ML20127E289

Rept on Sources & Effects of Electrical Transients on Electrical Systems of Commerical Nuclear Power Plants

ML20127E289
Person / Time
Issue date:	09/30/1992
From:	Rourk C NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:
Shared Package
ML20127E287	List: ML20127E284 ML20127E289 ML20128E839
References
RULE-PRM-50-56 NUDOCS 9210050355
Download: ML20127E289 (51)

Text

._ _

r t f e Report on the sources and Effects of Electrical Transients on the Electrical Systems of Comraorcial fluclear Power Plants Prepared by: Chris Rourk Engineering Issues Branch Division of Safety Issues Resolution Office of liuclear Regulatory Research U.S. 10uclear Regulatory Commission September 1992 flote : This report supports ongoing RES and liRC activities, and does not represent the position or requirements of the responsible 11RC program of fjcesF'-^ N -

I '

006024 q3e,pm

( -

v v f 4-Table of Contents

1. Summary . . . . . . . . . . . . . . . . . . . . . . . . 2

2. Introduction . . . . . . . . . . . . . - . . . . . . . .- . 3

3. Electromagnetic Pulse (EMP) . . . . . . . . . . . . . . 4

4. Geomagnetically Induced Current (G1C). . . . . . . . . . 6

5. Terromagnetic Effects . . . . . . . . . . . . . . . . . 7

6. Miscellaneous S?urces . . . . . . . . . . . . . . . . . 8

\

7. Switching Surges . . . . . . . . . . . . . . . . . . . . 9

8. Lightning Strike Phenomena . . . . . . . . . . . . . . . 10 8.1 Power Line Transients . . . . . . . . . .. . . . . . 10 8.2 Local Striko Effects on Power Plants . . . . . . . 14

9. Summary of Electrical Transient Phenomena . . . - . . . . 15

10. Lightning Events . . .. . . . . . . . .- . . . . . . . . 16 10.1 Power Line Transient Events. . . . . . . .. . . . 16 10.1.1 Line strike, no effect . . . . . .. . . 18 10.1.2 Line strike, plant effect reported . . . 19 10.1.3 Similar Lightning and Loss .of Line Events . . . . . . . . . . . . . . . . . 21' 10.1.4 Events with fire or loss of fire protection . . . . . . . .. . . . . . . 22 10.2 Local Strike Effects . . . . . . . . . . . . . . . 22 10.2.1 CRD Power Supply Overvoltage Protection Trips . . . . . . . . . . . . . . . . . . 22 10.2.2 APRM Spurious Signal . . . . . . . . .. . 29 10.2.3 Fire and Fire Suppression Failure Fvents . . . ... .. . . . . .. . . . 30-10.2.4 Events of Apparent Low Significance . . . 34

.11. Precursors to Potential Core Damage Accident Studies . . 35 12 . - ' Conclusion . . . . .. . . . . . . . . . . . . . . . . . 37' Rnittgangs , . . . . . . . . . . . . . . . . . . .. . . . . -39 o

t t f a Report on the Sources and Effects of Electrical Transients on the Electrical Systems of Commercial Nuclear Power Plants

1. Summary This report was written in response to if' .3-56, which requests that electrical transients.be added to the list of pheromena which licensed nuclear power plants and other nuclear facilities must be designed to safely withstand. In order to do this, the petition states that a comprehensive study must be made of all electrical transients. The petition states that this study must include a probabilistic risk assessment (PRA) of several representative plants, and that licensees should be required to perform the same level of analysis on each plant which applies for a new license or license renewal.

The fundamental assemption made in the petition is that electrical transients either can or should be studied in a PRA as a unified electrica] :ransient phenomenon. In order to determine if this approach is possible, the known sources and effects of electrical transients are briefly reviewed. It is demonstrated that electrical transients can be introduced to a system or-component by one of four methods; (1) by direct conduction of current over a power, control or data line; (2) by capacitive coupling to an electric field gradient; (3) by inductive coupling to a magnetic field gradient, and; (4) by irradiation of the system or component with ionizing radiation. Systems and components are typically shielded to the extent necessary to prevent misoperation or damage from electrical transient energy levels that have historically been encoritered. Problems are only encountered when the shielding is improperly installed or designed, or if all sources of electrical transients have not been properly taken into account.

Licensee event reports (LERs) are a source of information on operational experience involving shielding design or installation failure. LERs for lightning-related events which occurred between 1980 and 1991 are reviewed, to determine if such failures have occurred, and what impact they have had on plant safety.

Based upon these operatino events, it does not appear that the effects from electrical transients which have occurred could compromise the safe shutdown of licensed nuclear power plants.

It is concluded that there is no need for additional regulation of electrical transients as a class for existing licensed facilities, as requested in PRM-50-56.

2

)

1

.- . . _ . . _ _ _ _ _ _ _ _ _ _ _ _ _ v

< e f e

• e

2. Introduction PRM 50-56 requests that the Nuclear Regulatory Commission amend various regulations to account for the potentially harmful effects of all electrical transient phenomena. The petitioner states that "{t]o my knowledge, the design of no commercial nuclear power plant in the United States has ever been thoroughly studied to see what effect the range of electrical transients from lightning strikes, switching surges, or other sources, might be." The effects of lightning and switching surges have been thoroughly studied for over 50 years, and there are literally hundreds of technical papers on these two subjects in the technical literature. The purpose of these studies has been to protect electrical systems from the effects of lightning and switching surges. The results of these studies have been incorporated into the design of all power generation equipment and systems, not just those of nuclear power plants, as the owners of these plants have a powerful economic incentive to ensure catisfactory performance. Therefore, there is no basis for the contention that lightning and switching surges have not been thoroughly studied and that nuclear power plants were designed and built without any consideration given to protection from lightning and switching surges.

Although it is not stated until later in the petition, the term "other sources" is meant to include the wide variety of phenomena such as electromagnetic pulse (EMP), electromagnetic interference (EMI), geomagnetic currents, and " ferromagnetic effects". The fundamental assumption in the petition is that all electrical transients are the same, and that the effects of these transients on electrical systems have not been studied as a unified electrical transient phenomena. However, the petition does not establish why this approach is required. The approach taken by the industry has been to study the individual sources of transients and protect and shield electrical systems as needed for safety.

In response to the petition, the known sources and effects of electrical transients will be outlined. To begin with, it will be necessary to separate the sources of electrical transients from the effects. The predominant sources of electrical transients are lightning, line switching operations and faults, radiated electric and magnetic fields from various types of machinery, and geomagnetic phenomena. In addition, circuit conditions in the power transmission or distribution system can cause undesirable electrical transients. Finally, a high-altitude nuclear explosion can create an effect which is known as electromagnetic pulse (EMP) .

These sources of electrical transients will be individually described. in order to define how each source physically

. interacts with the power system and to define the effects of-each 3

~

r .

f e source of transient. This analysis will provide the technical basis for'a regulatory position on_the subject of' electrical transient' phenomena.

3. Electromagnetic Pulse (EMP)

A nuclear explosion releases a large magnitude burst of gamma radiation within a period of nanoseconds. If the blast occurs at a high altitude with low air density (e.g. 100-km), this radiation will form a roughly spherical shell with a thickness of several meters which expanda at the speed of light (Longmire).

Some of this radiation will eventually interact with the atmosphere,_and will be absorbed by molecules of oxygen, nitrogen, and other atmospheric gases. The excited atoms will-become ionized by emitting high energy electrons. The not reaction of all electrons produced-by this effect' explosion is called the Compton current.

The Con.pton current will travel in a net direction which is radially outward from the blast (Longmire). This will result in a compton. current vector-which is strongest and perpendicularly oriented towards the earth's surface directly under the blast.

This current vector will decrease in magnitude with distance from the point directly under the blast due to the decreasing-density of gamma radiation in the expanding shell as-it interacts with the atmosphere at a greater radius. In addition, the angle of the current vector will vary from perpendicular as a function of the blast radius vector and because of the earth's curvature.

As the Compton electrons travel through the air, they will cause secondary ionization of air molecules (Glasstone). These secondary electrons-ion pairs will travel a shorter distance and 1 recombine faster than the Compton electron-ion pairs. The separation and recombination of the Compton electrons from the positively charged ions and the secondary _ electron-lon pairs forms a large number of electric dipoles. A transverse field will radiate from each dipole, depending upon the separation. -

vector and time to recombination.

In addition, the Compton current vector direction also varies with time as a result of the influence of.the earth's magnetic field on the Compton electrons (Longmire). As the electrons travel through the earth's magnetic field, they will be deflected by the Lorentz force. The net effect is that the average Compton '

electron makes a looped path as.it travels away from then back' towards it's associated positive ion (which'is also affected by Lorentz forces, but is more massive and therefore harder to move). This effect results in an tangential component to the current vector, which creates a magnetic field in accordance1with-the Biot-Savart law. .This magnetic field is a function of.the deflection of the Compton current vector by the vector of the magnetic field produced by the earth (Lee) . This effect is 4

L:: e

t. n asserted to be coherent as a result of the gamma wavefront travelling at the speed of light (Lee, Longmire). 'The net Compton current also radiates an electric field due to this deflection which-is'similar to that created by a loop antennae.

This effect creates an electric field which is much_ greater than that from-a corresponding dipole antennae. .The first observations of these electric fields surprised researchers, as they were much greater than expected because-of the unforeseen geomagnetic deflection of the Compton current (Longmire).

These combined effects are what is typically referred.to as high-altitude EMP, or HEMP. In addition to HEMP, ionized particles =

are created by the nuclear explosion, such that the approximately sphere-shaped bubble formed by the blast has a net conductivity.

The effect of a' conducting sphere is to exclude a magnetic' field, such as the earth's_ magnetic field (Lee). This rapid change in the carth's magnetic field in the presence of the conducting surface of the earth will result in the generation of an-electric potential on the earth's surface in accordance with Faraday's law. This electric potential will create geomagnetically-induced currents (GIC), and is known'as magnetohydrodynamic (MHD)-EMP.

These currents can be of a significantly greater magnitude than typically experienced in relation to GIC from solar charged particles (Klein).

EMP radiation can induce a current pulse-on an electrical component from either electric induction, magnetic induction,_or from ionizing radiation (Glasstone). The electric _ induction effects are similar in nature to electromagnetic interference (EMI),-but are much more severe-for two reasons. The-first is that the field strength of an EMP electric field can be greater; than most EMI sources normally encountered. Second, the spectrum of radiation produced by EMP is broadband,-unlike-EMI sources

-such as radio transmitters.

. Transmission lines are also_ susceptible to_ electric and magnetic induction of currents. The rise time of_an EMP wave ont a power transmission-line can be very rapid. One source _gives a

" typical"' time ~of approximately 1.2 microseconds-(Glasstone).

-Lightning has a " typical" rise time of an order of magnitude!

greater than the EMP pulse, which creates a less severe effect on electrical insulation'(Lin).

~

This brief review-should clarify why it'is incorrect-to refer to pulsed electromagnetic-radiation as EMP, as'is done.in the petition._ .The actual mechanism.which is known as EMP is a complexLinteraction of radiation.from a nuclear explosion;with ,

the atmosphere and the earth's magnetic field, and which simultaneously affects' vast areas. Lightning, arc' welding, and- '

other conventional electrical disturbances create relatively minor electric: fields which diminish in magnitade with the square of-the distance from the source. d I  ;

5.

m

t e i e The petition suggests that computer programs which were developed to model the effects of EMP for military applications _could be used to study the effects of lightning and other electrical phenomena on nuclear power plant electrical syutems and components. However, EMP creates electric and magnetic fields that are uniform over areas of thousands of square miles. In addition, those programs include the effect of system-generated EMP (SGEMP), which involves ionizing radiation effects on electronic components. Because these effects are more severe than the effects of lightning and other typical electrical transient sources, the programs developed to model the effect of EMP on military equipment and installations might indicate a need for additional shielding which is not required for protecting equipment that will not be exposed to a near or direct nuclear strike.

The effect of high altitude EMP on nuclear plant safety systems has been studied by the NRC, and it was determined that the combined power line transient and EMI effects of EMP would not prevent the safe shutdown of a nuclear power plant (NUREG/CR-3069). Although low altitude explosions and ground bursts produce different EMP effects than high altitude explosions (Scharfman), the level of public exposure from a subsequent nuclear plant accident would be dwarfed by the level of exposure resulting from a nuclear weapon strike. Low altitude explosions and ground bursts do not appear to create a significant EMP threat to ground structures located outside of the radius of blast effects (Glasstone).

4. Geomagnetically Induced Current (GIC).

Geomagnetically induced current (GIC) can result from solar magnetic disturbances, which are created by the effect of charged particles of solar origin interacting with the earth's magnetic field. These particles cause severe compression of the magnetosphere. The earth's magnetic field is inductively coupled to the surface of the earth. The change in field intensity over time creates a potential in the earth's surface in accordance with raraday's law (Cleveland). This potential creates a very low frequency current field in the earth. This current has little in common with lightning. Lightning is a local phenomenon which creates electrical transients with the greatest magnitudes at high electrical frequencies (> 1kHz), and GIC is a global phenomena which produces electrical transients with low electrical frequencies (< 1 Khz).

GIC will flow in transmission lines, pipe lines, railroad tracks, or any other conductor which offers a path with a lower impedance than the surrounding earth. These currents are a potential problem due to transformer saturation and spurious relay operation and equipment damage (Kappenman). The adverse effects on a nuclear power plant would therefore be from a loss of 6

I

t i s a e

offsite-power due to system instability'and possibly from the ,

generation of high-frequency harmonic components of current produced by. transformer l saturation. However,_the probability _of

-a loss of offsite power is'much greater due to lightning than due to GIC, and power system instability does not pose any additional threat to'the safe shutdown of a nuclear power plant.

5. Ferromagnetic Effects The petition refers to " ferromagnetic effects" as a source of electrical transients, but it is not clear whether this means the-fact that nickel, cobalt and iron are magnetic, which is the correct meaning of ferromagnetic effects, or whether this is a reference to ferrorosonance, which is one of many phenomena which result from these ferromagnetic effects. Both iters will be addressed.

Ferromagnetism is the phenomena which occurs in some metals, such as nickel, cobalt and iron, and which results from an aligned electron spin moment (Skitek). In most-other elements, the electron spin moments oppose each other, and result in a,very low not electron spin moment. The ferromagnetic effects _are caused by magnetic domains, which.are created by the-strong interaction forces which occur between ferromagnetic atoms. In the: presence of a magnetic field vector, H, the domains will align and intensify the magnetic _ flux vector, B. If the magnetic field is time-varying, the magnetic flux will follow a hysteresis curve in-relation to the magnetic. field, which is caused by the magnetic domains re-aligning'as the magnetic field vector changes orientation. These ferromagnetic effects in and of themselvos_do- _

'not pose any hazard to the safe operation-of the plant. .

