the cause of rcic shutdown in unit 3

Attachment 3-5-1

Attachment 3-5

The cause of RCIC shutdown in Unit 3

1. Outline of the incident and issues to be examined

Unit 3 experienced a station blackout when it was hit by tsunami on March 11th, 2011, following

the reactor scram, and while the reactor core isolation cooling (RCIC) system had been

automatically shut down due to the high reactor water level signal at 15:25 after the system had

been manually started at 15:05. Even after the tsunami, DC power sources remained available,

and water injection was continued by restarting the RCIC manually again at 16:03. But at 11:36 on

March 12th, the status indicator lamp in the main control room (MCR) confirmed the RCIC had

been automatically shut down. Operators confirmed also in the RCIC room the situation that RCIC

had been shut down and relevant valves were reset on the MCR control board for standby and

restart. But immediately after attempting the restart, the stop valve trip mechanism was unlatched,

the valve closed and the RCIC failed to restart. Although the RCIC continued to function longer

than the designed 4 hours under the station blackout conditions, the cause of the Unit 3 RCIC

shutdown was examined, based on actual plant conditions observed and design information, in

order to interpret the accident progression and to contribute to further safety enhancements by

improvements of relevant equipment and operational procedures.

2. Relevant operational actions and local situations

Table 1 shows the chronological records of relevant operational actions taken and incidents

observed until the time of the Unit 3 RCIC trip at 11:36 on March 12th and that of operational

actions for the restart.

<RCIC startup at 16:03 on March 11th, 2011>

The RCIC was automatically shut down at 15:25 on March 11th, due to a high reactor water level

signal just before the tsunami arrival. It was manually started up at 16:03, as the DC power source

was still available after the tsunami hit. Thus, the main steam relief safety valve (SRV) and RCIC

could continue controlling the reactor pressure and water level of Unit 3.

Under these circumstances, the operators configured the water injection line through both the

RCIC injection line and its test line by valve operations on the RCIC control panel. This was done

to avoid battery depletion due to RCIC startup and shutdown, and to maintain the stable reactor

water level by using the condensate storage tank (CST) as the water source. In order to ensure

moderate transition of the reactor water level, the flow rate was controlled within a predefined

range by adjusting the test line valve apertures and the flow indicator and controller (FIC). Figure 1

shows the RCIC diagram and the operational status.

Attachment 3-5-2

ReactorMO

Main steam piping

Test bypass valve

Turbine

Test line

Water source switch line

Ope

ration line

S/C

FD

W s

yste

m

Containmentvessel

Feed w

ate

r line Minimum flowbypass valve

MO

MO

MOMO

MO

MO

MOAO

MO

MO

MO

HO FIC

FT

Pump

Condensate storage tank

Turbine inlet valve

Turbine steam stop valveControl valve

Flowcontroller

MO

Containment isolation valve

Flow rate meter

PＴ

Suctionpressure

gauge

PＴ

Discharge pressure gauge

PT

PＴ Inlet steam pressure gauge

Exhaust pressure gauge

Condensate pump

Glands of turbine / valves

Barometric condenser

Vacuum tank

Vacuum pump

Lube oil cooler

ReactorMOMO

Main steam piping

Test bypass valve

Turbine

Test line

Water source switch line

Ope

ration line

S/C

FD

W s

yste

m

Containmentvessel

Feed w

ate

r line Minimum flowbypass valve

MOMO

MOMO

MOMOMOMOMOMO

MOMO

MOMOMO

MOMOAO

MOMO

MOMOMO

MOMOMO

HOHOHO FIC

FT

FIC

FTFT

Pump

Condensate storage tank

Turbine inlet valve

Turbine steam stop valveControl valve

Flowcontroller

MOMO

Containment isolation valve

Flow rate meter

PＴPＴ

Suctionpressure

gauge

PＴPＴ

Discharge pressure gauge

PTPT

PＴPＴ Inlet steam pressure gauge

Exhaust pressure gauge

Condensate pump

Glands of turbine / valves

Barometric condenser

Vacuum tank

Vacuum pump

Lube oil cooler

Figure 1 RCIC system diagram

<RCIC shutdown at 11:36 on March 12th, 2011>

The RCIC status indicator lamp on the MCR control panel showed it had been shut down, and

the indicators showed zeros for discharged flow rate and pressure, confirming the RCIC shutdown.

