Text

LER 96-003-01:on 960115,alternate Shutdown Panel Inverter Resulted in Unavailability of Panel.Replaced Defective Inverter Alarm Logic Board
	ML18065A832
Person / Time
Site:	Palisades
Issue date:	08/01/1996
From:	Mathews C; CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.)
To:	;
Shared Package
ML18065A830	List: ML18065A829; ML18065A832;
References
	LER-96-003, LER-96-3, NUDOCS 9608070158
	Download: ML18065A832 (8)
	v • d • e

. .

NRC FORM 366 (4195)

U.S. NUCLEAR REGULATORY COMMISSION . APPROVED BY OMB NO. 3150-0104 EXPIRES 4/30/98 ESTlllAlEl IUIOEN PER RESPONSE TO COllPL Y WITH TMIS MANDATORY INFORMATION COUECTION REQUEST: 50.0 HRS. REPORTED LESSONS LEARNED ARE INCORPORATED LICENSEE EVENT REPORT (LER) IHTO niE LICENSING PROCESS ANO FED BACK TO INDUSTI!Y. FORWARD COllllEHTS REGARDING BURDEN ESTIMATE TO TME INFORMATION AND RECORDS llANAGEllEHT BRANCH (T~ F33), U.S. llJCLEAR REGULATORY COlllllSSION, WASHINGTON. DC 2055$.

. 0001, ANO TO niE PAPERWORK REDUCTION PROJECT (3156--0llW, OFFICE OF (See reverse for required number of digits/characters for each block) llANAGEllEHT AND BUDGET, WASHINGTON. DC 20503 FACILITY NAME (1) DOCKET NUMBER (2) Page (3)

PALISADES NUCLEAR PLANT 05000255 1of8 TITLE (4) LICENSEE EVENT REP.ORT 96-003-01, ALTERNATE SHUTDOWN PANEL INVERTER FAILURE RESULTS IN UNAVAILABILITY OF PANEL- SUPPLEMENTAL REPORT EVENT DATE (5) LER NUMBER (6) REPORT DATE (7) OTHER FACILITIES INVOLVED (8)

MONTH DAY YEAR YEAR I SEQUENTIAL NUMBER REVISION NUMBER MONTH DAY YEAR FACILITY NAME DOCKET NUMBER

'05000 FACILITY NAME DOCKET NUMBER 01' 15 96 96 - 003 - 01 08 01 96 05000 OPERATING THIS REPORT IS SUBMITTED PURSUANT TO THE REQUIREMENTS OF 10 CFR§: (Check one or more) (11)

MODE (9) N 20.2201(b) 20.2203(a)(2)(v) 50. 73(a)(2)(1) 50. 73(a)(2)(iii)

POWER 20.2203(a)(1) 20.2203(a)(3)(1) 50.73(a)(2)(ii) 50.73(a)(2)(x)

LEVEL (10) 100 20.2203(a)(2)(1) 20.2203(a)(3)(ii) 50.73(a)(2)(iii) . 73.71 .

r NAME

. : : ti 20.2203(a)(2)(ii) 20.2203(a)(2)(iii) 20.2203(a)(2)(iv)

Clayton M. Mathews, Licensing Engineer 20.2203(a)(4) 50.36(c)(1) 50.36(c)(2)

LICE.NSEE CONTACT FOR THIS LER (12)

50. 73(a)(2)(iv) 50.73(a)(2)(v) 50.73(a)(2)(viil

TELEPHONE NUMBER (Include Area Code) x Voluntary Report Specify in Abstract below or in NRC Form 366A

.. (616) 764-2035 COMPLETE ONE LINE FOR EACH COMPONENT FAILURE DESCRIBED IN THIS REPORT (13) '.'

CAUSE SYSTEM COMPONENT MANUFACTURER REPORTABLE CAUSE . SYSTEM* COMPONENT MANUFAqTURER REPORTABLE TONPRDS ' TONPRDS A El INVT X999 y

.: .. ~

SUPPLEMENTAL REPORT EXPECTED 114\ MONTH DAY YEAR I YES If yes COMPLETE EXPECTED COMPLETION DATE

,. x I NO. EXPECTED SUBMISSION DATE (15)

ABSTRACT (Limit to 1400 spaces, i.e., approximately 15 single-spaced typewritten lines) (16) ,.

