IR 05000259/1980032
| ML19331D367 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry  |
| Issue date: | 08/08/1980 |
| From: | Dance H NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML19331D366 | List: |
| References | |
| 50-259-80-32, 50-260-80-25, 50-296-80-26, NUDOCS 8009020225 | |
| Download: ML19331D367 (18) | |
Text
.
.
- e UNITED STATES l ',,,
' 't, NUCLEAR REGULATORY COMMISSION U
I REGION 11
.g 101 MARIETTA ST., N.W., SUITE 3100 e
O,
%r o[
ATLANTA, GEORGIA 30303
- ...*
Report Nos. 50-259/80-32, 50-260/80-25 and 50-296/80-26
.
I,icensee : Tennessee Valley Authority
,
500A Chestnut Street
.
Chattanooga, TN 37401 Facility Name: Browns Ferry
'
Docket Nos. 50-259, 50-260 and 50-296 Inspection at Browns Ferry Site near Athens, Alabama Inspected by: See Section 1 of the Report Approved by:
b-
"
D H. C. Jance, Section thief, RONS Branch Dat'e Signed SUMMARY Inspection on June 28 - July 4, 1980 Areas Inspected
.
This special, announced inspection involved approximately 300 inspector-hours on site in the areas of inspection and evaluation of the licensees investigation of the Unit 3 failure to fully insert control rods following a scram on June 28, 1980.
Results k
Of the area inspected, no items of noncompliance or deviations were identified.
.
.
... _. _
_ _ _, _ _ _,
- ~ ~ ~ ~ ---- - -
-
~~
-
_ _ _ _ _
,
- - _.
.- - - -, -..
-
--- -
- - - - - - - - - - _ _
_
.
.
.
.
DETAILS 1.
Persons Contacted Licensee Employees
.
- H. L. Abercrombie, Plant Superintendent
.
- J. L. Harness, Assistant Plant Superintendent
.
J. B. Studdard, Operations Supervisor E. Edmondson, Lead Electrical Engineer J. A. Teague, Maintenance Supervisor, Electrical M. A. Haney, Maintenance Supervisor, Mechanical J. R. Pittman, Maintenance Supervisor, Instruments R. G. Cockrell, Reactor Engineer B. E. Baggett, Shift Engineer, SRO J. D. Glover, Shift Engineer, SRO L. L. Kennedy, Shift Engineer, SRO M. Gant, Assistant Shift Engineer, SRO V. Johnson, Unit Operator, R0 R. Champion, Unit Operator, RO
-
E. Nave, Shift Technical Advisor R. T. Smith, Quality Assurance Supervisor L. Parvin, Quality Assurance Staff Other licensee employees contacted included numerous technicians, operators, mechanics, security force members and office personnel.
NRC Personnel On Site Individual Dates
- J.
P. O'Reilly, Director, Region II 7/2-4/80 H. C. Dance, Chief, Reactor Projects Section No. 1, R.gion II 6/30- 7/3/80
- J. Chase, Resident Inspector, Browns Ferry 6/28-7/4/80 C. Julian, Reactor Inspector, Region II 6/28-29/80 B. Moon, Reactor Inspector, Region II 7/1-3/80 K. Roberts, Senior Resident Inspector, IE 7/2 - 3/80
- R. Sullivan, Senior Resident Inspector, Browns Ferry 6/28-7/3/80 W. Ruhlman, Reactor Inspector, Region II 7/2 - 4/80 T. Ippolito, Chief, Operating Reactors Branch No. 3, NRR 7/2 - 3/80 W. Minners, Technical Assistant to Director of Safety Technology, NRR 7/2 - 4/80 F. Clemonson, Systematic Evaluation Program, NRR 7/2 - 4/80 S. Rubin, Office of Analysis and Evaluation of Operational Data, NRR 7/2 - 3/80
'
W. Mills, Senior Specialist, IE:HQ 7/2 - 4/80
- Attended exit interview
'
i
,
_.