Operators reset the relevant RCIC valves on the RCIC control panel for restart, and then

attempted to restart, but it was followed by immediate shutdown after the restart operations, and

then operators made an on-site check. They entered the RCIC room through the HPCI room. They

noticed that water had flooded the floor as deep as their ankles (about 10 – 20 cm) in both rooms,

but the air did not feel damp. Water drips were falling on the stop valve, etc., from the RCIC room

ceiling, but nothing abnormal was noticed on the turbine, pumps, piping, etc.

As the RCIC was confirmed locally to be out of service and no anomalies were noticed about the

mechanical structures of the steam stop valve, the operators tried again to restart the RCIC from

the MCR control panel by resetting and restarting valve operations, but the steam stop valve

closed immediately after and the RCIC shut down. While operators were struggling for

maneuvering the RCIC for status confirmation and restart actions, the high pressure core injection

system (HPCI) automatically started up due to low reactor water level signal and water injection

started again.

Attachment 3-5-3

Table 1 Chronology of RCIC operational actions and incidents observed at Unit 3

Date Time Incident observed Reference

3/11 14:47 Reactor scrammed automatically (1)

3/11 14:48 Emergency diesel generator (DG) started up automatically (1)

3/11 15:05 RCIC started up manually (1)

3/11 15:25 RCIC shut down automatically (high reactor water level signal) (1)

3/11 15:38 Station blackout (1)

3/11 16:03 RCIC started up manually

Operators configured the water injection line through both the

RCIC injection line and its test line by valve operations on the

RCIC control panel. This was done to avoid battery depletion due

to RCIC startup and shutdown, and to maintain a stable reactor

water level. In order to ensure moderate transition of the reactor

water level, the flow rate was controlled within a predefined range

by adjusting the test line valve apertures and flow indicator and

controller (FIC).

Except for essential systems required for monitoring and control,

all other non-urgent loads were disconnected to save the batteries.

Regarding monitoring instrumentations, only one channel out of

redundant channels A and B was kept connected. Emergency

lightings and clocks in the MCR were disconnected, as well as

fluorescent lights in other rooms.

(1)

3/12 11:36 RCIC shut down automatically

RCIC status indicator lamp showed “shut down.” Zeros on the

RCIC discharged flow rate and pressure indicators were

confirmed.

(1)

to

RCIC restart operations taken

Operators tried to restart the RCIC after the valve reset operations

on the RCIC control panel, which was followed by an immediate

shutdown after restart operations, and then operators were sent for

an on-site check.

The operators entered the RCIC room through the HPCI room.

They noticed water had flooded the floor as deep as their ankles

(about 10 – 20 cm) in both rooms, but the air did not feel damp.

Water drips were falling on the stop valve, etc., from the RCIC

room ceiling, but nothing abnormal was noticed on the turbine,

pumps, piping, etc.

As the RCIC was confirmed locally to be out of service and no

(1)

Attachment 3-5-4

Date Time Incident observed Reference

anomalies were noticed about the mechanical structures of the

steam stop valve, the operators tried again to restart the RCIC

from the MCR control panel by resetting and restarting valve

operations, but the steam stop valve closed immediately and the

RCIC shut down accordingly.

3/12 12:35 HPCI automatically started up due to the signal “low reactor water

level.”

(1)

(1) Fukushima Nuclear Accident Analysis Report, TEPCO, June 20, 2012

[http://www.tepco.co.jp/en/press/corp-com/release/betu12_e/images/120620e0102.pdf]

<Trip mechanism of RCIC stop valve and its resetting procedures (reference information)>

The RCIC turbine can be tripped by two different mechanisms: electrical trip and mechanical

overspeed trip. In an unusual situation, the turbine trip stop valve is closed to cut the steam flow.

Figure 2 shows a schematic diagram of the turbine steam stop valve and trip mechanisms.

A latch mechanism is built in the upper part of the valve where its stem is driven. In the electrical

trip mechanism, the trip signal (an automatic trip signal from the interlock or manual trip signal)

excites the trip solenoid and releases the latch. When unlatched, the spring in the valve cylinder

expands and the steam stop valve is fully closed.

In the mechanical trip mechanism, the tappet in the overspeed trip mechanism is pushed

upward, which works on the rod connected to the latch mechanism by spring forces, leading to the

unlatching. The electrical mechanism needs a DC power source just as the RCIC needs it for

operation and control, while the mechanical mechanism can work even if no DC power source is

available.