On January 15, 1~96, at 1330 hours0.0154 days 0.369 hours 0.0022 weeks 5.06065e-4 months , with the plant operating at 100% power, it was discovered I that on September 27, 1995, technicians . found '~hat the low voltage *cut-off for the Alternate I Shutdown Panel (ASP) inverter appeared to be set at 120 volts DC input. This condition, if it had I actually existed, would nave prevented the ASP from operating. Subsequent investigation has I*

shown that the actual cause of the inverter cutting off at 120VDC was ~failed inverter alarm circuit I board. The apparent 120VDC setting would be likely to show up on a failed inverter alarm board, I but is not indicative of the setting which existed before the board failure. During the September I 1995 repair, the failed inverter alarm board was replaced within the allowed LCO period, the new I board reset at 100.3VDC arJd the ASP returned to operable condition. When this condition was I originally reported under 10 CFR 50.72 on February 15, 1996, it was assumed that the measured I voltage cut-off at t20VDC was indicative of the previous cut-off setpoint and that the ASP could I

'have been inoperable for an undetermined time periqd. Since it is now known that the' observed I, low voltage cut~off value after the board failure is* not indicative of the low voltage cut-off set point, I it is logical to assume the ASP was operable prior to the September 1995 failure. Since the ASP is I not .normally energized, it is logical to conclude that the ASP was inoperable only during the five I day repair period. I 9608070158 960801 PDR ADOCK 05000255 S PDR

NRC FORM 366a U.S. NUCLEAR REGULATORY COMMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION 6l FACILITY NAME t1 \

l DOCKET12l LER NUMBER PAGE 131 YEAR SEQUENTIAL REVISION NUMBER NUMBER 2 OF 8 PALISADES NUCLEAR PLANT 05000255 96 - 003 - 01 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

EVENT DESCRIPTION * ,

On September 27, 1995, th~ plant was operating at 1-00% power and Technical Specification Test I Q0-23, "Auxiliary Hot Shutdown Panel Checks" was in progress. During the power-up sequence I for the Alternate Shutdown Panel (ASP), the DC input breaker to the panel tripped open. I Repeated attempts to power the ASP provided the same results: Q0-23 was aborted, the panel I declared inoperable, the appropriate LCO entered and an investigation into this failure to transfer I

.control to the ASP was initiated. The investigation revealed that the ASP inverter alarm board I

[EJ; INVT] had failed and needed to be replaced. I I

During this trouble shooting, technicians found the low voltage cut-off for the ASP inverter I appeared to be set at 120 volts DC input. Tl:iis condition,. if it had actually .existed, would have I prevented the ASP from operating when required by fires* in selected areas. The 125VDC battery I supply could not provide 120VDC to the ASP when the battery chargers.are deenergized. I.

Subsequent investigation has shown th~t the actual ca~se of the inverter cutting off at 120VDC ,..:

was a.failed inverter,alarm circuit*board. The apparent.120VDC setting on the tailed alarm boar.d, I according to the. inverter supplier, would be likely to show on a failed inverter alarm board, b.ut. is I not indicative of tl')e setting which existed before the board failure. * .. ,

I I

When this condition was originally reported on February 15, 1996, itwas assumed . thatthe I I measured voltage cut-off at 120VDC was indicative of the previous cut:-off setpoint and that the ASP could have been inoperable for an undetermined time period. Since it is riow known that the observed low voltage cut-off value after the board failure is not indicative of the low voltage cut-off set point, and that the inverter manufacturer sets the low voltage cut-off setpoint at 105VDC, it is logical to assume the ASP was operable from the time it was installed until a failure caused the September 1995 repair~ Since the ASP is not normally energized, except for testing dn a quarterly frequency, it is logical to conclude that the* ASP was inoperable only during the five day repair

period starting when the ASP was energized for testing and lastirig until _the repair was completed.

This time frame was within the seven-day aliowed outage time sp~cified in the Tech.nical Specifications.

CAUSE OF THE EVENT

1. lnit(al reviews determined that the low voltage cut-off setpoint for*:the ASP had not been I identified as a variable whose setting is necessary to assure operability of the panel. No I specific setting or tolerance was apparently specified during the initial panel installation and I checkout, and verifying this setpoint has not been a part of the periodic testing of the panel. I Inadequate post modification and subsequent surveillance testing is one root cause of the event.

NRC FORM 366a U.S. NUCLEAR REGULATORY COMMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT, CONTINUATION I

FACILITY NAME l1l DOCKET12\ LER NUMBER 6\ PAGEl3\

YEAR SEQUENTIAL REVISION

NUMBER NUMBER 05000255 . 3 OF 8 PALISADES NUCLEAR PLANT 96 - 003 - 01 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

2. When the repair work order was reviewed, there was a failure to recognize that the low voltage setpoint could be a safety significant issue. The low voltage condition was not addressed as part of the evaluation of the ~eficiencies repaired under the work order.