_ _ _ _....... -
.
... _.
_..... _..
... _... _... _.......-....
......
...
.
_
-.
_
.. _..
.. _
.
. _..
>
.
.
.
-2-
.
2.
Exit Interview The inspection scope and findings were summarized daily throughout the in-spection with those persons indicated in paragraph I above.
3.
Licensee Action on Previous Inspection Findings
Not inspected.
-
4.
Unresolved Items Unresolved items were not identified during this inspection.
5.
Summary Of Event On June 28, 1980, Browns Ferry Unit 3 began a normal shutdown to repair a leak in the 3B reactor feed pump discharge piping. Following established procedure, reactor power was decreased by lowering recirculation loop flows. Power was further reduced by fully inserting 10 control rods in a symmetric pattern.
From a reactor power of approximately 400 MWe, the reactor was shutdown by initiating a mans 1 scram at 0131 hours0.00152 days <br />0.0364 hours <br />2.166005e-4 weeks <br />4.98455e-5 months <br />. In re-sponse to the scram signal, all rods on the west side of the core fully inserted, but 76 rods on the east side of the core failed to fully insert, halting at various positions ranging from notch 46 to notch 2.
Six of these rods stopped at notch 4 or less and thus were effectively fully inserted from a reactivity standpoint. Nuclear instrumentation indicated that the reactor was shutdown and suberitical overall. An OD-8 process computer printout of the Local Power Range Monitor (LPRM) incore detectors was run at 0132 hours0.00153 days <br />0.0367 hours <br />2.18254e-4 weeks <br />5.0226e-5 months <br /> and demonstrated no areas of local supercriticality or excessive power generation. The recirculation pumps continued in opera-tion providing sufficient core flow to remove incore heat generation. All indications at this time were that the situation was a controllable one and
)
no prompt drastic emergency action was required.
The problem appeared to be hydraulic rather than electrical in nature. All indications were that all rods had received a normal electrical scram signal.
Operators observed that all rods displayed a blue " scram" light on the reactor control panel. These lights are energized only when both the scram water in-let and outlet valves are open for a given rod. Thus it appeared that both the inlet and outlet valves had opened admitting scram drive watar below the control rod drive (CRD) piston and allowing above piston water to be exhausted from the CRD mechanism.
Additionally, rods in the west side of the core had performed normally. Rods on the east and west side of the core are controlled by two separate banks of hydraulic control units (HCU) physically located outside the drywell, on the east and west of the reactor respectively. Water exhausted from above
-
the CRD piston collects in two separate scram discharge volume headers (SDVH)
located above the east and west HCU banks. Thus the probable failure mode was an inability to exhaust water from the east HCU bank into the east serra discharge volume headers.
f
..._
__
.
- -.... ~..
.;.
_
.
.
-3-
.
After approximately four and one-half minutes operators reset the manual scram.
Af ter 94 seconds, the reactor was manually scrammed again. All partially withdrawn rcds were observed to drive inward and then halt resul-ting in 59 rods remaining partially withdrawn.
In approximately one minute the scram was reset again'.
?"
a3 seconds the reactor was manually scrammed a third time. Again athdrawn rods moved in and halted with 47 rods partially withdrawn.
saly 35 of these rods.eere further than position 04 withdrawn, however.
Approximately three and one-half minutes later, the scram was reset again in preparation for another manual scram. After 160 seconds, the reactor received an automatic scram from high Scram Discharge Volume (SDV) water level. This scram was initiated by the reactor operator moving the SDV high water level scram bypass switch from the " bypass" to the " normal" position.
The switch had been placed in " bypass" as a prerequisite for the scram resets.
'
In response to this scram signal all the remaining rods inserted. The normal shutdown procedures were then resumed and investigations began.
.
During the efforts to correct the problem, the operators attempted to manually drive the control rods in.
This proved impossible as the Rod Sequence Control System (RSCS) would not allow a rod to be selected. The RSCS is designed to enforce a prescribed sequence of rod withdrawal when reactor power is less than 307,.