Latch mechanism

RCIC turbine steam stop valve

Limit switch

Trip solenoid（electricaltrip）

Motor

Mechanical turbine overspeed trip mechanism

Valve closure bylatch/unlatch mechanism

Spring

Rod

Tappet

Steam Inlet Turbine

Spring

Latch

Traveling nutCoupling

Position indicator lamp

disc

Latch mechanism

RCIC turbine steam stop valve

Limit switch

Trip solenoid（electricaltrip）

Motor

Mechanical turbine overspeed trip mechanism

Valve closure bylatch/unlatch mechanism

Spring

Rod

Tappet

Steam Inlet Turbine

Spring

Latch



disc

Figure 2 Schematics of RCIC turbine steam stop valve and its trip mechanisms

Attachment 3-5-5

Resetting of steam stop valve

In operation UnlatchedTraveling nut

rotated upward

Close turbine inlet valve

Reset steam stop valve

Open turbine inlet valve

Electrical trip

Restart RCIC

Reset

Motoropened position

Valve discrotated upward

Spring

Latch



disc

Discopened position



Motorclosed position

Discopened position

Discclosed position

Discclosed position

Resetting of steam stop valve

In operation UnlatchedTraveling nut

rotated upward

Close turbine inlet valve

Reset steam stop valve

Open turbine inlet valve

Electrical trip

Restart RCIC

Reset


Valve discrotated upward

Spring

Latch



disc

Discopened position



Motorclosed position

Discopened position

Discclosed position

Discclosed position

Figure 3 Schematics of RCIC turbine steam stop valve resetting

3. Conditions of DC power sources

Figure 4 shows schematics of DC power distributions and loads relevant to the RCIC and HPCI

of Unit 3. Batteries and DC power supply panels of Unit 3 were not destroyed by the tsunami,

since they had been installed in a sub-basement of the turbine building, differing from the situation

of Unit 1 and Unit 2.

Figure 5 shows schematics of the in-service condition of DC power loads, while Table 2 gives

the chronology of incidents and power supplies for each load. DC power required by the RCIC

logic circuits and trip signals for electrical trips is supplied from the battery DC125-3A (Power

Distribution Panel 3A-1). This DC power source for the electrical trips is considered to have been

secured, when the RCIC tripped at 11:36 on March 12th, since the RCIC operation and its resetting,

and the reactor water level monitoring had been controlled in the MCR, and no evidence of

depletion of the battery DC125-3A had been noticed.

The capacity of battery DC125-3A (1200 AH in terms of a 10-hour discharge rate) was set in

consideration for the following loads: instantaneous loads such as for starting up emergency diesel

generators (DGs) upon loss of the external power source, for operating circuit breakers of power

supply panels; and standing power loads for four-hour operations of RCIC, MCR control panels,

and DC lighting devices. In reality, however, the RCIC could have been in operation for about 20

hours after the station blackout due to the tsunami had occurred. This is considered as having

been made possible because the battery could have lasted without depletion until this time for the

following reasons: having a design margin (the actual power consumption is less than the

designed value), disconnecting non-essential DC lighting devices, avoiding RCIC trips due to high

reactor water levels, etc.

Attachment 3-5-6

But from the afternoon onward of March 12th, the battery DC125-3A is considered to have been

unstable, indicating signs of depletion in vacuum pump trips, instrumentations becoming defunct,

etc.

Attach

men

t 3-5-7

Figure 4 Schematics of DC power distributions and loads relevant to HPCI of Unit 3 (1/3)

RC

IC conde

nser vacuum pum

p

DC125V battery 3A 60cells 1200AH

（10-hr discharge rate）T/B sub-basement

・・・・

T/B sub-basement 125V main bus line 3A

R/B 1st floor 125V-R/B MCC 3A

RC

IC conde

nser condensate pum

p

RC

IC turbine

steam stop valve

RC

IC control

panel (9-4)

C/B 1st floor 125V distribution panel 3A-1

・・・・

To NO.2/3 To NO.3/3

RC

IC

condensate pum

p’sisolation

valve, R

CIC

valve

position indicators

for control

valve and

steam

stop valve

etc.

Com

puter input

of H

PC

I startup signals at low

reacto r w

ater level etc.