3. A contributing cause is the lack of a thorough evaluation of industry experience information.

As part of the evaluation of this low voltage event, it was discovered that a 1985 Industry Experience and Assessment document identified that a particular manufacturer's inverter had a low shutoff voltage problem. A more questioning attitude, with regards to this industry

experience report, could have led to the realization that the low voltage shut-off. set point *.

could be significant to inver:ter operation, and that we didn't monitor the low voltage cut-off setpoint for the ASP.

ANALYSIS OF THE EVENT 10 CFR 50.48, Appendix R requires alternate shutdown.capability tha~ is independent of normal I control stations to provided. This altemate shutdown station would be used to maintain the plant r in hot shutdown in the event that the control room could not be used as a result of a fire. The postulated fire. may make the control room uninhabitable or disable enough of the control room equipment to render it not useful to control the plant. At Palisades, this alternate shutdown .I capability is provided, in part, by our ASP (C-150). Th'e fires.that we have identified in this* I scenario result in postulated consequences that require.,using this ASP. During this postulated I scenario, the ASP is being powered from the station batteri~s and the battery chargers are maintaining a high voltage on the battery system: _If the postulated fir~ also renders both batt~ry chargers inoperable, then the ASP would be fed directly from the station batteries at a somewhat lower voltage. With the ASP inverter low voltage cut-off setpoint set to a voltage higher than what the* station batteries alone can provide, the ASP could not be energized per procedure.

The failure of the ASP to energize was discovered on September 27, *1995 during Technical Specification Surveillance Test Q0-23 "Auxil.iary Hot Shutdown Panel Check~".

A plant condition report was written to ,document this condition and troubleshooting was initiated.

At 2143 hours0.0248 days 0.595 hours 0.00354 weeks 8.154115e-4 months on September 27, 1995, the.ASP was declared inoperable and the plant was placed in a 7 day LCO per .the requirements of Technical Specification 3. 7. 5. Two separate issues were identified that could cause the panel to cutout at 120VDC: .

NRC FORM 366a U.S. NUCLEAR REGULATORY COMMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION I

FACILITY NAME l1 l DOCKETl2l LER NUMBER 16\ PAGE13l YEAR SEQUENTIAL REVISION NUMBER NUMBER PALISADES NUCLEAR PLANT 05000255 4 OF 8 96 - 003 - 01 TEXT (If more space is required, use additional copies of NRC Fann 366A) (17)

1. Component(s) within the inverter alarm undervoltage board failed causing the under voltage circuitry to spuriously activate even when the incoming DC voltage was held above the low voltage cut-off setpoint.

2. The low voltage cut-off setpoint could have been set higher than intended.

The following further explains these two issues:

ISSUE# 1 Two circuit boards make up the inverter control logic, the inverter logic board and the alarm logic board. The inverter logic board sets the frequency of the output waveform, regulates voltage, and controls inverter turn-on/turn-off. In addition, the inverter logic board supplies

+15VDC control power and +6.2VDC reference voltage to the alarm logic board. The alarm logic board is responsible for lighting lamps and controlling the status of alarm contacts to annunciate conditions existing within the inverter'. In addition to its alarm functions, the alarm.

logic board is also responsible for monitoring DC input voltage to the inverter and will initiate a low voltage shutdown signal to the inverter logic board. upon reaching the DC input voltage low voltage cut-off setpoint (adjustable on the alarm logic board).

l&C Technicians, in conjunction with a representative from the inverter manufacturer, and systems engineers were successful in troubleshooting the inverter failure down to com*ponent(s) on the inverter alarm board. The alarm board was signaling a low voltage cut-off shutdown to the inverter logic board even with the DC input voltage being held constant and well above the low voltage cut-off setpoint.

ISSUE# 2 During the initial troubleshooting of the inverter, it was discovered that the inverter low voltage cut-off was 120 VDC input. The low voltage cut.:.off of the replacement alarm logic board was set to the minimum acceptable setting of 100.3 VDC. The evaluation of the setting discrepancy was left for review under the plant condition report.