In this case, with numerous rods partially inserted and out of the normal insert sequence, the RSCS acts to prevent selection of any control rod for withdrawal or insertion.
Table I contains a sequence of events derived from the process computer,
interviews with the control room operators and plant staff.
6.
NRC Actions
.
Following this event the NRC initiated several actions.
a.
The Senior Resident Inspector arrived onsite within 2\\ hours of the event and a regional reactor core physics specialist arrived onsite the morning of the event.
A total of 13 NRC personnel under the direction of the Region II Director were onsite the following week to review the TVA investigetive efforts and to assess the event.
b.
Region II issued a Confirmation of Action letter to TVA on June 30, 1980, confirming TVA's investigative plan and an NRC review prior to restart. These actions were verified to have been completed, c.
Preliminary Notification PNO-II-80-119, Failure of Control Rods to
.
Insert During a Scram, was issued on June 30, 1980, alerting all NRC
'
offices to the event.
Other BWRs were informed of the event. An update of this PNO was issued on July 3,1980.
- - - -.. - -. - -
-
- -
- -
_.. -. -.
........ -. - -. -.. - --
_
. - - - -
.-.
.-
.
_
.
. -..
--
.
.
.
.
.
TABLE I BROWNS FERRY 3
-'
.
SEQUENCE OF EVENTS
,
FAILURE TO FULLY INSERT CONTROL RODS FOLLOWING A SCRAM ON 6/28/80 Hour Min. Sec.
Event Comment
31
Manual Reactor Scram Channel B (half scram)
First Manual Scram Manual Reactor Scram Channel A (full scram)
24 Reactor Trip Actuator, Channel A Auto Scram Signal from low reactor Reactor Trip Actuator, Channel B Water level after the scram (shrink)
Reactor Low Water level Channels, A, B, C & D Reactor Feedpump Turbine C Tripped Manual action
34 Scram Discharge Volume High Level Channel C Scram Discharge Volume High Level Channel D
37 Scram Discharge Volume High Level Channel B
40 Scram Discharge Volume High Level Channel A All SDV scram level swit.ches actuate
.
.
O
..... -, - ~ ~. - -. - - - - - - + = * --, -$.-
,
-,--e
-, --
r
- * =
-
,
. _, _
.
--
}
,9 w
.,- - - - ---
.
.
.
-2-
.
Hour Min. Sec.
Event Comment (Continued)
...
31
Turbine Malfunction Bus Energized
-
Turbine Stop Valve Closure Scram Channel A
.
Turbine Stop Valve Closure Scram Channel C Turbine Stop Valve Closure Scram Channel D Turbine Stop Valve Closure Scram Channel B Turbine Load Rejection Scram Channel B Turbine Load Rejection Scram Channel D Turbine Load Rejection Scram Channel A Turbine Load Rejection Scram Channel C Turbine Tripped Turbine tripped
.
---
manually
01 Reactor Low Water Level Channel A Clear Reactor Low Water Level Channel C Clear Reactor vessel level Reactor Low Water Level Channel D Clear recovered following Reactor Low Water Level Channel B Clear shrink
45 IRM Channel F High Level Trip
48 IRM Channel D Hign Level Trip Alarms caused IRM Channel D High Level Clear by operator down IRM Channel D High Level Trip ranging Intermedi-
,
IRM Channel F High Level Clear ate Range Monitor.
IRM Channel B High Level Trip (IRM) Channels
... -
-..
-.. _.
... - - -.
..
. -.
.
-
-Q
-
-
.w m
mmv-l-Q a
?
-
-
w
.
.
3-
-
..
Hour Min. Sec.
Event Comment (Continued)
34
IRM Channel B High Level Clear
.
IRM Channel B High Level Trip
,
IRM Channel B High Level Clear IRM Channel B High Level Trip IRM Channel B High Level Clear IRM Channel B High Level Trip IRM Channel D High Level Clear IRM Channel B High Level Clear
43 Manual Reactor Scram Channel A Clear Operator resets scram first time Manual Reactor Scram Channel B Clear
.