Ferroresonance refers to the overvoltage condition that can occur due to.the interaction between transformer inductances and-various capacitances as connected by-the transmission. system-(Miller). Because the inductive current of an a.c.1 voltage is 180 degrees out of phase with'the capacitive. current, it is-possible-to generate voltages of a large magnitude-across a series or parallel connected inductor and capacitor while drawing-very low current. The transformer is primarily inductive, but also contains some internal capacitances.. -Transmission line:

capacitive effects will predominate when the line-is= unloaded, and a. transmission line may also have capacitance' installed in series or shunt to_ improve voltage control.

If the circuit which is-formed by_ leaving an unloaded transformer connected to the transmission line is excited by a voltage at or near the resonant frequency of the circuit, the voltage will increase until protective' relaying orfsurge arresters operate.-

This rise can damage the transformer or capacitors,'or cause-repeated activation and subsequent failure of circuit breakers or lightning arresters. Ferroresonance usually only affects 7

4

r i. ,

f 4' 1

unloaded or: lightly loaded transformers, and should not posera Jhazard to aftransformer supplying power to or from a nuclear power plant (Smith).. ,

An additier.a1 electrical transient which was not mentioned in the petition or public comments is subsynchronous resonance (SSR).

SSR-is caused when the transmission system excites a: natural' frequency of the turbine generator shaft-system,.which predominantly happens on'large machines with several torsional-resonance modes in the subsynchronous frequency range (<60 hz)?. 1 It is-typically caused-by extensive use of capacitors to control; system voltage profiles'and the power transfer capacity of transmission lines-(Miller). SSR can'cause catastrophic failure:

of the turbine generator shaft, which can result in loss of life-and extended plant outages. -This effect was first observed in the early 1970's, and was addressed'and resolved by the utilities and manufacturers without intervention-from Federal or State regulatory agencies.

6. Miscellaneous Bources As previously discussed, FMP is the term given to the broad '

frequency ban'd magnetic and_ electric fields produced by a nuclear-explosion. Therefore, the term refers more to_the source of:the transient than to the effect. On the other. hand, EMI-is a-general term that refers to impairment of an electromagnetic signal by an electromagnetic disturbance (IEEE). This definition encompasses any; source of unwanted electrical signals, such:as-arc welders, commutator. brushes on motors and generators,-hand-l held radio transmitters, and circuit breakers. The definition also implicitly encompasses-any physical mechanism.by which an unwanted-electromagnetic disturbance can impair an electromagnetic signal, such as electric field effects,-magnetic field effects and_ power line. transients.

Essentially, an'NRC, position on EMI would be-the closest technically correct answer-to what was actually requested-in PRM-50-56.

The NRC is currently conducting research'at Oak Ridge National Laboratory (ORNL) under FIN L1951, " Regulatory Guide'and Acceptance Criteria for Electromagnetic. Interference in Digital' Systems." The-objective of the program is to develop the technical basis for a regulatory guide which will endorse Lsome car all of IEEE Standard:1050, "IEEE Guide for I&C Equipment l-Grounding.in Generating' Stations.'" LHowever, the' design guidelines in IEEE 1050 are' intended to minimize degradation of

instrumentation-and control signals-in generating. stations, as. '

opposed to providing protection from static charges, lightning, '

olectrical faults, or_other' gross transients. Implementation 1of IEEE 1050 would help to mitigate the effects:of'a lightning strike or similar disturbance on generating station-electrical systems, but additional measures would be needed in orderito protect generating = station structures, switchyard components,'the 8

i .

l f e

. i transmission system, and other systems and components.

7. Switching Surges A high-voltage transmission line will encounter faults for a variety of reasons. The predominant reason for a line fault la a lightning strike (Blackburn), which raises the potential of either the line and causes "flashover," or of the groundwire or tower and causes "backflash" (Golde). Other reasons include shorting between phases caused by birds, insulation breakdown caused by contamination, and random failure of surge arrestors.

When a fault occurs on a transmission line, it must be " tripped,"

which refers to the opening of circuit breakers to isolate the line. A high voltage transmission line will also be tripped infrequently even if there is no problem with the line, such as for maintenance or to take a generator offline. Tripping a transmission line creates a surge wave that is called a switching surge (Greenwood).

Some aspects of switching surges on power lines are less severe than lightning surges. The rate of increase of the voltage is lower, which produces less stress on the insulation, and the voltage and current will reach a lower peak magnitude than the surge from a direct' lightning strike. Other aspects of switching surges are more severe than lightning surges, as they can contain mere energy and create greater thermal stresses on surge arresters. There are additional problems associated with electric and magnetic field effects on equipment located in the switchyard due to the operation of circuit breakers (Greenwood),

although these problems should not affect safety-related equipment in the plant, by virtue of it's location.

Switching operations can create other problems on the power system which are not directly caused by lightning (but may be indirectly caused if lightning results in a line fault).

Switching operations can initiate ferroresonance. In addition, when a unit is tripped off line while the generator is providing power, the turbine and generator will undergo shaft torsional oscillations. Repeated occurrence of these oscillations can cause mechanical failure of the turbine generator shaft, especially if the unit circuit breaker recloses on the fault and excites one of the natural torsional modes of the turbine generator shaft (Joyce).

Switching surgos are also generated inside the plant by the switching of large motors. These surges can reach values as high as 4.6 per unit (i.e., 4.6 times the rated system voltage)

(Nishikawa). This reference reports on a study of inductive coupling between parallel runs of power, instrumentation and control circuits in a Japanese nuclear power plant. The surge currents created by these switching operations on the power cables did not generate significant voltages on the adjacent 9

l

mi .

' t. *~ t

c. ,

control and instrumentation circuits. This paper provides some

indication-of the amount ~of shielding-provided int internal cable shielding and raceway. .In' addition, the propagation of these-surges through the distribution system from the high_ voltage power systems-to the low voltage instrumentation and control' systems was not a concern, which indicates that such switching surges are attenuated during' transmission from higher voltage-levels to lower voltage levels in the plant.

In summary,-power line transients are the only'possible effect from switching _ surges which pose a threat to_the control systems of nuclear power plants, which is the concern stated in the petition, but_there are many other effects which can create a significant hazard to other. plant equipment,-including control systems in the switchyard. The statement in-PRM-50-56 thati "these transients are not considnred crucial for conventional-generating stations" is therefore incorrect. The potential impact of these transients is loss _of human life and extended-plant outages due to destructive failure of turbine generators, circuit breakers and transformers. In addition, system-instability could result in a regional blackout, such as'_the event which resulted in the New York City blackout on Julyfl3, 1977. The use of digital-controls in fossil futs generating stations has already become widespread. If switching surges ( or-any other source of electrical transients) created any problems with these devices,'these problems would become evident in operating experience and would probably have been addressed in the technical literature. The absence of any reported problem is likewise significant.

S. Lightning Strike Phenomena Lightning strikes create _ effects which can be divided into'two categories, power line transients and local strike effects. -The Advisory Committee on Reactor Safeguards .(ACRS) reviewed'the_ .

draft regulatory; guide referenced lin_PRM-50-56-on February 3,

'1981. From the transcript of this meeting, it is evident that the proposed regulatory guide did_notLaddress the potential effects of local strikes, and attempted!to compensate for this by imposing a.much stricter requirementffor sizing of surge arresters than is currently used.- As discussed-below, this-approach would not be effective. In' order to address the history-of this issue, it will be necessary_to describe,the possible effects of a lightning strike on a nuclear power _ plant.-

8.1 Power Line Transients Lightning can -interact;with' transmission lines in several ways.-

If the lightning strikes ground near the conductors,'but does not; actually strike a conductor, the lightning can still create a-surge wave on nearby transmission lines by magnetic induction ~

(Eriksson). This surge will be of relatively low magnitude 10 .

4

i- =..

(Yokoyama). If the lightning. strikes the shield wire:or grounded

= tower, the protected line can be involved in a back-flashoverL event,: where the potential of the tower as the: lightning-charge- >

flows to ground becomes. great enough to cause anLinsulation breakdown ofLthe line' insulators (Golde). The protected-line is-then shorted to ground, but at a time when.the-potential of ground.is momentarily greater than that of the line. As the potential rapidly decreases back towards_that of distant ground, the short will-remain due to the existence of an ionized conducting path (i.e. the arc).

An induced surge or a surge created by a back-flashover will be '

less severe than the surge caused by a direct strike. The effectiveness of shield wire protection of transmission-lines has-been thoroughly studied, and indicates that even the-best shield wire designs in use do not protect a line from overy direct strike (Brown II), although they do act to-limit the magnitude of-current in strikes which' reach the transmission line. For this reason and for other historical reasons,-transmission systems and generating stations have been built to withstand the. effects of a direct line strike. However, shield wires are used to minimize both the number of direct strikes and the magnitude of current in such strikesL(Brown I).

If lightning strikes a transmission line, a charge is essentially-

. injected into the line, and will create a' travelling wave.in both-directions on the line (Golde). Electrical current is a-physical phenomenon created by tlue movement of. electrical' charge. The discussion of the proposed-regulatory. guide in the ACRS transcripts _ appears to miss this fact, and focuses on the measured magnitude of lightning current-and not on transmission

~

line surge' currents. The magnitude _of the-transmission 1ine surge currents will be less than'the magnitude of-the lightning current which creates the power line transient, ,

current is measured in amperes, where cnua ampere represents the transfer.of one coulomb of charge in one second._of_ time;.

Consider an ideal case where a lightning stroke uniformly transfers;30 coulombs of charge in one microsecond.1 The current will have a magnitude of 30 million amperes. If this charge is deposited on a transmission line, it will split into two halves that will flow in either direction (Greenwood). If.the charge l uniformly transfers.to the transmission line, this transient will.

.have a current magnitude'of-15;million-amperes. However, ifLthe charge waveform is changed by the splitting process 1such that the-15 coulomb charge is now uniformly-transferredLinrthree microseconds, the magnitude of the current surge will have dropped to 5 million amperes.

The waveform of charge'as a' function of time (i.e. current) will change as t..e lightning charge transfers to the transmission.

line, both because of charge: splitting, and because of the 11

-s-  :.

?  %

. process of charge-transfer from the lightning channel to the transmission line. -In addition, the surge waveform:will be attenuated by resistive losses and corona as the surge moves down-the transmission line (Clayton). The surge 1will also lose: energy due to capacitive line charging and through_ inductive coupling-to-adjacent transmission lines.(Greenwood). When;the surge reaches a point where-it can flow to ground, such as a surge arrester, some of it wil3 flow past the arrester before it operates, some of it will flow to ground, and some of it will reflect off the arrester after it activates and travel in the opposite direction on the transmission line (Greenwood).

A comprehensive study of unshielded distribution line surge - ,

arresters found that only 2 out of 2800 arresters-(0.07 percent) with an average service life of 12 years exhibited-signs of.

having discharged a 100,000 ampere surge.1 over 99 percent of the discharges were 40,000 amperes or less (Gaibrois)'. Distribution lines are typically unshielded because of the'relatively large' cost, but high voltage transmission lines are almost universally built with shield wires. Line shielding will limit the magnitude of lightning current from a direct strike (Brown I). The maximum. '

stroke current which is expected for typical shield wire designs would be less than 90,000 amperes. This corresponds to a surge of less than 45,000 amperes after accounting.for charge splitting-and line attenuation.

The fundamental difference between lighting problems on distribution systems and on transmission systems must be kept in mind when reviewing technical references in the literature. Much of the work which has been performed with.regards to measuring: '

induced voltages, surge arrester currents, and transients at the service voltage level-was performed from the distribution end of the power generation system,-not at the generation end (Martzloff I and II, Yokoyama, Eriksson, Gaibrois, Odenberg). As noted, distribution lines do not' typically have_ shield wires. The transients measured on distribution and customer lines will'be ,

more severe than the transients experienced in power plants due to power line transients from lightning.

Even with this constraint, the transients measured on customer lines are not excessive. odenberg reports voltage peaks of approximately 4.0 per unit at typical service levels,.with a-strong dependence on the location of the supply in: the distribution scheme (main > subpanel > receptacle). This.

corresponds to a voltage " spike" of-approximately 500 volts for 120 volt service. A-20,000 ampere surge on a 30 kV distribution line having a typical surga impedance of 200 ohms would-have.a corresponding voltage peak _af 400 kV, or approximately'13.0 per unit. Thus, a surge is greatly attenuated in the process of propagating down in voltage level. The same surge ona 345.kV transmission line would be approximately 1.16 per unit. -This surge would create a negligible effect on a power plant control 12 3

s .

l system.

For these reasons, there is no technical basis for the position in the draft regulatory guide that surge arrestors should be designed for a discharge-current surge of 200,000 amperes, or even the modified position of 120,000 amperes. Even though there is slight probability of getting a lightning stroke with a current of this magnitude, the stroke current magnitude does not equal the magnitude of the current surge on a transmission line or through a surge arrester. The magnitude of current which can strike a shielded line is limited, and the surge waveform will be attenuated as the surge travels down the line. Finally, the entire surge will not flow to ground at the lightning arrester.

There is no support in the technical literature for the position adopted in the draft regulatory guide.

The surge current is a function of the_ charge deposited on the line by the lightning strike, but the surge voltage is a function of the surge current and the surge impedance of the line. The surge voltage waveform will have a leading edge waveform which is proportional to the surge current by a factor of the surge impedance (Greenwood). The surge impedance of a transmission line dependo upon physical design, but it is typically in the range of several hundred to several thousand ohms. Thus, the surge voltage for a 200,000 ampero surge would be 40 million volts or more. The dielectric strength of air is on the order of 500 kV/m, which means that a 40 million volt surge would flashover to any grounded structure within 80 meters of the line, a distance which exceeds the height of most transmission line structures. Even if this surge current amplitude is considered for the sake of argument, it is clear it would result in a flashover to the shield wire or other grounded structure.

If the surge does reach the unit transformer, the surge voltage on the high voltage side could be capacitively coupled to the low voltage side. However, it is possible to minimize this voltage by transformer design and by the use of surge arresters and capacitors on the transformer low voltage terminals. In addition, the voltage level at the generator terminals is on the order of 22 kV in nuclear power plants, and will be transformed down several times before ultimate use. These additional transformer stages and long cabling runs provide a significant amount of impedance to high frequency _ surges (Elgerd, Jackson).

This analysis indicates that high frequency power line transients will be damped-out prior to reaching digital control equipment in the power plant. In order to verify this assertion, operating events are reviewed in section 10 to determine if any power line.

surges have resulted in equipment damage or misoperation.

