications of damage.

The above recommendation is based upon the possibility of phase-to-phase generator short circuit currents through the failed transformer as high as 5pu.

The initial recommendation for the immediate generator inspection considered the possibility of higher

currents, resulting in several times greater end winding forces.

Physical evidence of high current forces at the transformer.low side was the primary driver for this recommendation, since measurements of generator currents were not available.

Our engineers continued to review the limited data available and subsequently concluded that currents high enough to do probable damage to the generator were not likely.

This conclusion was reached primarily by considering a

measurement of depressed generator voltage during the incident, inferring generator

currents, and specific capabilities of the generator design.

'k

Page II Mr., Anil K. Julka August 28, 1991 Should you have any questions regarding this recommendation, please contact me.

Very ruly yours J

eph

. Kir ch Ma ager Engine ring Services Pow r Generatio Services "SYRACUSE OFFICE JAK/bs JAK-059

.r I ~

T