46 Reactor Trip Actuator Channel A Clear auto scram signal Reactor Trip Actuator Channel B Clear reset
59 Scram Discharge Volume High Level Channel B Clear Scram Discharge Volume High Level Channel A Clear
12 IRM Channel D High Level Trip Auto half scram Reactor Trip Actuator Channel B Trip from IRM High IRM Channel D High Level Clear with companion IRM Channel D High Level Trip APRM downscale
17 IRM Channel D High Level Clear
45 Reactor Trip Actuator Channel B Clear
,
'
l l
'
,
l l
..- ____
_
..
--...... - - --
--
e-
--- -
e up-
-
.,
,- - -, - -,
,y-- - -,--,
-y,
--
y
--
-.--q--,
y
---
,,9
-, - -
,y-
,.
---y-
,m-
.
.
.
.
-4-
.
Hour Min. Sec.
Event Comment (Continued)
.
37
Manual Reactor Scram Channel B Operator initiates Mamaal Reactor Scram Channel A second Ma'nual scram Scram Discharge Volume High Level Channel A Trip Scram Discharge Volume High Level Channel B Trip
33 Scram Discharge Volume High Level Channel B Clear Scram Discharge Volume High Level Channel A Clear Scram Discharge Volume High Level Channel A Trip Scram Discharge Volume High Level Channel B Trip
19 Manual Reactor Scram Channel A Clears Operator resets scram Manual Reactor Scram Channel B Clears second time
32 Scram Discharge Volume High Level Channel B Clears Scram Discharge Volume High Level Channel A Clears
40 Scram Discharge Volume High Level Channel A Trip
42 Scram Discharge Volume High Level Channel B Trip l
12 Manual Reactor Scram Channel A Operator initiates Manual Reactor Scram Channel B third Manual Scram
22 IRM Channel F High Level Trip Reactor Trip Actuator Channel B Trips IRM Channel F High Level Clear
00 Reactor Trip Actuatoc Channel A Trips
.
_;
- - - -
- --) _ ; _.-.--
7 - - : - ;;;
-
- - - :- L :
, _ _ _ _ -.
. -
_
_
.
.
.
'
-5-
.
Hour Min. Sec.
Event Comment (Ccntinued)
...
42
Reactor Trip Actuator Channel A Clears Reactor Trip Actuator Channel B Clears Operator rests scram Manual Reactor Scram Channel B Clears third time Manual Reactor Scram Channel A Clears
17 Reactor Trip Actuator Channel A Trips Fourth reactor scram Reactor Trip Actuator Channel B Trips from SDV High Level scram out of bypass, all rods insert
36 Scram Discharge Volume High Level Channel B Clears Scram Discharge Volume High Level Channel B Trips
30 Manual Reactor Scram Channel B Operator inputs fol-Manual Reactor Scram Channel A lowup Manual scram
43 Reactor Trip Actuator Channel A Clears Reactor Trip Actuator Channel B Clears Operator resets final Manual Reactor Scram Channel A Clears scram per procedure Manual Reactor Scram Channel B Clears
04 Scram Discharge Volume High Level Channel B Clears Scram Discharge Volume High Level Channel A Clears
34 Scram Discharge Volume High Level Channel D Clears Scram Discharge Volume High Level Channel C Clears i
l i
.. _ _...
._
...
..
...
. -...
-
.
.
_
_
_
._-
.__
_
_..
.
.
-4-
.
d.
IE Bulletin No. 80-17 was issued on July 3, 1980 tv -.. BWR licensees specifying requirements to verify that the SDVHs were empty with vent and drain piping unobstructed, plant procedures addressed this concern, special training be conducted, and to demonstrate system operability by conducting a manual and an automatic scram.
Region II issued a confirmation of concurrence let er to TVA en July 14, e.