However, this review is essentially unnecessary, because the digital control equipment which is proposed for use in nuclear power plants has been increasingly relied upon in fossil power 13 a

i =.

e 4 plants and in' industrial applications, with no reports of widespread problems csused by electrical transients. There are no significant differences between fossil and nuclear power plants with regards to lightning surge protection, and any sensitivity of-digital control equipment to power line transients would have been exhibited at fossil power. plants.

8.2 Local Strike Effects on Power Plants Local strike' effects from lightning may be due to-the interaction-of magnetic and electric fields associated with the. lightning current, in addition to any potential for the lightning current to be carried through structures or components. _The current induced in a-nearby conductor _by.the magnetic field of an indirect lightning strike is a function of the speed-and magnitude of the charge transfer in the lightning, and the mutual magnetic flux coupling between the lightning channel and the conductor. For a transmission line which has a length measured in miles and a nearly straight path, this induced current'can reach a significant magnitude (although significantly less than a direct strike). For a cable.in a power plant with a length measured in hundreds of feet and a three-dimensional cable run, the current will be much less. In addition, these cables will be surrounded by steel reinforced buildings and metallic raceway enclosures, which will reduce the mutual magnetic flux coupling with a lightning channel to a negligible quantity.

The electric field effects of lightning are created by-capacitive coupling and are similar to electromagnetic interference (EMI).

Capacitive coupling refers to the potential which develops on an ungrounded metallic object which is capacitively coupled to both the lightning channel and the ground. The ratio of these capacitances will determine what fraction of the stroke 1 potential relative to local ground will be imposed on the_ object. The-magnitude of the capacitance will determine the amounttof charge which can transfer to the object. Most_ metallic structures in a power plant are grounded to prevent the buildup of static charges, but this will not prevent capacitive coupling to high;

~

frequency electric fields.

The frequenc; spectrum of. lightning-generated radiation extends' from the extra-low frequency band (ELF) to above the ultra-high frequency' (UHF)~ band, and can be determined from the fourier transform of the lightning pulse current waveform. A study which compared =the calculated local field from_ lightning at 50 meters to the uniform field intensity of high altitude nuclear EMP indicated that the field magnitudes are quite similar (Uman). It was determined in a separate study on the effects of EMP that _

structural steel-and metal raceway will.be effective at shielding safety-related circuits from these effects '(NUREG/CR-3069). In addition, the EMI effects from lightning will decrease by the square of the distance, whereas EMI effects from'EMP are uniform 14

+ .

over:an area such.as.a plant site. For these reasons, it--is unlikely that safety-related circuits could be damaged by_EMI.

effects from lightning.

The third local effect of a lightning striko-is caused by-the lightning current. Structural = damage can_ occur if lightning-strikes a material which is unsuited to' carrying electrical current, such as concrete. In addition, lightning can create a-  :

phenomenon known as around potential rise (GPR),_which is: caused  !

by the_ increase in ground potential in the vicinity of a lightning strike with respect to remote earth. This'can be particularly dangerous if the plant is grounded separately from  ;

the switchyard, and has control or instrumentation cables connected between the two grounds. GPR can induce high-voltage transient surge currents in control circuits where the entire plant is on one common ground (Mitani). -These voltages are reported to be due to capacitive and inductive coupling between cables and plant structural components (Ikeda). Increasing the.

efficiency of lightning protection on plant structures would not prevent GPR eventa from occurring, and might even increase the frequency of GPR events (Golde).

Of all the possible transient effects caused by lightning (i.e.

power line_ transients, magnetic induction, capacitive coupling, EMI effects and GPR), GPR appears to present the-most serious.

hazard to plant equipment, based upon operating events. All a equipment should be grounded to the station. grid, and thus all equipment could potentially be affected by the effects of GPR.

Operating events are reviewed in section 10 in an effort to determine the effects of local strikes on operating-plants'.

9. Summary of Electrical. Transient Phenomena Based upon this discussion, the transient effects of the various electrical transient sources can be tabulated.

Electric MagneticfPower Ground Pie 13 Field Line Potential Effects Effects Transients Rise EMP X X X GIC X Ferromagnetic effects X Miscellaneous sources X-Switching Events. X_ X X.

Lichtnina X X X X A quick inspection reveals that this table does not. completely _

represent the: effects of all transient sources. For example, both GIC and lightning are-attributed with creating power _line transients, yet the discussion in_ sections 4 and 8 clearly.

indicates that the means of protecting power. lines from these.two

- phenomena would be distinctly different.- However, it-is apparent from the table'that only lighting is capable of causing GPR.

15 E4 - _ ,

y -

t

.<' .c

s. .-

o .

Protection _offdigital. equipment from the' effects of GPR appears.

to be the most important aspect of. lighting' protection, yet .

implementation of.the. original _ draft of'the regulatoryLguide e

would not have.h?lped to protect-nuclear plants from this-phenomenont a fact which war. stated in the IEEE Power Engineering Society comments. It is noted that neither the' petitioner nor commentors addressed-GPR.

10. Lightning Events

- The draft regulatory guide for lightning protection was not issued previously, because:

1. the risk analysis indicated that events resulting in core damage due to lightning were not a significant contributor to the total risk from a nuclear power plant .

2. the technical issues which were raised in-house and by the industry were not resolved in the revised draft, and e apparently, a decision was made not to devote any_more resources to this issue.

LERs were reviewed for the purpose of addressing the concerns raised in the petition which could not be addressed by a review of information in the technical literature. This review first involved a determination of whether the event involved a local '

strike or a line strike, and it is possible.that an event was caused by both a local and line strike. 'If the event involved a line strike, it was necessary to determine whether_a power line transient caused any control equipment = damage or misoperation.

Control equipment misoperation is important, .because the' level of energy transmission which can damage digital equipment'might result in misoperation of electromechanical controls.

If the event-involved a-local strike, it was necessary to determine if the local strike caused _a concurrent power line  ;

transient, such as a local strike to the' switchyard. -It was'also necessary to determine if the local strikeEcaused damage.or- .

misoperation due to inductive coupling, capacitive couplingLor1 GPR. In most. cases, the amount of-information provided'by:the LER was. insufficient to determine the exact cause,.but in ,;

-general, the data-presented in these LERs indicates that GPR appears to be the most probable: mechanism wP 1-couples a-lightning strike to a power plant electricai .ystem.

10.1 Power Line Transient Ev0nts.

Lightning is a relatively common natural-phenomenon. _Unlike torn los,-earthquakes,-hurricanes, tsunami,.or other large scale and infrequent natural disasters,. lightning. frequently affects .

power systems and apparatus, and' utilities' routinely take

. measures'to-protect against lightning damage. However, this protection is_an economic trade-off between'the cost ofl damage:

and the cost of protection, Shield wires..for transmission _ lines 16 -3

-. e x

• provide a: good example.: -The best shield wire design will not '

prevent an- occasional direct strike to a protected line. But the presence of even a poorly designed shield _ wire will greatly reduce both the number of direct strikes to-a line and the current magnitude of those strikes (Brown I). . For this reason, a ,

utility might-choose to accept a less_ effective' shield wire design if it costs less than the best-available design, because the utility will have to protect the line from the effects of a direct strike in either case.

Inherent in this economic analysis is the recognition that it may; be-more cost effective to replace a damaged component ifiand when damage occurs than to protect the entire system from the worst possible condition. This economic reasoning does not extend to the generating plant for-several reasons. Mechanical failure of the turbine generator is extremely hazardous and-has caused loss of life. Electrical failure of the station transformer and.

generator is also hazardous, and failure of any major equipment' usually results in extended outages, loss of. revenue, and expensive repairs. For these reasons, generating' plants are heavily protected by design from the effects of lightning.

When lightning-related events are reported-in licensee event reports (LERs), the LER will_usually state that an event occurred while an electrical storm was in progress. Thel correlation between lightning and the event is inferred and-not direct. If the event.is caused by a transmission line trip during-an electrical storm, it can be assumed that lightning was involved.'-

llowever, when the event involves for example the blowing of a 1/16 ampere fuse in a single piece of equipment with no-other effects, the assumption of causality is less certain. A-fuse might blow for a variety of reasons which are undetectable, such as age or a defect in manufacture. In addition, the fuse might-blow from increased current flow resulting'from a drop in-voltage, if it is L fuse between a regulated independent l source and a line-fed source. In this case, the-lightning mayLbe responsible'for the blown fuse, but it was not physically coupled to the fuse, which is an important distinction..

In order: to determine if a power line transient was the' cause of an event, some basic information must be known.

~

Did lightning strike near the plant, cnr did it ctrike the _line? Did it strike; immediately before the failure? Did the_line voltage-decrease at the' plant?- Did a power line transient propagate to the control system power supply voltage? In most LERs,_a detailed analysis is not attempted. -In some LERs, the technical explanation for the causal link between the lightning and.the failure;is either

-totally absent or is superficial. It is only after several similar events that some utilities will. expend the engineering resources to attempt to' resolve the problem. Further, it is apparent in several cases that the random nature of lightning

^ makes it difficult to' determine if the problem has been resolved.

17

. c 1

.- .. u I

e . i

') .

'The petition =for rulemaking requests that a representative sample of plants be studied by_probabilistic risk assessment--(PRA)-in .  ;

order to determine the effect of all electrical transients. This I request assumes that the frequency of_ occurrence and magnitude of these transients could be predicted and meaningfully included in  ;

a PRA. However,. transients have been studied for over-50 years in order to prevent unwanted effects from occurring. Licensees ,

currently take whatever measures appear to be effective to- a prevent misoperation and damage of equipment from electrical-transients. -When problems occur, they are-unpredictable and unexpected. In cases where problems are recurring, licensees take steps to avoid repetition of those problems in the future.-

Therefore, if these transients could be-estimated, it would make more sense to protect against them, rather than to analyze them to unnecessary detail.

It is possible to estimate the potential impact of transients based upon the gross effects of known transient sources. For example, the loss of offsite power (LOOP) can occur due to a lightning strike. It is therefore possible to estimate the impact on plant safety due to lightning strikes causing the loss of offsite power by determining what percentage of LOOP events~

are caused by lightning. This approach can be applied to other J l transient sources (e.g. LOOP events caused by GIC), or to other

- events (e.g. loss of main feedwater caused by lightning). Events caused by lightning were the only events reviewed in response to the petition, due to the existence of other NRC programs which study the other significant sources of transients.

Oak Ridge National' Laboratory (ORNL) performed a search'of the Sequence Coding and Search System (SCSS) data base, which' identified 142 LERs for the eleven-year period from 1980 to 1991 which reported a lightning related event. In addition, 10 LERs- a were discovered through other sources. These events were grouped into categories of power line transient and local strike event.

These two groups were analyzed to-determine.what plant effects appear to be caused by lightning, and whether these_ effects _could be used to estimate risks which may exist due to ineffective. .

, lightning protection. This data is presented in Tables l'through 3,-and is described in greater detail in the following sections.

10.1.1 Line strike, no effect These LERs reported events which did not involve any complications or equipment failure. Most involved the inadvertent start of a diesel generator or-loss of a required ,

offsite line. These events involved line strikes which were-located beyond the plantfswitchyard. These events correlate well-with the theory _of lightning _related surges in the technical

-literature, namely, that line and transformer attenuation decrease the magnitude of-power line surges to a point where~they.

do not pose a hazard to plant equipment.

18

i .

10.1.2 Line strike, plant effect reported The abstract for these LERs suggested that equipment misoperated or was damaged in a way that indicated that-a power line' surge transient might be responsible. The LERs were then reviewed to determine if a surge voltage could have propagated into the plant and caused the equipment misoperation or damage. Of these LERs, 22 involved strikes to high voltage transmission lines, 3 involved strikes to the station switchyard, and 5 involved damage to transformer bushings or surge arresters. These are estimates, because most LERs do not provide enough information to positively establish the location of the strike. However, it would be expected that the 22 events involving strikes to the transmission line would have less of an effect on the plant due to power line surges than the 3 events involving strikes to the switchyard or -

the 5 events involving damage to transformers.

The 22 events involving off-site, high voltage transmission line-strikes contain good examples of poor reporting of lightning ,

related events. Two of these events (LERs 382-91-013 and 322 008) referred to damaged circuit cards and one (LER 382-91-013) involved a spiking chlorine monitor. No attempt was made to provide an adequate technical explanation of the cause of the damage, which could have been due to a drop in line voltage, GpR from a local strike, or a power line surge overvoltage. The worst event report (LER 298-80-031) states that a line fault in the 345 kV system " induced" an erroneous signal in the turbine generator digital electric hydraulic (DEH) computer. This spurious signal indicated that the unit circuit breaker had tripped. However, if a line fault occurred, it is possible that protective relaying transmitted a trip signal which was blocked at the generator circuit breaker by.a time delay. It is highly unlikely that a power line surge could induce a spurious signal inside the logic circuits.of a computer without causing any other -

damage or spurious signals.

The remaining 18 LERs provide enough information to determine the cause of equipment misoperation, which seems to be primarily due to low voltage. These events are:

1. Failure of component prior to lightning strike (254-90-004),

(388-85-025).

2. Low voltage, etc. (254-90-013), (346-87-020), (370-85-5),

(387-88-010), (388-85-020), (397-90-024), (454-88-006),

(454-89-007), (454-91-002), (458-85-063), (528-87-021).

3. Lightning strike stated, cause of trips was not explained (254-91-008).

4. Equipment misoperated during or after event, then cleared up without explanation (259-80-59), (395-86-012).

19

1 l . .

5. Surge or " spike" stated as cause of event, when actual cause was probably low voltage (395-83-074), (416-91-006).

The 3 events which involved a lightning strike to the plant switchyard should have a greater probability of causing equipment damage due to a power line surge than a strike beyond the switchyard, because the surge has a shorter distance to travel and will be attenuated less from the line losses. There will also be GPR cffects from that part of the surge which is conducted to ground by the transformer surge protection.

However, it must also be kept in mind that the magnitude of current or charge in a lightning stroke is a stochastic event, and that a sample size of three events does not provide an adequate statistical tasis to draw a firm conclusion.

These 3 events do not provide any indications of plant equipment or misoperation caused by line transients. Two events (LERs 348-91-009 and 456-91-005) did not report any concurrent equipment damage or misoperation. The third event (LER 278-85-018) involved the existing or simultaneous failure of a DC solenoid valve in the MSIV controls, a failure which was not attributed to lightning. In additica. the LER also reported that a circuit breaker inadvertently opened due to GPR offects on a manual control cable running between the switchyard and the control room. Manually trioping the circuit breaker was noted to cause a similar effect, which would be expected.