RC

IC control

panel (9-30)

Trip channel panel

of ES

F (I) (9-78)

HP

CI control

panel(9-41)

HP

CI control

panel (9-32)

RC

IC

logiccircuits,

RC

ICturbine

control panel,

RC

ICturbine spee

d control

Reactor w

ater level instrum

entation for ES

F

(reactor water level high / low

switch [for R

CIC

and HP

CI trip

/ start or AD

S actuation])

Trip channel panel

of ES

F (I) (9-88)

Instrumentation for E

SF

(pressure / tem

perature sw

itchfor R

CIC

trip or H

PC

I isolationetc.)

HP

CI isolation logic etc.

DC125V battery 3B 60cells 1400AH

（10-hr discharge rate） T/B sub-basement



ESF: Engineering safety features

Attach

men

t 3-5-8


HP

CI condensate vacuum

pump

・・・・

T/B sub-basement 125V main BUS line 3B

R/B 1st floor 125V-R/B MCC 3B

HP

CI condenser conden

sate pump

HP

CI auxiliary oil pum

p

HP

CI control

panel (9-3)

C/B 1st floor 125V distribution panel 3B-1

・・・・

To NO.1/3

HP

CI

condensate p

ump

isolation valve,

HP

CI

valveposition indicators for controlvalve and steam

stop valveetc.

RC

ICisolation

logic circuits

HP

CI control

panel (9-39)

RC

IC control

panel (9-33)

・・・・

C/B 1st floor 125V distributionpanel 3B-2

HP

CI

logic circuits,

HP

CI

turbine trip, H

PC

I instrumentation

Trip channel panel

of ES

F (II) (9-78)

Reactor

water

levelinstrum

entation for

ES

F(reactor

water

level high

/low

sw

itch [for

RC

IC

andH

PC

I trip

/ start,

AD

Sactivation etc.])

・・・・

C/B basement 125V distribution panel 3B-4

Trip channel panel

of ES

F (II) (9-89)



To NO.3/3





ESF: Engineering safety features

Instrumentation for E

SF

(pressure / tem

perature sw

itchfor R

CIC

trip or H

PC

I isolationetc.)

Attach

men

t 3-5-9


RC

IC system

mortor-driven valves

（except turbine steam stop valve

）

T/B sub-basement 250V main BUS line 3A

R/B 1st floor 250V-R/B MCC 3A

・・・・

RC

IC

instrumentation

(for m

onitoring on

the

RC

IC control panel in the

MC

R)

RC

IC

instrumentation

circuit panel (9-1

9

）

C/B 1st floor Vital AC 120/240V distribution panel

T/B sub-basement 250V main BUS line 3B

・・・・ R/B 1st floor 250V-R/B MCC 3B

・・・・・・・・

Vital AC CVCF source

・・・・ Distribution panel for process computer







To NO.2/3 To NO.1/3

Distribution panel for transient recorder

HP

CI system

mortor-driven valves

Attach

men

t 3-5-10

RCIC

RCIC vacuum pump

RCIC condensate pump

Monitoring indicators

Temperature recorderon the AM panel

Water level indicatoron the AM panel

Water level indicatoron the main panel

HPCI

HPCI vacuum pump

HPCI condensate pump

HPCI auxiliary oil pump

SRV solenoid

March 130:00

Date0:00

March 1114:00

March 1212:00

16:03 RCIC manual startup 11:36 RCIC tripped - restart operation 5:08 RCIC restart operation

3:37 tried to start vacuumpump but failed

12:55 vacuum pumptrip confirmed

after 19:00 chart recorder stopped

20:27 indicator stopped

20:36 indicator stopped (details unknownon the order of water level indicators ChannelsA and B being stopped, etc.)

No recordwhen stopped

4:06 condensatepump power supplyswitched off

3:39 manually shutdown

Reactor pressure maintained at about 7MPa

12:35 HPCI automatic startup 2:42 HPCI manually shutdown

2:45 turning on all switches to open SRVs,but the reactor pressure did not decrease

around 9:00 rapidreactor depressurization

SRV behavior unknown although the reactor pressure was maintained at about 7MPa

Figure 5 Schematics of DC power loads

Attach

men

t 3-5-11

Table 2 Chronology of incidents and power distributions to each load Battery

DC125V-3A Battery

DC125V-3B Battery

DC250V-3A Main bus line panel

Main bus line panel 3A Main bus line panel 3B 3A 3B

Date Time

Incident

MCC 3A

PDP 3A-1

PDP 3A-2

MCC 3B

PDP 3B-1

PDP 3B-2

MCC 3A

MCC3B

Vital

Remarks or presumptions (X shows the power supply

source to related load)

PDP: Power distribution panel

3/11 16:03

RCIC manual startup X X X

― Flow control by FIC while monitoring relevant parameters in MCR X

Unnecessary loads were disconnected from the vital power distribution panel except that for monitoring parameters relevant to RCIC.