1980, confirming plans to continuously monitor the SDVH, install an atmospheric vent on the east SDVH similar to that located on the west side, provide instructions to operating personnel that specify actions to be taken when water is detected in the SDV, and to expedite review of modifications required to the SDV, including additional vents on Units 1 and 2.
7.
Reactor Physics Investigations An NRC reactor physics specialist from Region II arrived at the site the morning of the event. The inspector interviewed cognizant Reactor Engineers and plant personnel including all licensed operators present during the event.
All available strip chart recorders reflecting plant parameters during the event were examined. Process computer printoe+= of core parameters prior to and during the event were reviewed.
.
Indications are that the first manual scram rendered the core subcritical.
There is no evidence of a return to criticality during the event. There is no indication of local area power generation in excess of available core cooling capability.
Incore neutron detectors indicated that local criti-cality was possible.
A core physics computer code study, modeling the control rod position following the first scram, will be necessary to deter-mine if areas of local criticality actually existed. Licensee representatives stated that their preliminary conclusions were as follows:
There was not an accurate estimate of Keff following the scram but the a.
minimum shutdown margin of 0.0038 AK/K, required by Technical Specifi-cation 4.3. A.I. was not immediately available.
b.
Reactor coolant samples taken showed normal activity levels following a scram. There was no evidence of fuel damage as a result of this event.
Based on intermediate range monitor nuclear instrumentation following c.
the first scram, heat generation rates in the fuel on the east side of the core were less than one percent of full power heat generation rates.
The inspector found no evidence to dispute these conclusions. The failure to have the required shutdown margin was properly reported to the NRC. No.
deviations or items of noncompliance were identified in this area.
-
.
_-_u..
..--... -.... -.-..
..
....
--
.....
.-
-. -
_
.
.
.
-5-
'
,
8.
Procedures and Operator Actions Review of precedures listed below were made by the inspector to determine adequacy and to see thtt they were followed in this event:
30I-101,Section V. c.. Shutdown by Manual Scram OI 85,Section IV, Abnormal Operation (Control Rod D'ive System)
r 0I-63, Standby Liquid Centrol System
.
Iba procedures were determined to be adequate but did not specifically cover symptons in this event.
Temporary instructions were provided to address the issues on an interim basis. Subsequently, E0I-47, Failure of Control Rods to Fully Insert During Scram, was written to identify the control rod and standby liquid control system restrictions contained in IEB 80-17.
Operator response during this event was judged by the inspectors to be most satisfactory. Rod positions were obtained after each of the scram actions.
Deliberate actions and understanding of the system were demonstrated.
No deviations or items of noncompliance were identified in this area.
9.
Mechanical System Review The apparent cause of the control rods on the east side of the reactor to insert is the east scram discharge volume header (SDVH) was partially filled with water prior to the scram. This is also supported by the SDV Hi level (scram) switch which activated in 18 seconds after *.he initial scram.
Review of 12 previous scrams indicated normal times of 42-56 seconds.
Figure 1 is a schematic of the system. The design intent is that the SDV system remains empty, with vent and drain valves open, during normal opera-tion. Upon a scram, the vent and drain valves close, and the scram discharge water from above the CRD piston empties rapidly into both SDVH chambers.
With the east SDVH partially filled, the pressure quickly equalized to that of the CRD scrag inlet water, and the CRD's stopped motion and settled into the nearest notch.
Upon reset of the scram signal, the scram outlet valves from the CRD's close, the vent and drain valves of the SDV open, and water begins to drain from the SDV providing free volume. When the reactor was scrammed again, scram discharge water again flowed into the SDV and all partially withdrawn rods drove in.
When pressure equalized above and below the CRD pistons, the rods not fully inserted again stopped and settled into the nearest notch. By repeating this process of reset and scram, three times, the rods were finally able to fully insert. The periods of reset, during which the SDV was allowed to drain, were 94, 53, and 160 seconds. The amount of control rod insertion after each scram signal roughly corresponds to the relative length of drain times. The fact that the final scram which fully inserted the remaining rods was initiated from the automatic circuitry is apparently not significant. At that point, the SDV had been allowed to drain for 160 seconds and a manual scram at that point would have accomp-lished the same results.