The 5 events which involved strikes to or flashover of the unit transformer bushings or damage to the surge arresters should have the greatest probability of causing equipment damage due to a power line surge wave, because of the conbined effects of inductive and capacitive coupling of the-transformer low-voltage winding to the lightning surge. However, two of these events (LERs 529-89-001 and 302-81-033l did not report any additional -

equipment failure or misoperation. The third event (272-91-024) states that a manual / auto controller for an atmospheric release valve failed, but does not indicate what the cause of the failure was. The fourth event (LER 301-87-002) states that some meteorological equipment was interrupted, but does not state whether any damage occurred.

The fifth event report (LER 317-87-015) states that a transformer bushing flashed over during a snowstorm, and that lightning had been seen in the area. The only equipment failure was the plant computer, which became non-operational at approximately the same time as the bushing flashover. This observation is apparently the basis for asserting that lightning was the cause, and it is not stated elsewhere that lightning was seen or heard. This unit is near the coast, and it is therefore possible that the transformer bushings flashed over due to atmospheric contamination and/or a local strike causing a rise in ground potential. The computer failure can be explained as resulting 20

Q 'e from low voltage / power interruption or GPR. No technical explanation of the computer failure was provided in the LER, and-no damage to computer components was described.

In summary, these events provide no compelling evidence that a transmission line strike, switchyard strike, or flashover at the-unit transformer bushings will allow a surge'overvoltage to. enter the plant via the power line. Where equipment misoperation or-failure occurred, the cause was not technically analyzed, and can be explained from other known effects of a lightning strike. For example, low voltage can cause fuses and circuit breakers to actuate due to overcurrent conditions, and low voltage can cause equipment to trip and reset to default values. _GPR can create _,

~

surge overvoltages due to grounding configurations and may create other effects which are currently being studied. These events support the expectation that transformer inductance and surge arrester protection will mitigate the effect of power line surgos.

10.1.3 Similar Lightning and Loss of Line Evnnts These 6 events (identified as group A) occurred at Susquehanna Unit 1, and resulted in plant effects which were similar to 4 ovents (identified as group B) which involved a loss of line with-no concurrent lightning strike. These plant effects are.also similar to many effects reported in LERs where events were alleged to have been caused by a " voltage spike" or " power line surge."

The group A events (LFRs 387-84-028, 387-84-029, 2 events on-387-86-028, 387-87-020, and 387-88-014) all involved Zone III HVAC isolation, and train B stand-by gas treatment system (SBGT) actuation. In addition, these LERs report that-various radiation monitors tripped and chillers isolated,_but do not provide any further identification. Recirculation pump runback is reported on four of the six LERs, A and B train reactor recirculation-scoop tube lockup was reported for three events, and river water.

clean-up (RWCU) system isolation is reported for_two events.

The group B cvents (LERn 387-83-092, 387-85-035, 387-87-007, and 387-87-015) were caused by transformer failure, offsite surge arrester failures, and an unidentified'fuult not attributed to lightning. Two of these four events involved Zone III HVAC isolation, and three of the.four involved train B SBGT actuation. '

Two of these events state-exactly which radiation monitors tripped and which chillers isolated, while a third doesn't identify the exact units. Recirculation pump runback is reported on two of-the four LERs, A and B train reactor recirculation-scoop' tube lockup was-reported for two events, and-RWCU isolation.

is reported for two events. .

The similarity of these events suggests that low voltage / loss 1of 21 4

• e

power is responsible for many of those equipment trips.

Destructive failure of surge _ arresters, transformer failures, and line faults would not be able to create a uteep-fronted, high-current surge effect like that created by lightning.

Unfortunately,;the licensee _does not identify what caused the equipment-to' trip or isolate, i.e. Iow voltage,-load sequencing, etc. However,-it seems relatively certain that the mechanism which is causing the equipment to trip or otherwise1 fail from offsite line_ faults to' ground not caused by lightning-is the same mechanism which causes equipment to trip or faj' from an offsite line fault to ground caused by lightning.

10.1.4 Events with fire or loss of fire protection These events involved 2 line strikes which resulted in a fire due to surge arrester failure, one which resulted in a fire due-to an exploding circuit breaker, and one which_ involved loss of1 fire protection equipment. They are discussed in Section 10.2.3 of this report in conjunction with 8 other' events caused by.a local.

strike which resulted in a loss of fire protection or a fire.

10.2 Local Strike Effects The LERs included 66 events involving local lightning strikes.

This number of_ events is estimated, due to the unclear and inadequate explanation in some of these LERs. As previously described, it is possible that these events could be caused by magnetic induction, capacitive coupling, EMI or GpR, and-these mechanisms will be considered in each event analysis.

10.2.1 CRD Power Supply Overvoltage Protection Trips Reactor trip caused by spurious rod drops into the core 10 an-event with a large number of occurrences (i.e., for lighting-related events, not relative to total number of unit trips).

These rod drops were caused by-the operation of the overvoltage (OV)' protection on both the primary and back-up control. rod. power drive (CRD) power supplies for 20' events. These events occurred at-the following times and plants.

Year of Occurrence, Number of Incidents

'79 '80 '84 '85 '86 '87 '88 '89 #90 '91 Braidwood 1 1 1 1 Braidwood 2 1 2

-Byron 1 1 2 1 Comanche Peak 1 1 Farley 2 1 1 1 Vogtle 1 1 Zion 1 1

' Zion 2 1 2 1 The occurrence of this event has been relatively widespread and 22

=, .

..}

. frequent, and provides a good example of how the industry tends to address 1 lightning-related events. The utility explanations from_each LER are summarized below, and the explanation'is analyzed in regards to the six critoria listed. The description given of each criterion. reflects the usual content of the LER with regards'to that criterion. For eaxample, most LERs did not state whether a lightning strike was observed immediately-prior to the event. Any additional information, such as if and where lightning was seen, will be given after each event description, and will be identified by category.

1. Causal - it is not stated whether a lightning strike.was observed immediately prior to the trip.

2. Power Line Transient - the LER does not state that a power line transient or voltage dip was recorded.

3. EMI - the LER does not state that EMI was considered as a possible cause or contributing factor for this event.

4. Magnetic coupling .the LER does not state that magnetic induction was considered as a possible cause or contributing factor for this event.

5. Capacitive coupling - the LER does not state whether the equipment housing was grounded.

6. GPR - the LER does not state that GPR was considered as a possible cause or. contributing factor for this event.

1,2 - Zion Unit 1 and 2,_-August-17, 1979 (PNO-79-347)

This event was reported on a preliminary notification of_ event =or unusual occurrence (PNO). IE Information Notice 85-86 provides-more detail of the event.- A lightning strike in close proximity to the plant caused the overload protection devices'for the control rod drive (CRD) do power supply-cabinets to trip. Tests verified that noise induced on the ac input to.one power supply would actuate the overvoltage protection trips on-the main and auxiliary power supplies. One power _ supply was damaged by the _

incident. Several actions were taken, including using the CRD motor-generator as the-power supply for the auxiliary pover:

supply.

2. Power Line Transient -;the analysis states that noise on the ac power supply replicated the event, but it is not stated whether a power line transient occurred. It.is.not' stated whether the overload protection device'which tripped was the overvoltage protection device, the setting of which was increased from 27 V to 29 V.

3,4 - Zion _ Unit 2, April 3 and July 16, 1980..

Information on this event was obtained-from IE Information Notice 85-86. The notice states that the cause of.the-trip on August-p 17, 1979 was the same as-the cause of the trips on April 3 and July.16, 1980, and _that these trips occurred - before ar.y 23

modifications were made to Unit 2. No additional damage occurred during these events.

The amount of information in-IN 85-86 is barely sufficient to establish.that these events were probably CRD power supply overvoltage protection trips. There is no information about the location 4 or timing of the lightning strikes.

5 - Farley Unit 2, March 27, 1984 (LER 84-004).

The utility reported that a lightning strike caused a' surge on the plant's AC distribution systam. This surge caused the primary CRD power supply to trip on overvoltage. Capacitive coupling between.the primary and the backup power supply allegedly caused the backup CRD power supply to trip, as the backup is fed from a motor-generator set. The control rods dropped into the core, and the unit tripped on negative neutron flux rate.

2. Power Line Transient - trip of motor-generator fed supply ,

rules out power line transient.

5. Capacitive coupling - it was not stated whether the equipment housing was grounded. The theory about a power surge and capacitive coupling is undeveloped, and seems to.

be more of an. excuse than an explanation. The magnitude of voltage surge needed to induce a significant signal on adjacent equipment through capacitive coupling would probably have caused equipment damage to the primary. supply.

6 - Dyron Unit 1, July 13, 1985 (LER 85-068).

The utility stated that the~ reactor trip was apparently caused by lightning. It was stated that the lightning had hit containment, and had been conducted to ground through structural steel components. The surge had induced voltages into cables passing.

.through containment penetrations. These surges caused the CRD power supplies to fail. The response at Byron ~1 was to improve the containment structural-lightning protection.

1. Causal - it.is not stated why the strike was known to have struck containment (i.e. it wr.s oLacrved,- it lef t burn-marks, etc.).

4. Magnetic coupling - no attempt was mhde to quantify the.

mutual inductance between the cables and the suspected current path.

7 -

Farley Unit 2, July 15, 1985 (LER 85-010).

This event resulted in the independent failure of the CRD power-i supplies and the 4160 volt bus transfer scheme, which-caused-the simultaneoun loss of all three reactor coolant pumps. The utility reported that the event was apparently caused by-24 l .-

c- .-

i l

-lightning, and that no cause for the bus transfer failure could l be determined. The CRD power supply failure was not addressed. l No corrective actions were reported. 1 8 - Zion Unit 2, June 27, 1986 (LER 86-016)

The trip was attributed to lightning striking the_ containment  !

lightning rods. Eight out of ten overvoltage protection devices on the CRD power supplies tripped. Five resistance temperature detectors (RTDs) were damaged during the event. The reactor tripped due to overtemperature delta-T readings caused by the failed RTDs. The cause of the event was believed to be that the lightning surge current path included the containment cable penetrations, which somehow induced a voltage on the cables passing through the penetration.

1. Causal - actions were taken in 1980 to prevent recurrence, but are not discussed here.

4. Magnetic coupling - no attempt was made to estimate the mutual inductance between the cables and the suspected current path.

9,10 - Byron Unit 1, July 29 and July 31, 1987 (LER 87-017).

Lightning is given as the cause of these events. The power supply overvoltage protection had tripped on nine out of ten circuits on the first event, and on three out of ten circuits on the second event. The only modification which the licensee made was to change the power supply grounding to be similar to Unit-2, which had not experienced any similar events. The lightning protection was inspected, and found to be adequate. The licensee also compared the= plant design and operating record with the Zion unit, which has similar equipment. Westinghouse was contacted, and determined that no action should be taken to make the power supplies less susceptible-to voltage surges.

1. causal - Westinghouse advised no action, even though this same event had occurred on at least 8 previous occasions at other plants.

5. Capacitive coupling - the LER states that the equipment housing was grounded.

11 - Vogtle Unit 1, July 31, 1988 (LER 88-025).

The utility reported visual verification that lightning struck the containment building immediately prior to the reactor trip.

Simultaneous problems with the Emergency Response Facility, Fire

-Protection, and Security computers were reported. An inspection of the containment revealed no physical signs that lightning had struck. The lightning protection on the containment building was foundrto be incomplete, as only one down conductor (usedLto conduct lightning to ground from the top of a building) had-been 25

.[ O'

' *' Je

^

tconnected. tooth'e stationigrounding structuref(contcinment buildings-_usuallyLhaveifour or_more evenly spaced down 1 conductors). EThe,CRD' power supplyJovervoltage_ protection ~had-activated. NoLtechnicaljexplanation forothe: event:was_provided.1 1 Surge. arresters were put:in:the CRD ac power < supply; lines..

1.- Causal -this is the.first LER which supports afcausal 4 relationshipLbetween lightning and the event. .

2. Power Line Transient - the LER does not> state.that a-line ,

transient' occurred, but surge arresters were placed-in the s ac power supply to the CRD. ,

12,13 - Braidwood Unit 1 and Unit 2, October 17,-1988f(88-023).

The reactor trip at both units was attributed to 1ightning.

There was an_ independent failure:of reactor: parameter transmitters (pressurizer pressure, coolant flow,-and pressurizerL level). Ten out'of ten overvoltage protection ^ circuits,on the CRD power supply of Unit 1 had activated, and'six-ofLten overvoltage protection _ circuits on the Unit 2 CRD power supply ' ,

activated. Station grounding was inspected and found to be '

adegoate. No corrective actions were reported.

14,15 - Braidwood Units 1 and 2, July 18, 1989 (LER-89-006).-

The reactor trip at both units was directly attributed toi '!

lightning. Ten out of ten overvoltage protection'-circuits actuated on Unit 1. -No1 details were given for. Unit-2.~-The unit trips 1 occurred at separate times, approximately-6 minutes apart..

No-corrective actions were reported.

1. Causal - the-report states that the trips. wore "directly attributed"'to lightning,' but it_is not< stated 1whether:a lightning strike was observedelmmediately' prior to the trip.

16 - Braidwood1 Unit 2,' September ~7,L1989'(LER 89-004).~

A video recorder was used to monitorfthe site for: lightning strikes. There were three' strikes'_tolthefplant'andsone~to the switchyard within a seven minute' period.- oneJof the~ strikes' hit

~the Unit 2 containment building' shortly before Unit 21 tripped?due to actuation 1of the_overvoltage protection on the CRD. power supply. The cause was-believedito be due to a rise in_the ground'

~

4 potential from the lightning strike.. .A time-delay _wasLadded: ton the overvoltage protection circuits.

1. Causal - A-video recorder was:used to establish-the causal:

. relationship betweenia local strike and this_ event. .The1, time delay had-previously been1 rejected by. Westinghouse:

_(Byronil,- .LER 87-017) .

Power Line: Transient'- the-LER doesn't establish.that a-

~

2.

power line transientiis nottassociated'with-this event, 26- .

.7

,....---.# * .s .,

'because it is not explicitly stated that nolline transient occurred.

6. GPR - this is the first LER to consider GPR as a possible cause or contributing factor for this event. The mechanism-by which GPR would effect the CRD OV protection is not described.

17 - Braidwood Unit 1, June 8, 1990 (LER 90-008).

The utility reports that lightning is the cause of the event, but states'that the location of the strike was not known. It is not stated whether lightning was observed or heard immediately prior to the reactor trip. The OV protection in three CRD power supply cabinets tripped. No immediate corrective actions were taken.

1. Causal - the impact of the time delay on the CRD OV protection which was added after the previous event is not discussed.

6. GPR - the utility-tried surge arresters to combat-GPR, but then takes no further action when these surge arresters _are apparently ineffective. No discussion of GPR. ,

18 - Byron Unit 1, August 19, 1990 (90-011).