Our reviews also determined that a 1985 Industry Experience and Assessment, General Electric Service Information Letter NO. 418, "Topaz Inverter Low Voltage Shutoff", chronicles a similar low voltage cut-off scenario for a Topaz inverter. It was identified that the review at the time only addressed the manufacturer of the inverter. Since we had no Topaz inverters on-site, no further reviews for this issue were recommended. This can be viewed as a missed opportunity to identify that the low voltage cut-off problem could have generic implications, which might have led to the discovery that the low voltage cut-off setpoint for the ASP was not being properly monitored.

NRC FORM 3S6a U.S. NUCLEAR REGULATORY COMMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION I

FACILITY NAME l1l DOCKETf2\ LER NUMBER 16\ PAGE 13\

YEAR SEQUENTIAL REVISION NUMBER NUMBER 50F8 PALISADES NUCLEAR PLANT 05000255 96 003 01 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

SUBSEQUENT INVESTIGATION I I

Subsequent investigation (after the submfssion of the original Licensee Event Report) led to the I following additional information regarding the details of this event: I I

VALUE OF THE LOW VOLTAGE CUT-OFF SETTING I I

According to the ASP (C-150) inverter supplier, "The low voltage cutoff set point adjustment would have been factory preset and tested to the value specified in the (vendor) manual, 105 VDC, prior to shipment." *

The c.-150 panel fabricator received and installed the inverter without te~ting of adjustment.

Palisades Nuclear Plant installed the co,mpleted C-150. panel.without testing or adjusting the set point. * * * .

After'the installation of the panel, no record exists of changing or adjusting the setting. Nor does any record exist of previous damage or failure of the inverter.

Therefore, the reasonable conclusion has been reached that the C-150 panel low voltage cut-off set point was correctly set by the supplier in accordance with the vendor manual and remained at .105 voe setting.

READINGS OF LOW VOLTAGE CUTOFF FROM FAILED BOARD The actual cause of the C-150 panel failure was .the failure of its inverter alarm logic board.

This board was replaced as a unit.

Input versus output readings were taken on the failed inverter and seemed to indicate a low voltage cut-off at 120 VDC. However, the inverter vendor has stated, "A determination of the original low voltage cut-off setpoint cannot be made, while adjusting the input voltage to the inverter and measuring the input voltage level at which. the inverter shuts down, if the alarm logic board installed is defective." Therefore, the measured inverter cut-off voltage does not provide a reliable indication of the true low voltage cut-off set point.

INOPERABLE INVERTER TIME The inverter was noted as last being operable during an August 10, 1995, surveillance. The inverter was not po.wared again until the September 27, .1995, surveillance, when the inverter

i .

NRC FORM 366a U.S. NUCLEAR REGULATORY C<;>MMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION I

FACILITY NAME 11 \ DOCKET12l LER NUMBER I 6\ PAGE 13\

YEAR SEQUENTIAL REVISION NUMBER NUMBER 60F8 PALISADES NUCLEAR PLANT 05000255 96 003 01 TEXT*(lf more space is required, use additional cop'ies of NRC Form 366A) (17) alarm logic board failure was detected. Therefore, our conclusion is that the alarm logic board failed when the C-150 panel was powered for the September 27, 1995,_ surveillance. The inverter wa$ declared operable after repair on October 2, 1995, which me~ns that the inverter was out of service for less than five days (and within the time frame of a decl_ared seven-day LCO). ..

. SUBSEQUENT INVESTIGATION CONCLUSION The C-150 panel would have performed as intended upon loss of the battery chargers except during the less than five-day out-of-service time of the failed board.

SAFETY SIGNIFICANCE Since the C-150 panel was operable except for the brief period during repair, the actual satety .I significance of the failure was minimal. Even if the panel had been ino'perable for an extended period, I the following discussion shows the safety significance would." still not have been great. I

. . I If the low voltage cut-off for the ASP inverter had been set at approximately 120 volts DC and a I known initial battery terminal voltage of less than 120 volts DC was experienced, it is reasonable to I conclude that the panel would not"have operated. Therefore, in the event of a fire in the Electrical Equipment Room, 1-D Switchgear Room, Cable Spreading Room, or the Auxiliary Buildfr,g*59'c)*

corridor concurrent with a loss of offsite power and the loss*of the station battery chargers, the ASP would not have been available. For a fire in *any of thes.e areas, both the existing and new .Appendix R analysis credit the ASP as being available. *