.. - - -. - - - - _
__
. _.. - -. - - - - -.. -.. - - - - - - -
sir, e me e
8e
.
--
- -. - -
,,, _,
.,-
,-
w.
,
y
_ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ -
_
.
. _ _
,
Figure
-
.
,
.
St. RAM DISCHMGE VOLUME
,
onre: 7/,ufso
,.i l l
,,
6" SCRAM DISC HAnsa unAntns.
re.J
!.'!
!
~
,
'
[~
l *'
's
--
i,]
i a
K
'
.i c
-
CRW j II la
~ 7, ta 3 alls 588EllllllIII N UED'S
\\,
i
,
e a,aaaaaaaeaaaaAIIIII y
- l 50 '
-
' '
,
- _ _ _. _ _
.
l I i !
krest i l I
C 64 0 W
-
s StamM
- A l'
Host Cow urrow
[ ren semam"usamen
!,;
'n
-
.;
OUTtt T A
rvp 93
{
r..
. c.,.
,
si t re, y,,,,,
}
-f
H E s.orn j
f
'I
'e
s
',;,ggyg
/ 91 CRD's l
)
,,,,,,,,,_
,
y, r
j h
- : y g g y ii l
y (h
~ f na
{
,
[
A
[,
_l ll ls ti e n a s a a f i a e i la
___
,!._'I
,[
I I b
y is
!
'
=
"
i
-
'
,
"
L wis,
~ ~ ~
is s
I j
e2 r,sa
~
l
1,,
vo
.
a f
xus Q
'
'
-
Oofit r 8"
tru E
n e. 92.
' -
,,
m
-
!
CLEAN RAD,
3 Ii Vle.WE cnw
i!!W LorP)
.
.
.
.
-
>35D o'
I
.
. __
.
.
.
.
.
6-
-
.
The exact cause for the east SDVH being partially full is not known. The licensee performed many tests and inspections in accordance with approved procedures to determine the cause in the days following the events. Many
)
of these evaluations were witnessed by NRC inspectors and the results of
,
the tests were evaluated by the NRC. Table II lists these efforts.
Only two anomalies were found as a result of the tests and inspections.
a.
When a vacuum pump was first connected to the east vent line on the drainage side, a vaccum of 8" of mercury was pulled for a short period which dropped sharply to 2*'.
The reason for this temporary vacuum was not determined and the transient vacuum was not observed during repeti-tion of the test.
b.
The low level (3-Gallons) and rod block (25-Gallons) level switches did not activate during the first calibration fill after the scram.
Operator observations did indicate that these two switches worked at least during part of the event. The third protective level, 50-Gallons level scram, composed of four separate switches worked properly. IE Bulletin 80-14, Degradation of BWR Scram Discharge Volume Capability, addresses the operability of these switches.
,
It was hypothesized that. the water in the west SDVH prior to the scram either failed to drain after the prior unit 3 scram on June 7,1980, was forced in through the vert line from the clean rad waste (CRW) vent and drain h::dcr-eystcs, er accumuleted free ler.eage L ne CRB8s-through one--
or more leaking scram outlet valves. No abnormality in the CRW was observed between June 7 and 28 of a magnitude to force water up into the SDV system.
If such an abonormality had occurred, the water should have emptied back out of the SDV via the drain line. Leakage of thermally hot water from the CRD's through the scram outlet valves would have caused a high temperature alarm in the control room on the leaking CRD and no such alarm was observed.
Potential blockage ir %e 2 inch drain line connecting the east SDVH with the instrument volume and the drain valve was investigated. This piping was cut at six locations and inspection yielded no evidence of blockage.
The east SDVH was flushed and inspected with a boroscope from the cut made in the 2" drain line. The instrument volume was also inspected with a boroscope. No anomalies were noted.