It is stated that a lightning strike immediately prior to the-trip caused the event. Nine out of ten CRD power supply overvoltage protection circuits actuated. Surge, protection had been added to the overvoltage protection'for the CRD power supplies. The corrective action stated was that the existing power supplies will be replaced with a different model that incorporates a time delay and automatic trip reset.

It is not stated how it was established that lightning struck immediately prior to the event. Although surge protection was attempted, it does not appear that any measurements were taken_to determine the actual effects of the lightning strike. The replacement of the power supplies may_bela generic, recommendation by_ Westinghouse; this is important~for event 20.

1. Causal - the LER does not describe how it was established that lilhtning struck immediately prior to the event.-

Instead of trying to find-out what: mechanism is responsible for CRD OV protection actuation, and_possibly other effects, the solution reached here_is to replace CRD power supplies.

6. GPR - even though GPR was raised-as a possible cause for-this event at the Braidwood station, which is owned by_the, same utility,-no further mention is'made of it-here.

19 - Comanche Peak Unit 1, September 8, 1990-(LER-90-028).

It is~ stated that lightning was' believed to have' caused the event. Two out of ten OV protection circuits for.the CRD actuated. The--licensee stated that a surge on the 120 volt ac:

27

.i-

I system caused the event, and that surge suppressors were being installed.

1. causal - The mechanism for this trip appears to be the same as the earlier trips, but the utility is attempting to resolve the problem with 120 volt AC surge suppression, which was already tried by other utilities. No apparent communication between the utility and other utilities or Westinghouse on the other events.

2. Power Line Transient - the licensee attempts to resolve the problem with power line surge arrestors, but does not state if a power line surge occurred, or why it didn't affect other equipment.

20 - Farley Unit 2, August 6, 1991 (LER 91-005).

The utility states that the event was caused by a lightning induced transient. Control roda dropped into the core on loss of CRD power. It is not stated that any overvoltage protection tripped. Tr.e proposed mechanism for the trip is an induced surge in the power supply cables which temporarily caused the CRD power to go to zero, allowing the rods to drop. Service water equipment and security system failures were reported. No corrective actions were reported, and the extent of these failures was not given,

1. causal - If the utility replaced the power supplies based on a generic recommendation by Westinghouse, it is possible that the OV protection automatically reset. However, this is not stated.

2. Power Line Transient - although this event is attributed to a transient in the power supply, it sounds like the utility means a magnetically coupled surge. A power line surge is not explicitly stated. -

4. Magnetic induction - no attempt is made to calculate the mutual inductance between the lightning path and the cables.

After 20 events over 12 years, the root cause of the activation of CRD overvoltage protection due to a local lightning strike is still unknown. Apparently, utilities have tried improving lightning protection of containment and other structures, and this has had no effect. In addition, surge protection and shielding have been scrutinized and improved, which has also had no effect. However,.there is one apparent source of lightning induced transient which does not appear te have been investigated in depth by the utilities, which is GPR. If the overvoltage protection device for these power supplies is using instrument ground as a reference, it is possible that a transient on instrument ground would cause the overvoltage protection to actuate. Other explanations are possible. Improving the grounding of equipment and structures would not prevent problems caused by GPR.

28

e .

l 10.2.2 APRM Spurious Dignal The sano critoria which woro used to ovaluato tha CRD power supply OV trips woro also used to ovaluats those five events, which occurred at boiling water reactors (DWRs). Those ovents are not as wide pread geographically or temporally as the CRD power supply OV protection trips.

1-4 Grand Gulf Unit 1, August 15, 1988 (LER 88-012), July 22, .

1989 (LER 89-010), November 7, 1989 (LER 89-016), August 10, 1991 (LER 91-03^'

These four ovents all report that the averago power rango monitors spiked high, resulting in a reactor trip. Additional equipment misoperation occurred oa three of the four events, consisting of two events of spurious reactor coro isolation cooling (RCIC) activations and tt.o events of spurious high pressure coro spray (llPCS) trip. The correctivo action for the first three events was to install a " lightning dissipation system." The fourth event occurrod after the lightning dissipation system had been installed, so it listed a corrective action of o>;tonding the lightning dissipation system.

1. Causal - The lightning dissipation system is described as lightning dissipation arrays which reduce the potential of lightning str' king the plant sito and other vulnerable plant structures. This explanation is totally lacking in technical merit, and sounds similar to the 18th century conception that lightning rods could be unod to " dissipate" static chargo before it formed lightning. The offectivonoss of such atrays has boon largely dismissed by technical exports (Goldo). If the arrays are actually grounding equipment, they don't reduce the potential for lichtning to strike the oito, but provide a low-impodance path to ground.

This protects structures by decreasing the probability that a structure will take a direct strike from lightning, but does not provent GPR offects.

5. - Brunswick Unit 2, September 10, 1984 (LER 84-025).

Lightning is given as the cadno of an RPS actuation caused by APRM high neutron flux signal. The unit was in a refueling / maintenance outage at the time of the ovent. Very little information was provided about the Unit 2 trip, apparently because it occurred while the unit was not operating.

The interesting point about these five events is that the first Grand Gulf LER (88-012) reports that the ground trip of a large fan caused the trip of (c3 APRM channels on June 8, 1988, resulting in a half-scram. No investigation was made into the similarity between a lightning strike and a ground fault of a large fan, but it is likely that both events would tend to raiso 29

10.2.2 APRM Spurious Signal The same criteria which were used to evaluate the CRD power supply OV trips were also used to evaluate those five events, which occurred at boiling water reactors (DWRs). These events are not as widespread geographically or temporally as the CRD power supply OV protection trips.

1-4 Grand Gulf Unit 1, August 15, 1988 (LER 88-012), July 22, 1989 (LER 89-010), November 7, 1989 (LER 89-016), August 10, 1991 (LER 91-010)

These four events all report that the average power range monitors spiked high, resulting in a reactor trip. Additional equipment misoperation occurred on three of the four events, consisting of two events of spurious reactor core isolation cooling (RCIC) activations and two events of spurious high pressure core spray (llPCS) trip. The corrective action for the first three events was to install a " lightning dissipation system." The fourth event occurred after the lightning dissipation system had been installed, so it listed a corrective action of extending the lightning dissipation system.

1. Causal - The lightning dissipation system is described as lightni.ng dissipation arrays which reduce the potential of lightnt.ig striking the plant site and other vulnerable plant struc* ares. This explanation is totally lacking in technical merit, and sounds similar to the 18th century conception that lightning rods could be used to " dissipate" static charge before it fermed lightning. The effectiveness of cuch arrays has been largely dismissed by technical experts (Golde). If the arrays are actually grounding equipment, they don't reduce the potential for lightning to strike the site, but provide a low-impedance path to ground.

This protects structures by decreasing the probability that a structure will take a direct strike from lightning, but does not prevent GPR effects.

5. - Brunswich Unit 2, September 10, 1984 (LER 84-025).

Lightning is given as tha cause of an RPS actuation caused by APRM high neutron flux signal. The unit was in a refueling / maintenance outage at the time of the event. Very litt;e information was provided about the Unit 2 trip, apparently because it occurred while the unit was not operating.

The interesting point about these five events is that the first Grand Gulf LER (88-012) reports that the ground trip of a large fan caused the crip of two APRM channels on June 8, 1988, resulting in a half-scram. No investigation was made into the similarity between a lightning strike and a ground fault of a large fan, but it is likely that both events would tend to raise 29 i

the ground potential in the local aron relative to distant ground.

It should be noted that no events woro reported whero spurious APRM signals or CRD power supply OV trips resulted from a lino strike. Other explanations are possible. Becauso all affected PWRn are Westinghouco units, it is possible that they all have the namo power supplies for the CRD. One LER identified theco power supplies as being provided by Lambda Electronics. In an unrelated event at Arkansas lluclear 1, Unit 1 (LER 313-86-004), a Lambda power supply feeding the turbine electrohydraulic controllor (EHC) misoperated, immediately after a lightning striko in the plant switchyard was observed. However, the power supply overvoltage protection was not the cause of the trip, which was apparently due to a change in voltago sotting from 16 .- .

to 15 V dc.

These events could be related to the magnetic coupling of cabling runs to the lightning current, but because lightning strikes at random locations, this would not explain the large number of ovents at one unit of a multiple unit plant (Byron) with the absence of events at the other unit. EMI is similarly unlikely, since there should be other CRD power supply events or spurious APRM signal events reported which are unrelated to lightning.

However, none of the affected facilition reported any similar events resulting from causes other than lightning. GPR of the station ground is the only effect which would account for a similar plant responso for lightning strikes at different locations around the plant, but a different responso for two units at the same site.

10.2.3 Fire and Firo Suppression Failure Events The most potentially serious event would appear to be a lightning strike causing a loss of firo protection capability and concurrently initiating a fire. Events which involved a failure of fire protection or a fire are summarized below.

1. Three Mile Island Unit 2, (LER 320-82-010), June 1, 1982 The AIT halon system actuation spuriously actuated during a thunderstorm. The event was attributed to activation of an ultraviolet light detector in the AIT. No corrective action was taken at tho timo.

2. Three Mile Island Unit 2, (LER 320-82-023), June 29, 1982 This event was the second spurious AIT Halon system actuation at TMI 2. Louvnrs were installed on the AIT structure openings as a corrective action.

3. McGuiro Unit 1, (LER 369-82-076), November 4, 1982 Lightning damage to an offsito insulator on a transmission line caused the loss of the second of three main fire pumps. The 30 1

. w first pump was out of service at the timo.

4. Three Mile Island Unit 2, (LER 320-83-043) August 27,1983 to Top). ember 12, 1983 tr intake Tunnel (AIT) halon system was disarmed during a t naoratorm on multiple occasions, in order to provent spurious actuation of the AIT halon system. Spurious _ actuation of the AIT halon system had occurrod on two previous occasions. The reactor was defuelod and inoperable at this time, so the safety significance wac minimal.

5. Zion Unit 2, (LER 304-86-016), June 27, 1986 This event is also reported in section 10.2.1 and involved the failuro of fivo RTDs and tripping of the CRD power supply overvoltago protection circuits. In addition to thoso and other failures, the battery room firo alarm also spuriously alarmed.

This event is of minimal important from a loss of fire protection standpoint.

6. Dusquehanna Unit 1, (LER 387-88-015), July 17, 1988 1 partial loss of fire protection occurred due to the failure of  !

two transponders and a DC transmission circuit card. The failed components appear to have boon microprocessor chips, but an exact description of the failuro mechanism or failed component was not given. There was also no description of the location of the lightning strike or other plant effects (loss of power, local strikos, etc.).

7. Vogtle Unit 1, (LER 424-88-025), July 31, 1988 This event was also reported in section 10.2.3 and involved the tripping of the CRD power supply overvoltage protection. The firo protection computer alarmed, but was apparently not damaged.

8. Summer Unit 1, (LER 395-88-010), October 6, 1988.

Lightning struck a meteorological tower, which apparently resulted in the fire protection computer misoperation. The fire protection computer was offlino for 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br />. No actions were apparently taken to provent recurrence of the computer failure due to lightning strikes on the meteorological tower.- This is a possible GPR ovent.

9. Cooper Unit 1 (LER 298-90-007), June 17, 1990 The halon system spuriously actuated during a lightning storm, and a " control board" in the halon system control circuitry failed, due to the failure of one transistor and throo integrated; circuits. The cause of the failed components and the location of the lightning strii;e which caused the damage were not reported.

The following events concern lightning strikos which resulted in an onsite fire.

31

1. Oyster Creek Unit 1 (LER 219-87-030), November 5, 1987 This LER reports that a voltage transient was caused by a surge arrester failure. Thore is no indication given of whether a storm was in progross at the time. However, tho surgo arrester exploded and caught on fire. This event is of arguable importancu, since the LER does not specifically attribute the failure to lightning. No other equipment failuro was reported.

2. Dequoyah Unit 2, (LER 328-88-034), August 15, 1988 A flachover occurred in the 6.9 kV switchgear during an electrical storm, but apparently after the storm had passed. The flashover was attributed to moisture and particular contanination in the switchgear. Other equipment tripped or isolated duo to the lightning activity, but no equipment was damaged by the lightning strikes.

3. Waterford Unit 3, (LER 382-90-012), August 25, 1990 A circuit breaker exploded and burned when it failed to clear an offsite fault. The breaker failuro protectivo rolaying failed to operato. When a circuit breaker fails to successfully extinguish the arc which is created when the contactors separate, a potentially dangerous amount of energy is released in the circuit breaker which can lead to destruction if other circuit breakers do not isolate the fault. Debris from the explosion was thrown on neighboring equipment.

4. Yankee Rowe Unit 1 (LER 029-91-002), July 16, 1991 A lightning strike at an unstated location on the 115 kV lines serving the plant caused a surge arrester on the unit transformer to fail and burn. Additional equipment tripped or suffered blown fuses, but this appears to be due to low voltage. Tolophone and microwave telecommunications were also damaged, apparently by other lightning strikes.

It is interesting to note that surge arrester failures have occurred for these events and the events in section 10.1.3 without concurrent lightning. It 5 percent of all power line surges exceeded 120,000 amperes as alleged in the February 3, 1981 ACRS meeting, the majority of arresters should fail due to lightning, as the maximum rating for arrestors in 65,000 ampores.

However, for the events reviewed in this report, only 1 event (Yankee Rowe, 029-91-002) reported the concurrent failure of a surge arrester with a lightning strike. In fact, this event resulted in the concurrent failure of several arresters and flashover of a second lino at a disconnect switch, and was probably due to multiple strikos.

Although some of these events do not involve serious failures, or cannot be verified to have involved lightning, it is likely that there have been additional lightning-induced fires or losses of fire protection that were not reportable events. Therefore, it is not unreasonable to estimate the frequency of a lightning 32

f i

induced-fire as-4 out of 177 ovents (for a probability of 0.023 firos/striko-ovent), and to estimato the loss of fire supprossion  !

system as 8 out of 177 ovents, excluding intontional de-activation of the Halon-system at TMI (for a probability of 0.045 failuros/striko-ovont). The probability product of those two ,

events gives an estimate of the probability that a lightning strike could start a tiro and that a second striko would cause ,

fire suppression system damago or failure (this ovent has novor- .

occurred). This probability is 0.001 ovonts/striko-ovent. It is l noted that this is only an estimato, assuming that both ovents. I would be independent (i.e., one lightning striko doesn't cauno i both events).

During the porlod. 1980-1991, 46 licensees rocoived opqrating '

licensos. Taking into account the year in which a licansoo was given an operating license, thoro waro approximately 967 operating years for the 111 licensies with operating and licensed ,

facilities at the end of 1991. This yields a probability of-177 striko-ovents for 967 operating years, or 0.183 striko-ovents per operating year.- Multiplying this probability with tho .

probability of a firo and loss of firo supprossion por strike i ovent yields the probability of firo and loss of' fire nupprossion due to a lightning striko, which is 1.8E-4 ovents/roactor operating year.