The areas where a single fire may cause the loss of both battery chargers are Appendix R alternate shut(jown areas. The Safe Shutdown Analysis is required by regulation to assume loss of offsite

a power du"ring the fire. In addition, the Appendix R fire is postulated to non-mechanistically affect all safe shutdown components in the area, even if the fire loading is not sufficient to cause such damage. Realistically, there is reasonable assurance that a fire in these areas would not progress to a state where safe shutdown equipment would be damaged. Any fire would be controlled by the installed automatic suppression system or. fire brigade manual suppression such that either a loss of offsite power would not occur or that only one battery charger would be lost. Thus, the probability of a realistic fire causing both a loss of offsite power and a loss of both battery chargers in these areas is very*low. The specific features for Fire Area 2 - Cable Spreading Room; Fire Area_3 D Switchgear Room; Fire Zone 13A -Auxiliary Building 590'-0" Corridor;*a,nd Fire Area 21 - Electrical Equipment Room are addressed *below.

i . e .

NRC FORM 366a NUCL~

, , 4195 U.S. REGULATORY COMMISSION LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION I

FACILITY NAME 11 \ DOCKETl2l LER NUMBER 6l PAGE 131 YEAR SEQUENTIAL REVISION NUMBER NUMBER 7 OF 8 PALISADES NUCLEAR PLANT 05_000255 96 003 01 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

Suppression Actual Fire Design Spray Suppression Fir,e fi[e A[ea LQadiog Oeosit~ Sc[a~ Oeosit~ oetei;;tiQo Cable Spreading Room - FA 2 1.39 Hr 0.30 gpm/ft 2 *0.60 gpm/ft 2 Yes 1-D Switchgear Room - FA 3 51 Min 0.20 gpm/ft 2 0.40 gpm/ft 2

Yes Aux Bldg 590' C9rridor - 33 Min None in area

N/A Partial FZ 13A Of concern .. Area Elec Equip Room - FA _21 7 Min 0.30 gpm/ft2 .0.38 gpm/ft2 ... Partial Area ,

From the above information, th~ two areas w_ith a higher fire loading are the Cable Spreading '

Room and 1-D Sw'itchgear Room. The combustibl~ loa_ding in both of these areas is composed of electrical cabling and electrical panels. However, both areas have automatic fire.suppression provided over the cabling with an actual spray density of twice the design requirements. Also,.

separate automatic smoke detection is provided for these tw6 areas that alarms in the Control Room. The postulated real\stic fire in these areas.1would be a localized cable fire that should be quickly extinguished or controlled by the automatic suppression with twice the design spray density. The suppres~i6n spray would likely limit the fire size such that the redundant cables or battery chargers would not both be 'lost due to a single_ fire. *

The Electrical Equipment Room has minimal fire loading and has a greater than 1O feet clear space between the redundant components and circuits located in this room. The automatic .

suppression*and the partial detectiqn is located directly over the in-situ combustibles. There is minimal likelihood of a realistic fire affecting redundant circuits in this room due to the physical separation and positioning of automatic suppression directly over the i:!reas of concern. ..

. ~.

NRC FORM 366a U.S. NUCLEAR REGULATORY COMMISSION 4195 LICENSEE EVENT REPORT (LER)

TEXT CONTINUATION I

FACILITY NAME 11 l DOCKETl2l LER NUMBER 6l PAGE 13l YEAR SEQUENTIAL REVISION NUMBER NUMBER PALISADES NUCLEAR ~LANT 05000255 8 OF 8 96 003 01 TEXT (If more space is required, use additional copies of NRC Form 366A) (17)

While the non-mechanistic Appendix R fire is postulated to damage the redundant (;ircuits or components in the alternate shutdown areas described above, it is highly unlikely that a realistic fire. could cause such damage. In addition, the .areas with a higher fire loading have automatic:

suppression systems with as-built water spray densities of twice that required by the design specifications.

In addition, the evaluation has shown that even with an inoperable C-150 panel, the plant would I still be able to reach and maintain a stable hot shutdown condition in the event of *an Appendix R I fire. Sufficient time and local control options exist so that all required safety functions could be I preserved. ' . ** I CORRECTIVE ACTIONS CORRECTIVE ACTIONS TAKEN'

1. The defective inverter alarm logic board;was replaced and the low.voltage cut-off set point I was adjusted to the minimum acceptable setting of 100.3 V. These a,ctions corrected the I identified problem. *

2. Developed.a procedure *to include periodic testing and. adjustment of the* low volt_age cut-off setpoint for the ASP inverter.

3. Determined whether other similar equipment setp'oints exist which can cause inverters or -.

battery chargers to shut dowr:i, and verified that appropriate* testing is performed~. j

ML18065A832

Text

Navigation menu

Search