Tests indicate that a failure of the vent valve to open or a blockage of the vent line to the CRW system indeed inhibits draining of the east SDVH.
Except for the transient vacuum observed during the first test, no evidence of a plug exists. Tests and inspection after disassembly proved the vent valve to be normal. Subsequently, five cuts were made in the east SDVH vent piping. The piping was clear. A 1.3 cfm vacuum pump attached to the vent pipe was unable to establish a vacuum after fifteen minutes.
.
.
No deviations or items of noncompliance were identified in this area.
'
D**]O
"
D) S X]li%
T M
oaN oN
._..--.
- - - -
.. -........ - - -....... -. -.. -.. -.
. -.
.. -
-
+7-.
.
-
,-m
- -..,
wy-w
-,
-
--
-, +
.w g-v
-
.
.
-
.
TABLE II BR0kNS FERRY UNIT 3 INVESTIGATIONS AND TESTS CO."PLETLD
.
.
1.
VERIFIED CORRECT HYDRAULIC C0hTROL UNIT VALVE ALIGh5.NT
-
2.
EAST BANE VEST VALVE VERIFIED OPERABLE 3.
FRICTION TESTF' ;5 RODS - PERFOR.".ED NORMAL 4.
VERIFIED CALIBRATION OF 3-GALLON, 25-GALLON, AND 50-GALLOL LEVEL SWITCHES ON INSTRLT.NT VOLUME 5.
COMPLETED RADIATION SURVEY OF DRAIN LINES TO DETERMIhT. IF HOT SPOTS EXIST INDICATING BLOCKAGE
'
-
COMPLETED RADIATION SURVEY OF #3 EQUIPMEhT DPAIN SUMP 7.
' ? PLED #3 EQUIPMENT DRAIN SUMP A
8.
SAMPE REACTOR C00LAhT SYSTEM - FOUND NOR"AL 9.
VERIFIED THAT OFF GAS RADIATION LEVELS LT.RE NOR".AL 10.
COMPLETED VISUAL Ah3 MECFANICAL INSPECTIONS OF VEhTS AhT DRAINS IN SCRAM DISCHA*GE VOLUME 11.
VERIFIED THAT NO MAIhTENANCE OR MODIFICATIONS IN PROGRESS OR RECESTLY PERFORMF.D THAT WOULO AFFECT C0hTROL ROD DRIVES 12.
REVIEWED SCRAM HISTORY FOR PREVIOUS FAILURES 13.
PERFORMED PRESSURE, FLOW AND DRAIN TIME TESTING ON EAST AND WEST HEAEERS 14.
GE ENGINEERS PERFOR.".C EXTENSIVE EVALUATIONS AND INSPECTIONS.
15.
CLT Ah3 INSPECTF.D 2-INCH LINE Ah'D VENT HEADER PIPING WITH BOROSCOPE INSPECTION OF 6-INCH HEADERS Ah3 SCRAM DISCHARGE INSTRUMEhT VOLUME
.
.
.
16.
GE ISSUED TWO SERVICE IhTOR".ATION BULLETINS
.
.
.... -. - -
. --.-- --
_ _ -. -. - - -
.-
.. -. :.... -,.
.--
.
.
,
.
-2-t 17.
GE PERFORMED TEST ON CONTROL R0D DRIVE Ah3 PRESSURE VESSEL MOCKUP IN SAN JOSE 18.
PERFORMED TESTS TO CONTIRM ADEQUACY OF ULTRASONIC TESTING TO EASURE WATER LEVEL IN SCRAM DISCHARGE VOLUME 6-INCH HEADERS
,
19.
PERFORMED DRAIN TEST TO DEMONSTRATE THAT TE SYSTEM VILL DRAIN IN A PREDICT-
.
ABLE MANNER FROM A NORMAL ALIGh3EhT 20.