This probability noods to be further offset by estimates-of the ability of the plant fire brigado to detect and extinguish.a firo-with a partially or totally disablod firo protection system, the- '

probability that a lightning-croated firo.will occur in an area with failed firo protection, and other factors. However, it assigning a value to those probabilities would seem speculativo r at'best, and could easily croato results that would support either the need for regulation or no need for-regulation. For examplo, it could be asserted that tho fire brigado has a 90-percent chance of being notified of any fire by observation, and that a fire has a 10 percait chance of starting in a vital area, such as a cabic spreading room. These arbitrary probabilities- '

would cause the probability of a firo-related event to drop to .

1 8E-6 ovents/ reactor operating year.

. Even if tholprobability of core damago was a certainty in this event, the frequency of core damage would be below the Commission's proposod safety goal policy (Taylor) . ,

L l- -A possiblo resolution for provention-of lightning related= fires l- would be to develop a regulatory guido'for structural--lightning:

protection. _However, it should be noted that 3Lof the 4 reported ovents involving fires' wore due to lino strikos, which would not be abated by building protection (or increased levels of surge-arrestor ampacity).- The-events resulting=in loss of fire protection appeared to result primarily from local-strikonf(7 of i 8 events). Adding structural lightning protection could. increase the number of plant strikes-(Goldo), possibly resulting in an l l 33

-j q

1:

l _ a _- ._ . _ _ . _ _ _ .

_ . _w- _a;

. _. _ .x;

. l I

increase in lightning events reculting in a loss of firo protection. It is evident that the need oxists to develop more information on the effects of local strikos, particularly regarding the phenomenon of GPR.

10.2.4 Evonts of Apparent Low significance of the 66 ovents which appeared to involve local strikes, 20 involvt.J CRD power supply trips, 5 involved APRM spurious signals, 10 involved firos or fire protection, and 31 involved events of apparently low significance. Thoso events woro not investigated in depth, in part because of the lovel of uncertainty involved when compared with the ovent loading to a ,

firo concurront with loss of fire supprossion,-which is itself i too uncertain to provido any clear basis for regulatory action.

It is not possible to meaningfully include thoso-ovents in a PRA, because most of them involvo unique occurrencos with an unknown initiating event frequency. For examplo, if ono lightning event j in 11 years at a single unit results in concurrent failure of ,

Jive hot log RTDs, an accumulator lovel transmitter, two hot log  !

low lovcl amplifiers, an R11R loop temperature amplifier, and simultaneous spurioun activation of 8 out of 10 CRD power supply ovary cago protection circuits (LER 304-06-016), what probability should that ovent roccivo? In addition, all components failed in a conservative direction. It would appear to be more cost offective to study the mechanism by which a local lightning striko is affecting plant equipment and develop .

protectivo measures, rather than try to model possible offects from a mechanism which is not understood.

1. Transient Indications on Process Indicators Two spurious pressurizar low pressure signal trips (LER 250 <

019), (LER 250-86-032); control building air intake radiation monitor spurious actuation, multiple events (LER-331-84-020).

2. Damage to Process Indicators Sovoro transient on seven main steam pressure transmitters causing two failures (LER 272-80-031); simultaneous failuro of two out of four refueling water storage tank lovel transmitters (LER 413-89-023).

3. Failure-of External-Beismio, Gas Stack, or Weather Monitoring Instrumentation Failure of soismic monitors; failure of wind direction sensort-failure of off-gas flow recordor; failuro of wind speed sensor; failure of air temperature indication. LERs - 220-80-020, 259-83-002, 259-81-026, 259-82-001, 259-82-015, 259-62-043, 259 058, 271-80-028, 271-84-104, 293-81-018, 302-82-048, 302-83-032, 302-83-045, 320-82-019, 321-81-074.

34 n , _

.~ -

e .c o .

4. Equipment Failure Two similar f ailures of electrohydraulic control (EllC) system components (LERs 313-86-004 and 313-87-002); fuse blow in D.C.

power supply to radiation monitor (LER 263-84-022); potential transformer fuses blow (LER 255-86-028); failure of output amplifiers for forced balance acceleromotor attributed to '

lightning (LER 416-82-003); control room monitoring instrumentation unavailablo (LER 389-83-059).

l

5. Multiple Failures

a. (LER 412-90-011) spurious containment high' radiation, seismic monitor, and looso parts monitoring signals, and failure of meteorological equipment.

b. (LER 346-80-068) failure of wind direction sonsor and damage to meteorological-computer electronic components.

c. (LER 265-87-007) lightning struck a meteorological tower, damaging the local power transformer and distribution panel.

In addition, the undervoltage relays operated for the bus feeding the meteorological tower transformer.

d. (LER 458-87-016) automatic start of annulus mixing system I and gas treatment system. Not caused by spurious actuation  ;

signal. ,

o. (LER 155-83-013) blown static inverter fuse, damage to telephone systems, security system equipment,-and domestic water controls.

11. Precursors to Potential Core Damage Accident Studies Ton lightning-related events woro identified in the NUREG/CR-3591 and NUREG/CR-4674 studies of procursors to potential core damage accidents. These are-reviewed for potential impact.

1. NUREG/CR-3591, 1980-1981.

A. Indian point 2, LER 247-80-006 - The conditional core damago probability resulting from this event was estimated' to be 1.3E-5. The only effect caused by_ lightning.was the L loss of offsite power.

B. Crystal River 3, LER 302-81-033 - The conditional core =

L damage probability resulting'f7om this event was estimated l to be 3.7E-4. The only effect caused by lightning was the.

loss of offsite power.

- C.' Prairie-Island 2,. LER 306-80-020 -- The conditional core; damage probability resulting from this event was'estimatedf i 35 1

Y ..

+

..-_.n --. _ ,

l

- s to be 4.7E-5. The only offect caused by lightning was the loss of offsito power.

2. NUREG/CR-4674, 1985.

A. Turkey Point 3, LER 250-85-019 - The conditional core damago probability resulting from this event was estimated to be 8.96E-4. Lightning is attributed with generating a spurious low pressurizer pressure signal, which resulted in i a plant trip. Subsequent to the trip, the plant experienced >

additional problems with valves due to moisture in the .

Anstrument air linos. The only effect which might have been-caused by lightning was the unit trip due to the spurious -

low pressurizer-pressure signal.

B. Farley Unit 2, LER 364-85-010 - The conditional core damage probability resulting from this ovent was estimated to be 1.41E-5. This event was caused by the control rod drive power supply overvoltago protection actuating, which appears to be due to ground potential rise (GPR) - o f fects from local lightning strikos. This event was complicated by.

the early opening of the generator " output breakers." The cause for the misoperation of the breakers could not be determined.

C. Byron Unit 1, LER 454-85-069 - The conditional coro damage probability resulting from this event was estimated to be 7.15E-7. This event was caused by an instrument technician inadvertently grounding a: power lead to a power supply. The power supply was found to be failed as a result of an earlier lightning strika. No explanation was provided-regarding how lightning _ caused the power supply failuro or what part of the power supply failed. The earlier lightning.

strike resulted in the activation of the control rod drive power supply overvoitage protection,.which appears to bo due to GPR.

3. NUREG/CR-4674, 1986.

A. Lacrosso Unit 1, LER 409-86-023 - The conditional core damage probability resulting from this event was~ estimated to be 2.0E-5. The only effect caused by lightning was the loss of offsite power.

4. NUREG/CR-4674, 1987.

r A. Calvert Cliffs 1, LER-317-87-015 - The conditional; core damage probabilityLresulting from this event was estimated to be 7.0E-8. The event was attributed to' lightning because a storm was in progress, but was caused by flashover of the-main transformer bushing. This unit is located on the coast, and the. bushing flashover could have been caused-by l 36

. r salt or other contamination. The " plant computer" became  :

unavailable concurrent with the bushing flashover. The -

cause-of the plant computer unavailability was not stated.  ;

5. NUREG/CR-4674, 1989.

A. Crystal River 3, LER 302-89-025 The conditional coro ,

damago probability resulting from this event.was estimated to be 6.3E-5. The ovent resulted from a loss of offuite-power caused by lightning. The event was complicated by the '

unavailability of one of the diesel generators, which was being serviced. .;

B. Grand Gulf 1, LER 416-09-016 - The conditional coro  !

damage probability resulting from this' event was outimated to be 1.2E-6. This event was caused by a local lightning $

strike which created a spurious signal on three out of eight average power range monitors (APRMs), which may be a.rosult of GPR. The ovent was complicated by the unavailability of the reactor core isolation cooling (RCIC) system, due to surveillance testing.

The events with the highest conditional core damage probability i involved additional equipment failure which was not attributed-to lightning. Several of the events were complicated by equipment damage or misoperation which may have been caused by GPR from a' ,

local 1 11ghtning strike (2.C, 4.A). .However, those events had low conditional core damage probabilities (7.15E-7 and 7.0E-8, respectively). Most of the events were not the~ result of an -

absence of lightning protection, but rather woro caused by the effects of lightning even when state-of-the-art protection is ,

used.(i.e. loss of offsite power).

The only area where additional research and regulation might' result in a decrease in public risk is in protection ~of plant systems from the effects of GPR. However, evon if it assumed that every local lightning event during the period-1980-1991.(71 events) had the average calculated conditional core damage probability for a GPR-related event (CDF ='3.9 E-7), and that regulation of GPR protection.would climinato all GPR-related events, then the decrease in core damage frequency (CDF) for the industry would only be 2.9E-8. Under the Commission's safety.

goal policy (SECY-91-270), it does.not appear that regulation of- J lightning protection would result in a decrease in CDF of greater-than 1.0E-5.

12. Conclusion PRM-50-56 states that nuclear power plants require additional regulation because of the potential for electrical transients in  !

- instrumentation, control, and power circuits,--and:that this-regulation.should. apply to all.new license applications as well 37 .

t y + 6 .4,. -"i-. -- . ~ ~ , + 4 , , , , C- p . - . -

._ - - , - - r. ,. y-,-- r- .4.. ---,-<ew-,.. . -.

as licenso renewals. Based upon the analysis in this report, there does not appear to be any basis for increasing the regulatory requirements associated with electrical transients as a class. All sources of electrical transients have been studied or are currently being studied by the 11RC, with the apparent exception of GPR caused by local lightning strikes. Ilowever ,

based upon operating events and the Commission's safety goal, regulation of lightning protection does not appear to bo justified on the basis of safety significanco, llowever, in light of the amount of effort which has been expended on this issue in the past, and the increasing reliance on digital controls, it seems prudent to considor changes to regulatory requirements for futuro plants.

Regulation of system protection from EMI is currently being -

studied under a separate program, and does not need to be addressed in response to this petition. The structural and power lino protection practicos currently used by licensees appear to adequately protect licensed facilition from the offects of direct strikes based upon the operating experiences reviewed in this report. Therefore, existing standards could be used as the technical basis for consideration of any new regulation for structural and power lino protection. The effect of GPR on power plant systems is not well documented in the technical literature, and further research will be required to establish a technical basis for regulation.

38

)

r e .

l e .

e Ref.919.1meJi o Blackburn, J. L. , " Ground Fault Relay Protection of Transmission Lines" AIEE Transactions Part III, Power Apparatus and Systems, August 1952 o Erown, G.W., " Lightning Performance II, Updating Backflash Calculations," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-97, No. 1, January / February 1978, o Brown, G.W., " Lightning Performance I, Shielding Failures-Simplified," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-97, No. 1, January / February 1978.

o -Clayton, R.E., et. al. " Surge Arrester-Protection and Very Fast Surges," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-102, No. 8, August 1983. ,

o Cleveland, F., et. al. " Solar Effects on Communications," ,

IEEE Power Engineering Review, September 1991.  !

o Elgerd, O. Electric Enercy Systems Theory - An Introduction, Second Edition, 1982, McGlaw-Hill.

o Eriksson, A.J., Stringfallow, M.F., Meal, D.V., " Lightning-Induced overvoltages on Overhead Distribution Lines," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-103, No. 4, April 1982.

o Galbrois, G.L., " Lightning current Magnitude through L Distribution Arresters," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-100, No. 3, March 1981.

o Glasstene, S The Effects of Nyclear Weapons, Third-Edition, is?7, United States Department of Defense.

o Golde, R. H. Liahtnina l

o Greenwood, A., Electrical Transients in Power Systang, 1971, t Wiley-Interscience. +

o IEEE Dictionary of Standard Terms o Ikeda, I., Tani., M.,-Yonezawa, T., " Analytical Technical of Lightning Surrjes Induced on Grounding Mesh of PWR Nuclear

. Power Plant," IEEE Transactions on Energy Conversion, Vol.

5, No. 1, March, 1990, o Jackson, D.W., et. al . , " Survey of Failures of Surge L -Protective Capacitors and Arresters on AC Rotating _ :1

! Machines," IEEE Transactions on Power Delivery, Vol. 4, No.

3,. July 1989.

39 l

e- ..

o Joyce, J.S., " Torsional Fatigue of Turbine-Generator Shafts I caused by Different Electrical System Faults and Switching Operations," IEEE Transactions on Power Apparatus and i Systems, vol. PAS-97, pp. 1965-1973, Sept./Oct. 1978.

o Kappenman, J.G., Albertson, V.D., " Cycle 22: Geomagnetic '

Storm Threats to Power Systems Continue," IEEE Power Engincoring RevioW, September 1991. .

o Klein, K.W., Barnes, P.R., Zaininger, H.W., " Electromagnetic  ;

Pulso and the Electric Power Network," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-104, No. 6, June 1985. ,

o Lee, K.S.H. "EMP Interaction: Principles, Techniques and . '

Reference Data," Air Forco Weapons Laboratory, AFWL-TR "

402, December 1980.

o Lin, Y.T. et. al. " Characterization of Lightning Return Stroke Electric and Magnetic Fields from Simultaneous Two-Station Measurements," Journal of Geophysical Research, Vol. '.

84, No. Clo, October 20, 1979.

o Longmire, C. "On the Electromagnetic Pulse Produced by Nuclear Explosions," IEEE Transactions on Antennis and Propagation, Vol. AP-26, No.1, January 1978.

o Martzloff, F.D., " Coordination of Surge Protectors in Low-Voltage AC Power Circuits," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-99, No. 1, January 1980.

o Martzloff, F.D., "The Propagation and Attenuation of Surgo i Voltages and Surge currents in Low-Voltage AC Circuits," '.

IEEE Transactions on Power Apparatus and Systems, Vol. PAS-102, No. 5, May 1983.

o Miller, T.J.E. B_eactive Power Control in Electric Systems,-

1982, Wiley-Interscience.

o Mitani,_H., " Magnitude and-Frequency.of Transient Induced Voltages in Low Voltage control Circuits of Power Stations- '

and Substations," IEEE Transactions on Power Apparatus and .

l Systems, Vol. PAS-99, No. 5 September / October '1980.