RCIC automatically tripped X*1) 3/12 11:36 RCIC flow rate/discharge

pressure indicated 0 X

― RCIC resetting and restart attempted

X X

― RCIC tripped immediately after restart

X*1)

*1) No signs of DC power source depletion confirmed at this point. Voltage for electrical trip was considered to be secured.

3/12 12:35

HPCI automatic startup (low reactor water level)

X X X

to 3/12 about20:00

Monitoring indicators ・ PCV temperature chart

recorder on AM panel ・ D/W pressure, S/C

pressure and S/C water level indicators on AM panel

・ Reactor water level indicator (wide range, fuel range)

X X

As of 3/12 11:36, all indicators were working and no signs of DC power source depletion were confirmed. After 3/12 20:00, signs of DC power source depletion were noticed such as out-of-service indicators. DC power did not seem to have been stable.

Attachment 3-5-12

4. Examination of the cause of repeated shutdown

4.1 Possibility of electrical trips

It is highly possible that the RCIC was not tripped by mechanical overspeed, but electrically at

11:36 on March 12th, as seen in the afore-mentioned monitoring and operating situations in MCR

and field observations, based on the following grounds.

The turbine steam stop valve equipped with the trip mechanisms had been closed, tripping

the RCIC;

The operational procedures of turbine steam stop valve had been completed to the

resetting procedure from the MCR panel; and

DC power for RCIC operation and control had been available.

Figure 6 shows the interlock logic of RCIC electrical trips.

Figure 6 The interlock logic of RCIC electrical trips

The following subsections describe the feasibility conditions of each interlock logic, among

which the highest likelihood is for high turbine exhaust pressure.

Manual trip button

High reactor water level (L-8；5655mm above TAF）

High turbine exhaust pressure （0.29MPa[gage]）

Low pump suction pressure （－0.0508MPa[gage]）

Turbine trip （Steam stop valve

closed）

RCIC automatic isolation signal isolates the

RCIC by detecting steam leakage or exhaust

rupture disk damage.

Turbine overspeed （electrical trip： rated×110% rpm)

OR

OR

High RCIC turbine pump room temperature （93℃）

High flow rate in steam piping （±120.7kPa)

Low steam supply pressure （0.344MPa）

High pressure between the turbine exhaust rupture diaphragms

（69kPa)

OR

RCIC automatic isolation signal

Attachment 3-5-13

4.2 Possibility of high turbine exhaust pressure interlock working

Figure 7 shows the RCIC turbine exhaust pressure changes observed by the operators. Due to

loss of heat removal functions, the exhaust pressure was increasing with the increase in both the

D/W and S/C pressures, and as of 11:36 on March 12th, when the RCIC tripped, they were at a

level very close to the preset high exhaust pressure trip level. The exhaust pressure reading was

0.25MPa[gage] as of 11:25 on March 12th, and had not reached the preset trip level

(0.29MPa[gage]). Further, there is no record on the exhaust pressure at the very instant of the

RCIC trip, and so no direct evidence of observation exists concerning the RCIC trip due to high

exhaust pressure.

0.0

0.1

0.2

0.3

0.4

3/1118:00

3/1121:00

3/120:00

3/123:00

3/126:00

3/129:00

3/1212:00

3/1215:00

3/1218:00

Exh

aust

pre

ssur

e (M

Pa[

gage

])

0.101

0.201

0.301

0.401

0.501

PC

V p

ress

ure （

MP

a[ab

s])

RCIC exhaust pressure(measured)HPCI exhaust pressure（measured）D/W pressure(measured)S/C pressure（measured）

Figure 7 RCIC exhaust pressure change and PCV pressure change

It should be noted, however, that the D/W and S/C pressures showed an increasing trend again,

after remaining stable at one level for almost 3 hours before the RCIC trip, in the particular

circumstance in which the PCV pressure should have simply increased under the situation of no

final heat sink. Consequently, it is possible to consider that the turbine exhaust pressure reached

the trip level at 11:36 on March 12th, based on the following.