L

, o Nishikawa, H., et..al., " Vacuum ~ Circuit-Breaker Switching L Surge' Influence on' Low Voltago Instrumentation Circuits," .

IEEE Transactions on Power Apparatus <and' Systems, Vol. PAS-

, 102, No. 7, July 1983.

l o' NUREG/CR-3069, " Interaction of Electromagnetic Pulse with L Commercial ~ Nuclear Power Plant Systems," U.S. Nuclear L Regulatory Commission.

40

't

~~.- , , . , ,, .,- ,. _ u,,

- -- -.a ., .,,;-, - , _.

s .

o Odenberg, R., Braskich, B.J., " Measurements of Voltage and Current Surgos on the AC Power Line in Ccuputer and Industrial Environments," IEEE Transactions on Power Apparatus and Systems, Vol. PAS-104, No. 10, October 1985.

o Osaman, L.R., IEEE Surge Protective Devices Committee, letter to G.W. Knighton, NRC, " Risk Assessment of Proposed Regulatory Guide on " Lightning Protection for Nuclear Power Plants."

o Rabinowitz, M. "Effect of the Past Nuclear Electromagnetic Pulse on the Electric Power Grid Nationwide: A Different View," IEEE *ransactions on Power Delivery, Vol. PWRD-2, No.

4, October 1987.

o Scharfman, W.E., Vance, E.P, Graf, K.A., "EMP Coupling to Power Lines," IEEE Transactions on Antennas and Propagation, Vol. AP-26, No. 1, January 1978, o SECY-92-270, August 27, 1991, " Interim Guidance on Staff Implementation of the Commission's Safety Goal Policy."

o Skitek, G.G., Marshall, S.V., Electromannetic CRacepts and hPI211_c.ilt i on s , 1982, Prentice-Hall.

o Smith, D.R., Swanson, S.R., Borst, J.D., "Overvoltages with Remotely-Switched Cable-Fed Grounded Wye-Wye Transformers,"

IEEE Transactions on Power Apparatus and Systems, Vol. PAS-94, No. F, September / October 1975.

o Uman, M.A., Master, M.J., Krider, E.P., "A Comparison of Lightning Electromagnetic Fields with the Nuclear Electromagnetic Pulse in the Frequency Range 10' - 10 7 Hz,"

IEEE Transactions on Electromagnetic Compatibility, Vol.

EMC-24, No. 4, November 1982.

o Yokoyama, et. al., " Advanced Observations of Lightning Induced Voltage on Power Distribution Lines," IEEE Transactions on Power Systems, Vol. PWRD-1, No. 2, April 1986.

41

e Inkle _One - LER Data _

i Unit Dkt Year No Type of LER -

1 Yankee Rowe 29 1983 19 Line striko 2 Yanhoo Rowo 29 1982 19 Line strike 3 Yankee Rowe 29 1986 4 Lino? Local?

4 Yankee Rowe 29 1983 22 Line striko 5 Yankoo Rowo 29 1901 2 Fire, Fire Protection ,

6 Big Rock Point 155 1983 13 Line? Local? l 7 Haddam Neck 213 1983 14 Line strike 8 Oyster Creek 1 219 1987 30 Fire, Fire Protection 9 Nine Mile Point 1 220 1980 20 Mot. Tower 10 Nine Mile Point 1 220 1988 15 Line strike 11

• Indian Point 2 247 1980 6 Line strike 12

• Turkey Point 3 250 1985 19 Local Strike ,

13 Turkey Point 3 250 1986 32 Local Strike 14 Quad Cities 1 254 1990 13 Line strike, other offects '

15 Quad Cities 1 254 1991 8 Lino strike, other effects 16 Quad Cities 1 254 1987 14 Line strike, other effects 17 Quad Cition 1 254 1990 4-Line strike, other effects 18 Palisados 255 1986 28 Local Striko 19 Browns Ferry 1 259 1980- 2 Met. Tower 20 Browns Ferry 1 259 1982 58 Mot. Tower 21 Browns Ferry 1 259 1981 26 Met. ToWor >

22 Browns Ferry 1 259 1980 59 Line striko,-other effects 23 Trowns Farry 1 259 1982 43 Met ~. Tower 24 Browns Ferry 1 259 1982 15 Met. Tower 25 Browns Ferry 1 259 1982 1 Mot. Tower '

'26 Robinson 2 261 1985 18 N/A.?

27 Monticello 263 1984 22 Lino? Local?

28 Monticello 263 1991 19 Line strike '

29 Quad Cities 2 265 1987' 7 Local Striko 30 Vermont Yankee 271 1984 14- Met. Tower-31 Vermont Yankee 271 1980 28 Mot. Tower 32 Vermont Yankoo 271 1991 14 Line strike, other offects. <

33 Salem 1 272 1980 31 Local Strike 34-Salem 1 272 1991 24 Line strike, other effects. y 35 Salem 1 272 1987 7 Line striko 36 Peach Bottom 2 277 1987 12 Line strike 37 Peach Bottom 3 278 1985 18 Lino strike, other_offects 38 Poach Bottom 3 278 1991 10 N/A ?

39 Peach Bottom 3 278 1986112 N/A ?

L 40 Surry.2 .

281.1991 6 N/A 7 41 Indian Point 3. 286 1980 8'Line-strike

42. Pilgrim 1 293 1983.45. Line strike- ,

43 Pilgrim 1 293 1981 18-Mot. Tower 44 Zion 1 a IN 85-86 295'1979 CRD. Power Supply OV tripL 45 Cooper 298.1987 17 Line strike l- 46-Cooper 298 1987 18 Lino ' strike L

47 Cooper- 29811990 7 Fire,-Fire Protection '

48 Cooper 298 1986 20 Line strike ,

42

., , ~.e.. . ~ a , ,,.w.on,

- , . . nn , 4 w - , N- , 4 e , -

e e l

49 Cooper 298 1987 16 Line strike 50 Cooper 298 1986 15 Line strike 51 Cooper 298 1980 31 Line striko, other offects 52 Point Beach 2 301 1987 2 Line striko, other offects 53 Crystal River 3 302 1983 32 Mot. Tower 54 Crystal River 3 302 1982 48 Mot. Tower 55

• Crystal River 3 302 1989 25 Lino striko 56 Crystal River 3 302 1983 45 Mot. ToWor 57 Crystal River 3 302 1981 34 Met. Tower 58

• Crystal River 3 302 1981 33 Line striko, other effects 59 Zion 2 - IN 85-86 304 1979 CRD Power Supply OV trip 60 Zion 2 - IN 85-86 304 1980 CRD Power Supply OV trip 61 Zion 2 304 1986 16 Fire, Fire Protection 62 *Prairio Island 2 306 1980 20 Lino striko 63 Maino Yankoo 309 1983 14 Line strike _

64 Maino Yankee 309 1983 25 Line strike 65 Arkansas Nuclear 1 333 1986 4 Local Strike 66 Arkansas Nuclear 1 313 1987 2 Local Strike 67 Cook 1 315 1981 49 Line striko 68 *Calvert Cliffs 1 317 1987 15 Lino? Local?

69 Three Milo Island 2 320 1982 18 Fire, Fire Protection 70 Three Milo Island 2 320 1983 31 Firo, Fire Protection 71 Three Mile Island 2 320 1982 23 Firo, Fire Protection 72 Three Mile Island 2 320 1983 43 Fire, Firo Protection 73 Three Mile Island 2 320 1983 25 Fire, Firo Protection 74 Throo Mile Island 2 320 1983 43 Fire, Fire Protection 75 Three Mile Island 2 320 1982 19 Met. Tower ,

76 Ilatch 1 321 1981 74 Mot. Tower 77 Shoreham 322 1990 8 Line striko, other offects 78 Shorobam 322 1985 *0

, Line striko 79 Brunswick 1 325 1984 15 Spurious APRM indication 80 Sequoyah 2 328 1988 34 Fire, Fire Protection 81 Duano Arnold 331 1984 20 Local Strike 82 Duano Arnold 331 1991 8 Lina striko 83 Fitzpatrick 333 1982 33 Line striko 84 North Anna 2 329 1986 9 Line strike 85 Davis-Bosso 1 346 1980 68 Met. Tower 86 Davis-Besso 1 346 1981 8 Line striko 87 Davis-Bosse 1 346 1937 10 Line strike, other offects 88 Farley 1 348 1991 9 Line strike, other offects 89 Farley 2 364 1991 5 CRD Power Supply OV trip 90 *Farley 2 364 1985 10 CRD Power Supply OV trip 91 Farley 2 364 1984 4 CRD Power Supply OV trip 92 Arkansas Nuclear 2 368 1982 40 Line? Local?

93 Arkansas Nuclear 2 368 1985 16 Lino? Local?

94 McGuiro 1 369 1984 17 Line strike 95 McGuire 1 369 1985 20 Line strike 96 McGuiro 1 369 1982 46 Line? Local?

97 McGuire 1 369 1985 17 Line striko 98 McGuire 1 369 1984 10 Lina striko 99 McGuire 1 369 1982 76 Fire, Fire Protection 100 McGuiro 2 370 1985 5 Line strike 43

- e t

101 Waterford 3 382 1985 54 Line strike  ;

102 Waterford 3 382 1990 12 Fire, Firo Protection  :

103 Waterford 3 382 1991 13 Line strike, other offects i 104 Susquehanna 1 387 1988 14 Line Strike .

105 Susquehanna 1 387 1986 28 Lino Strike l 106 Susquehanna 1 387 1987 20 Line Striko }

107 Susquehanna 1 387 1988 15 Fire, Firo Protection ~

108 Susquehanna 1 387 1984 28 Line Striko l' 109 Susquehanna 1 387 1984 29 Lino Strike 110 Susquehanna 1 387 1988 10 Lino striko, other offects ill Susquohanna 2 388 1985 20 Line strike, other offects  !

112 Susquehanna 2 388 1985 25 Line striko, other offects i 113 St. Lucie 2 389 1983 63 N/A 7-  !

114 St. Lucie 2 389 1983 59 Local Strike 115 Summer 1 395 1983 74 Line strike, other offects  ;

116 Summer 1 395 1988 10 Firo, Fire Protection i 117 Summer 1 395 1986 12 Line strike, other effects 118 WPPSS 2 397 1990 24 Line strike, other effects 119

• Lacrosse 409 1986 23 Line strike -

120 Beavor Valley 2 412 1990 11 Local Striko 121 Beaver Valley 2 412.1988 10 Line strike 122 Catawba 1 413 1989 21 Local Strike 123 Catawba 1 413 1985 34 Line strike ,

124 Grand Gulf 1 416 1984 27 Line strike d 125 Grand Gulf 1 416 1991 6 Line strike, other offects 126 Grand Gulf 1 ' 416 1982 3 Met. Tower 127

• Grand Gulf 1 416 1989 16 Spurious APRM inC. cation' ';

128 Grand Gulf 1 416 1989 10 Spurious APRM indication 129 Grand Gulf 1 416 1991 10 Spurious APRM indication 130 Grand Gulf 1 . 416 1988 12 Spurious APRM indication '[

131 Vogtle 1 424 1988 25 CRD Power Supply, fire .

132 Vogtle 1 424 1987 30 Line strike 133 Comancho Peak 1 445 1990 28 CRD Power Supply OV trip  !

134 Comancho Peak 1 445 1991 21 Line striko 135 Comancho Peak 1 445 1991 19 Line striko ,

136 Byron 1 454 1990 11 CRD Power Supply OV trip: ';

137

• Byron 1 - 454 1985 68 CRD Power Supply OV trip .

138 Byron 1 454 1987 17 CRD PoWor Supply OV trip  !

139 Byron 1 454'1988 6 Line striko, other effects 140 Byron 1 454 1991 2 Line strike, other effects-  :

141' Byron 1 - 454 1989 7 Line strike, other offects i 142 Braidwood 1 456 1990 8 CRD Power-Supply OV trip.

143 Braidwood 1 456 1989 6 CRD Power Supply OV~ trip 144-Braidwood 1 456 1991 5 Line strike, other offects 145 Braidwood'1 456 1988 23 CRD. Power. Supply ~OV trip-146 Braidwood 2 457-1989 1 4 CRD Power Supply OV trip ,

147'Riverbend 1 458 1987'16 Local-Strike 148 Riverbond 1 ;458 1985.63 Line-striko, other effects- 1 149 Wolf Creek 482-1985 71 LineLstriko-150 Wolf Creek 482 1985 55- Line strike 1511Palo' Verde 1 528 1987 21 Line strike,.other offects 152 Palo Verde 2- - 529L1989 'l Line strike, other effects; 44 '

k

+  %-- ,e w . ,-w-- w , , , . . . . ,-

< . s o .

?

Table _2 ,__LER_.DA1A Unit Number "ower Cause of Trip events sovel r 1 Yankoo Rowo 1 100% No reactor trip ,

2 Yankoo Rowe 1 100% RPS trip 3 Yankoo Rowo 1 101% RPS trip 4 Yankoo Rowo 1 100% Turbino trip- '

5 Yankoo Rowe 1 88% Turbino trip 6 Dig Rock Point 1 75% No reactor trip 7 Naddam Neck 1 100% No reactor trip -

8 Oyster Crook 1 1 0% No reactor trip 9 Nino Hilo Point 1 1 97% No reactor trip 10 Nino Mile Point 1 1 0% RPS trip 11

• Indian Point 2 1 100% Turbino trip "

12

• Turkey Poilt 3 1 100% RPS trip 13 Turkey Point 3 1 53% RPS trip 14 Quad Cities 1 1 100% No reactor trip 15 Quad Cition 1 1 0% No reactor trip ,

16 Quad Cition 1 1 99% No reactor trip ,

17 Quad Cities 1 1 98% Turbino trip 18 Palisados 1 0% No reactor trip 19 Browns Ferry 1 1 0% No reactor trip 20 Browns Ferry 1 1- No reactor trip -

21 Browns Ferry 1 1- No reactor trip 22 Browns Forry 1 1 73% No reactor trip ,

23 Browns Ferry 1 1- No reactor trip 24 Browns Ferry 1 1- No reactor trip 25 Browns Forry 1 1- .

No reactor trip ,

26. Robinson 2 1 98%' No reactor _ trip 27 Monticello 1 0% No roactor. trip 28 Monticello 1 100%. RPS trip 29 Quad Citios 2 1 90% No reactor trip 30 Vermont Yankoo 1 0% No reactor trip 31 Vermont Yankoo 1 89% No reactor trip ,

32 Vormont Yankoo 1 100% Turbino trip 33 Salem 1 1 100% RPS trip 34 Salem 1 1 100% Turbino trip 35 Salem 1 1 100% .Turbino trip 36 Peach Bottom 2 1 0% No reactor trip .;

37 Peach Bottom 3 1 100% RPS trip 38 Peach Bottom 3 1 97% Turbino trip 39 Poach Bottom 3 1 81% Turbine trip 40 Surry 2 1 100% No reactor trip 41 Indian Point 3 1 60% No reactor trip 42 Pilgrim 1 1 0% No reactor' trip  :

43_ Pilgrim 1 1._93%- ENo reactor trip 44 Zion 1 - IN 85-86 1 45 Cooper ~1 100% No reactor trip 46 Cooper 1 100% No roactor. trip -

47 Cooper 1 94% No reactor trip 48 Cooper 1 72% No' reactor trip 45 .

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

s i 4 . >

t

• A 49 Cooper 1 0% No reactor trip '

50 Cooper 1 94% No reactor trip 51 Cooper 1 86% Turbine trip 52 Point Beach 2 1 100% Turbine trip 53 Crystal River 3 ' 1 0% No reactor trip 54 Crystal River 3 1 90% No reactor trip i 55

• Crystal River 3 1 0% No reactor trip 56 Crystal River 3 1 98% No reactor trip ,

57 Crystal River 3 1 0% No reactor trip 58

• Crystal River 3 1 100% Turbine trip 59 Zion 2 - IN 85-86 1 ',

60 Zion 2 - IN 85-86 1 61 Zion 2 1 68% RPS trip 62

• Prairie Island 2 1 100% Turbine trip 63 Maine Yankee 1 100% No reactor trip ,

64 Maine Yankee 1 100% No reactor trip 65 Arkansas Nuclear 1 1 82% RPS trip 66 Arkansas Nuclear '. 1 38% Turbine trip 67 Cook 1 1 100% No reactor trip 68 *CP.lvert Cliffs 1 1 100% Turbine trip 69 Three Mile Island 2 1.0% No reactor trip.

70 Three Mile Island 2 4 0% No reactor trip 71 Three Mile Island 2 1 0% No reactor trip 72 Three Mile Island 2 3 0% No reactor trip 73 Three Mile Island 2 11 0% No reactor trip 74 Three Mile Island 2 3 0% No reactor trip 75 Three Mile Island 2 1 0% No reactor trip 76 Hatch 1 1 984 No reactor trip 77 Shoreham 1 0% No reactor trip ,

78 Shoreham 1 1% No reactor trip 79 Brunswick 1 1 99% Spike on APRM 80 Sequoyah 2 1 98% No reactor trip 81 Duane Arnold -1 57% No reactor trip 82 Duane Arnold 1 100% No reactor trip 83 Fitzpatrick 1 100% No reactor trip 84 North Anna 2 1.100t RPS trip 85 Davis-Besse 1 1 0% Po reactor trip-86 Davis-Besse 1 1 0% 'No reactor trip 87 Davis-Besse 1 1 100%. Turbine trip ,

88 Farley 1 1 100% RPS trip 89 Farley 2 1 100% High negative flux 90 *Farley 2 l' 99% .High-negative flux 91 Farley 2 1 100%' High negative flux 92 Arkansas Nuclear 2 1 1% No reactor trip 93 Arkansas Nuclear 2 1 100% RPS trip 94 McGuire 1 1 50% No reactor trip .;

95 McGuire 1 2 0% LNo reactor' trip 96 McGuire 1 1 75% No. reactor trip 97 McGuire 1 1 0% No reactor trip 98 McGuire'l 1 0% No reactor trip '

l. 99.McGuire l' 1 50% No reactor trip 100 McGuiro 2 1 0% No. reactor trip 46 l

l s

• e e e . >

101 Watorford 3  ? 0% No reactor trip 102 Waterford 3 1 100% RPS trip 303 Waterford 3 1 26% RPS trip 104 Susquehanna 1 1 100% No reactor trip 105 Susquehanna 1 2 100% No reactor trip 106 Susquehanna 1 1 100% No reactor trip 107 Susquehanna 1 1 97% No reactor trip 108 Susquehanna 1 1 100% RPS trip 109 Susquehanna 1 1 100% RPS trip 110 Susquehanna 1 1 100% Turbine trip 111 Susquehanna 2 1 0% No reactor trip 112 Susquehanna 2 1 100% Turbine trip 113 St. Lucio 2 1 34% No reactor trip 114 St. Lucio 2 1 94% No reactor trip 115 Fammer 1 1 95% No reactor trip 116 Hummer 1 1 100% No reactor trip 117 9ummer 1 1 100% No roactor trip 11b WPPSS 2 1 92% No reactor trip 119 *Lacrosso 1 0% No reactor trip 120 Boavor Valloy 2 1 0% No reactor trip 121 Boavor Valloy 2 1 100% No reactor trip 122 Catawba 1 1 100% No reactor trip 123 Catawba 1 1 0% No reactor trip 124 Grand Gulf 1 1 0% No reactor trip 125 Grand Gulf 1 1 100% No reactor trip 126 Grand Gulf 1 1 04 No reactor trip 127

• Grand Gulf 1 3 100% Spiko on APRM 128 Grand Gulf 1 1 100% Spiko on APRM 129 Grand Gulf 1 1 100% Spiko on APRM 130 Grand Gulf 1 1 100% Spike on A?RM 131 Vogtle 1 1 16% High negativo flux 132 Vogtle 1 1 100% Turbino trip 133 Comancho Peak 1 1 38% High negativo flux 134 Comanche Peak 1 1 100% No reactor trip 135 Comancho Peak 1 1 100% No reactor trip 136 Byron 1 1 78% High negative flux 137

• Byron 1 1 11% High negativo flux 138 Byron 1 2 98% High negativo flux 139 Byron 1 1 98% No reactor trip 140 Byron 1 1 100% No reactor trip 141 Byron 1 1 99% No reactor t?lp 142 Braidwood 1 1 100% High negative flux 143 Braidwood 1 2 86% High negativo flux 144 Braidwood 1 1 0% No reactor trip 145 Braidwood 1 2 0% No reactor trip 146 Braidwood 2 1 90% High negativo flux 147 Riverbend 1 1 91% No reactor trip 148 Riverbond 1 1 21% Turbine trip 149 Wolf Crook 1 0% No reactor trip 150 Wolf Crook 1 92% No reactor trip 151 Palo Vordo 1 1 0% No reactor trip 152 Palo Verde 2 1 100% No reactor trip 47

t

• o o o e

+

• b Inkle 3 - LER_ Data ll Unit System Effects 1 Yankoo Rowo one offsito lino lost 2 Yankoo Rowo one offsite lino lost i 3 Yankoo Rowo  !! oater drain level controllor failuro 4 Yankoo Rowo one offsito lino lost 5 Yankoo Rowe IDOP , firo, lost comm., sec., non-1E UPS l 6 Big Rock Point Line and local striko offects  !

7 !!addam Neck One offsito lino lost j 8 Oyster Crook 1 Voltage trans. - lightning arrestor 9 Nino Milo Point 1 Meteorological tower 10 Nino Milo Point 1 One offsito lino lost 11 *1ndian Point 2 Loop - all lines (shield wire failuro) 12

• Turkey Point 3 Spurious pressurizer pressure signals 13 Turkey Point 3 Spurious pressurizer prosauro signals 14 Quad Cities 1 Voltage transient J 15 Quad Cities 1 Voltage transient i 16 Quad Cities 1 Voltage transient ,

17 Quad Citics 1 Generator protectivo relaying misop.

18 Palisados Poor reporting, exact offect unknown 19 Browns Ferry 1 Motocrological tower 20 Browns Ferry 1 Motoorological tower 21 Browns Ferry 1 Meteorological tower 22 Browns Ferry 1 One offsito lino lost ,

23 Browns Ferry 1 Motoorological tower 24 Browns Ferry 1 Motoorological tower 25 Browns Ferry 1 Meteorological tower 26 Robinson 2 Computer damage attr. to lightning 27 Monticello Poor report, local or lino strike 28 Monticello One offsite lino lost 29 Quad Cities 2 RWCU actuation due to loss of power 30 Vermont Yankoo Meteorological tower 31 Vermont Yankoo Meteorological tower 32 Vermont Yankee Transmission lino relaying failuro 33 Salem 1 .Prossure transmitters damaged 34 Salem 1 Lino flashover at main transformer 35 Salem 1 One offsito lino lost 36 Poach Bottom 2 One offsito line lost 37 Poach Bottom 3 Lino strike, existing failed de solenoid 38 Peach Bottom 3 Block switch failure attr. to lightning 39 Peach Bottom 3 N/A ?

40 Surry 2 Level transmitter failed 41 Indian Point 3 One offsito line lost 42 Pilgrim 1 One offsito lino lost 43 Pilgrim 1 Meteorological tower 44 Zion 1 - IN 85-86 45 Cooper one offsito line lost 46 Cooper One offsito lino lost 47 Cooper One offsito lino lost 48 Cooper One offsite line lost 48

t ' c r e o 49 Cooper One offsite lino lost 50 Cooper One offsito lino lost, diesels started 51 Cooper Turbine DEH malfunction 52 Point Beach 2 Loss of RCPs, bus transfer failed 53 Crystal River 3 Motoorological tower 54 Crystal River 3 Motoorological tower 55

• Crystal River 3 One offsito line lost 56 Crystal River 3 Meteorological tower 57 Crystal River 3 Meteorological tower 58

• Crystal River 3 LOOP - all lines 59 Zion 2 - IN 85-86 60 Zion 2 - IN 85-86 61 Zion 2 RTD damago, spurious fire alarm 62

• Prairie island 2 Two offolte lines lost 63 Maine Yankoo Two offsite lines lost, diosols started 64 Maino Yankee Two offsito linos lost, diesels started 65 Arkansas Nuclear 1 TG control pwr supply problems (Lambda) 66 Arkansas Huclear 1 EHC Card failed during storm 67 Cook 1 One offsito lino lost 68 *Calvert Cliffs 1 Bushing, comp. fail. attr. to lightning 69 Throo Milo Island a Halon system actuated 70 Throo Milo Island 2 Halon system disarmed during storm 71 Throo Milo Island 2 Halon system actuated 72 Three Mile Island 2 Halon system disarmed during storm 73 Threo Milo Island 2 Halon system disarmed during storm 74 Throo Mile Island 2 Halon system disarmed during storm 75 Throo Mile Island 2 Motoorological tower 76 Hatch 1 Motoorological tower 77 Shoreham Voltage transient, damaged circuit cards 78 Shoreham Voltage transient 79 Brunswick 1 RCIC Actuation HBO Sequoyah 2 Switchgear flashover and fire 81 Duane Arnold Voltage transient, rad. monitor alarms 02 Duano Arnold Voltage transient, diesels started 83 Fitzpatrick One offsite line lost 84 North Anna 2 Circuit breaker misoperation 85 Davis-Desno 1 Meteorological tower 86 Davis-Bosso 1 SU transformer lightning arrester failed 87 Davis-Bosse 1 Breaker reclosing caused T-G shaft vib.

88 Farley 1 Start-up transformer trip 89 Farley 2 CRD OV protection 90 *Farley 2 CRD OV protection, loss of all RCPs 91 Farley 2 CRD OV protection 92 Arkansas Nuclear 2 CEAC failed, attr. to lightning 93 Arkansas Nuclear 2 Falso.RCS paramotors, attr. to lightning 94 McGuire 1 Voltage transient, diesels started 95 McGuiro 1 Voltage transient, diesels started 96 McGuiro 1 Flow controller' changed state 97 McGuire 1 Voltage transient, diesels sta'rted 98 McGuire 1 Voltage transient, diccols started 99 McGuiro 1 One offsito line lost 100 McGuire 2 Voltage transient, diesels started 49

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

g

• w*

s ..

6

• y 101 Watorford 3 LOOP - all lines 102 Waterford 3 Station circuit breaker exploded 103 Waterford 3 SG controls failed 2 hrs after lightning 104 Susquehanna 1 One offsito line lost, ESF actuation 105 Susquehanna 1 One offsito line lost, ESF actuation 106 Susquehanna 1 One offsito lino lost, ESF actuation 107 Susquehanna 1 FPS microprocessors falloc - lino, local?

108 Susquehanna 1 One offsite lino lost 109 Susquehanna 1 One offsito lino lost, HPCI, RCIC, ESP 110 Susquehanna 1 Protective relaying failuro n/a Itng 111 Susquehanna 2 One offsito line lost 112 Susquehanna 2 Protectivo relaying failure 113 St. Lucio 2 Logic circuits " scrambled" 114 St. Lucio 2 Loss of annunciators attr. to lightning 115 Summor 1 Ltng tripped diosol during surveillanco 116 Summor 1 Loss of FPS computer for over 7 days 117 Summer 1 One offsite lino lost 118 WPPSS 2 Voltage transient 119

• Lacrosse LOOP - all lines 120 Boavor Valley 2 Local strike offects, damaged mot, eq.

121 Beaver Valley 2 One offsito lino lost 122 Catawba 1 FWST level transmitters failed 123 Catawba 1 Voltage transient, diosols started 124 Grand Gulf 1 One offsito lino lost, diosolo started 125 Grand Gulf 1 One offsito lino lost 126 Grand Gulf 1 Meteorological tower 127

• Grand Gulf 1 2 HPCS channels tripped, RCIC actuation 128 Grand Gulf 1 RCIC Actuation 129 Grand Gulf 1 2 HPCS Channels tripped 130 Grand Gulf 1 none 131 Vogtlo 1 CRD OV protection, FPS computer alarmed 132 Vogtle 1 One offsito lino lost, relay failure 133 Comancho Peak 1 CRD OV protection 134 Comanche Peak 1 Voltage transient 135 Comancho Peak 1 Voltage transient 136 Byron 1 CRD OV protection 137

• Byron 1 CRD power supply " failed,"

133 Byron 1 CRD OV protection 139 Byron i Voltage transient 140 Byron i Volatge transient 141 Byron i Voltage transient 142 Braidwood 1 CRD OV protection 143 Braidwood 1 CRD OV protection 144 Braidwood 1 Voltage transient 145 Braidwood 1 CRD OV protectien, transmitters damaged 146 Braidwood 2 CRD OV protection 147 Riverbond 1 SBGT and annulus mixing auto starts 148 Riverbond 1 TG protective relaying failuro 149 Wolf Creek Voltage transient, entrl room isolation 150 Wolf Crock- Voltage transient, rad. monitor alarmed 151 Palo Verde 1 Voltago transient 152 Palo Verde 2 LOOP - all lines to ESF transformer 50 I

.- - - - _ - _ _ . - _ _ _ - _ _ _ _ _ _ - _ _ -