(1) Instrumentation accuracies

As is shown in Figure 8, exhaust pressure gauges for the MCR indicator and the logic circuit

input are different, and about 0.02MPa difference could exist in these two instrumentations due to

instrumentation accuracies and reading errors. Even taking these inaccuracies into account, the

3/12 11:36RCIC trip

3/12 12:35 HPCI automatic startup

RCIC exhaust pressure Trip setpoint 0.29MPa[gage]

S/C spray ON

3/12 11:25Exhaust pressure

reading 0.25MPa[gage]

Attachment 3-5-14

pressure reading of 0.25MPa[gage] at 11:25 on March 12th is considered not to have reached the

trip level (0.29MPa[gage]).

S/C

MO

MO

HO

Turbine inlet valve

Turbine steam stop valve

Control valve

PT

PT

MCR display

Input tologic circuit

About ±0.02MPa differences can exist when indicator and reading accuracies are considered

Condensatepump

Glands of turbine or valves

Barometriccondenser

Vacuumtank

Vacuumpump

S/C

MOMOMO

MOMOMO

HOHOHO

Turbine inlet valve

Turbine steam stop valve

Control valve

PTPT

PTPT

MCR display

Input tologic circuit

About ±0.02MPa differences can exist when indicator and reading accuracies are considered

Condensatepump

Glands of turbine or valves

Barometriccondenser

Vacuumtank

Vacuumpump

Figure 8 Positional illustration of RCIC exhaust pressure gauges

In the meantime, the D/W and S/C pressures increased from 11:25 through 11:36. According to

the RCIC test operation that had been done during the scheduled maintenance, the turbine

exhaust pressure under the reactor rated power condition (about 6.9MPa[gage]) was about

0.05MPa when the S/C pressure was atmospheric pressure. This represents the pressure loss in

the turbine exhaust piping under rated power operation. The S/C pressure measured from 11:25

through 11:36 was about 0.36MPa[abs], and when 0.05MPa is added as the pressure loss, it

reaches 0.41MPa[abs] (about 0.31MPa[gage]), exceeding the trip level.

It is considered, therefore, the turbine exhaust pressure might have reached the trip level when

the D/W and S/C pressures increased from 11:25 through 11:36, although the exhaust pressure, at

11:25 on March 12th had not reached the trip level, but was close to it (0.29MPa[gage]).

(2) Explanation of the successful RCIC turbine resetting operation

As was shown in Figure 7, the turbine exhaust pressure reading was about 0MPa[gage] after

the RCIC trip at 11:36 on March 12th. This is because the steam in the exhaust piping was

condensed and because the S/C pressure (back pressure) was blocked by the check valve in the

downstream side of the exhaust piping.

Concerning the incident of repeated trips of the RCIC upon steam injection in restarting attempts

after successful resetting, it can be interpreted by the following scenario: If a trip is triggered by the

high turbine exhaust pressure, the exhaust pressure on the indicator decreases, the trip condition

is cleared once, and later upon steam injection the RCIC trips again due to the increased

pressure.

Attachment 3-5-15

On the other hand, the scenario is still unclear concerning the incident of D/W and S/C pressure

changes, i.e., monotonously increasing upon loss of heat removal functions with no final heat sink,

then staying once at a level from 09:30 through 11:25 on March 12th, and again starting to increase.

At that time, the reactor steam was being released to the S/C via the SRV and RCIC, and the

water inside the S/C was likely to have thermally stratified. The incident might have been caused

by some unknown phenomena in the S/C.

4.3 Possibility of other interlocks working

Other interlock conditions are unlikely to be met, as discussed below.

<Manual trip>

No traces of this were confirmed, which indicates the manual scram button had been pushed.

<High reactor water level (L-8: TAF+5655mm)>

Figure 9 shows the chart record of reactor water level (narrow range) and Figure 10 shows the

measured water level changes which operators read in the indicators. The RCIC flow rate was

being controlled by the FIC while monitoring the reactor water level, and as of 11:36 on March 12th,

when the RCIC tripped, the reactor water level had not reached the level L-8. An RCIC trip due to

high reactor water level interlock is not possible.

<Low pump suction pressure (-0.0508MPa[gage])>

This could occur if the condensate storage tank water level decreased. But immediately after,

the HPCI could have been started up and its operation continued (using the same trip signals).

Further, if the tank water level really dropped below the preset trip level, the RCIC resetting

operation would not have been successful. These are inconsistent with the facts as they are

known.

<High turbine overspeed (electrical trip: rated x 110% rpm)>

It is unlikely that the overspeed trip occurred, since the DC power for RCIC control was available,

and the steam control valve was functional while the operator was monitoring the reactor water

level and controlling RCIC flow rate using FIC.

<Automatic isolation signal>

No records exist showing the status indicator lamp indicated the RCIC isolation valve closure.

Further, the PCV isolation valve on the outer side was closed from the MCR at around 11:00 on

March 13th.

Attachment 3-5-16

Red: Reactor water level (narrow range)

Green: Reactor pressure (wide range)

Figure 9 Chart record of reactor water level/pressure

-10

-8

-6

-4

-2

0

2

4

6

8

10

3/1118:00

3/120:00

3/126:00

3/1212:00

3/1218:00

3/130:00

3/136:00

3/1312:00

Rea

ctor

wat

er le

vel (

m)

Wide range water level(measured)

Fuel range water level(measured)

Figure 10 Reactor water level changes

5. Related countermeasures

In the development of tsunami accident management procedures under station blackout

conditions for the Kashiwazaki-Kariwa Nuclear Power Station, a procedure to bypass the high

RCIC turbine exhaust pressure interlock is being developed to continue water injection using the

RCIC even when the S/C pressure increases. This is based on a concept to prioritize continuing

the RCIC operation for flexibility even if the turbine exhaust pressure exceeds the trip level, since

such an operation will not damage the RCIC equipment immediately.

The RCIC exhaust piping has a vent line equipped with a rupture disc for preventing

3/12 11:36RCIC tripped

3/12 12:35 HPCI automatic startup

3/13 2:42 HPCI manual shutdown

3/12

12:

00

3/13

00:

00

L8 setpoint

3/12 11:36 RCIC tripped

L8 setpoint

TAF

BAF

Attachment 3-5-17

overpressurization. This is because the RCIC is designed so that its pressure is released when the

RCIC rupture disc physically ruptures, if the RCIC turbine exhaust pressure continues to increase

as the S/C pressure increases, and reaches the rupture disc working pressure. Practically,

however, the S/C pressure is controlled, since the PCV is cooled by means of the PCV spray or

PCV venting by alternative water injection lines (the pressure limit for PCV venting is set lower

than the RCIC rupture disc working pressure).

There are other countermeasures, too, for strengthening high pressure water injection functions,

and some examples are listed below.

Reinforced DC power source (capacity intensification, elevated positioning).

Additional installation of a high pressure alternative cooling system (HPAC).

Development of local RCIC manual startup operating procedures upon loss of all AC and

DC power sources.

6. Summary

It is highly possible that the RCIC was not tripped by mechanical overspeed, but was electrically

tripped at 11:36 on March 12th, based on analysis of the monitoring and operating situations in the

MCR and field observations, for the following reasons.

The turbine steam stop valve, equipped with trip mechanisms, had been closed, tripping

the RCIC.

The operational procedure of the turbine steam stop valve had been completed to the

resetting procedure from MCR panel.

DC power for RCIC operation and control had been available.

Among possible interlock conditions for electrical trips, the one with the highest likelihood was

for the high turbine exhaust pressure. As of 11:25 on March 12th, the turbine exhaust pressure

reading was 0.25MPa[gage], and below the preset trip level (0.29MPa[gage]). But the D/W

pressure and S/C pressure were increasing after remaining stable at one level before the RCIC

trip, and it was possible, from the following, that the turbine exhaust pressure reached the trip level

at 11:36 on March 12th, the time of RCIC trip.

When the pressure loss in the RCIC exhaust piping was added to the S/C pressure reading,

the turbine exhaust pressure would be at the level exceeding the trip level. Therefore, the

turbine exhaust pressure had to be close to the trip level.

It was possible that the turbine exhaust pressure reached the trip level when the S/C

pressure increased from 11:25 through 11:36 on March 12th.

It was possible to interpret the incident development as that the RCIC resetting could be

done (the trip condition was cleared once), but it tripped immediately afterward upon steam

release again.

Attachment 3-5-18

Other interlock conditions were unlikely to be met.

(End)

the cause of rcic shutdown in unit 3

Documents