PERFORMED VACUUM HOLD TEST TO DEMONSTRATE THAT A BLOCKED VEhT PATH WILL RESTRICT DRAINAGE OF TE SIX-INCH SCRAM DISCHARGE PIPING 21.
FRICTION TEST - DEMONSTRATE NORMAL INSERT - WITHDRAWAL OPERATION OF THE
,
DRIVES IN TE EAST BANK 22.
SCRAM TESTING
.
A.
FULL SCRAM TEST AT RATED CONDITIONS FROM ZERO POSITION TO VERIFY PROPER OPERATION OF ELECTRIC COMPONENTS AND HYDRAULIC CONTROL UNITS B.
INDIVIDUAL ROD SCRAMS AT VARIOUS CONDITIONS FROM POSITION 48 TO VERIFY SCRAM CAPABILITY WITHIN TECH SPECS TIMES (EAST BANK RODS)
.
23.
PERFORMED SCRAM TIME TEST OF FIVE RODS THAT FULLY INSERTED AND FIVE THAT DIDN'T ON THE EAST BANK - NO DIFFERENCE 24.
CHECKED FOR SCRAM DISCHARGE VALVE LEAKAGE EVERY 200 POUNDS WHILE INCREASING REACTOR PRESSURE 25.
GE RECOMMENDED RESTART OF REACTOR TO PERFORM SCRAM TESTS 26. TVA ON-SITE SAFETY REVIEW COMMITTEE RECOMMEhTED RESTART
- 27.
TVA NUCLEAR SAFETY REVIEW COMMITTEE RECOMMEhTED RESTART
.
.
28.
NRC ISSUED IE BULLETIN 80-17 29.
LICENSEE PERFORMED ACTIONS OF IE BULLETIN 80-17 30. NRC ISSUED CONTIRMATION OF CONCURRENCE LETTER TO ALLOW RETURN *;0 P0kT.R
.-..- -
-
-. -
. _; -...
.......... -... -
.......
.
_
,
_
_
_
._._
_
-
. _..
,
_
.
.
.
-7-
.
10.
Electrical System Review NRC inspectors on site reviewed the licensee's efforts and performed indepen-dent inspections to determine if the cause of the failure to fully insert control rods was of electrical origin.
The inspector reviewed the referenced Reactor Prote tion System (RPS)
drawings and inspected the installation to determine whether possi.bilities of a common mode failure exist within the system. In addition, the inspector traced some of the licensee's post incident trouble shooting efforts.
References:
Browns Ferry Unit 3 Reactor Protection System Drawing, 730E915 sh 8,
-
11, 13, 15 and 16.
Browns Ferry Unit 3 Manual Control System Drawing, 730E321 sh 4.
-
Browns Ferry Unit 3 CRD Hydraulic System, 47W616-85.
-
-
Browns Ferry Unit 1, 2, and 3 Sequential Event Recording Wiring Diagrams, 45N621.
Browns Ferry Unit 3, Sequence of Event Recording on 6/28/80.
-
Findings:
a.
Scram Valve / Vent and Drain Valve Control - The RPS logic for this control is designed such that loss of instrument air to the hydraulic Control Modules (HCU) will result in a scram.
Instrument air is needed to keep the inlet and outlet scram valves closed. A full scram removes air from these valves, causing them to open. Each scram valve has a position switch associated with it to show that it actually changed position. Both inlet and outlet scram valves for each control rod must be in the open position to illuminate a blue scram indicator light for the individual control rod. The licensee stated that their operators verified illumination of all blue lights following the first manual scram. This confirms that all the scram valves responded to the manual scram signal.
During a scram, the scram pilot valves de-energize and block the flow of instrument air to the scram valves. The s: ram pilot valves are designed fail-safe since a loss of power to the solenoids, due to an initiation of a scram (manual er automatic), causes a loss of instru-ment air. The inspector verified from the sequence of events that the loss of power occurred during the first manual scram. Subsequently, automatic scrams occurred from reactor process instrument signals such,
as Lo Reactor Water Level and Scram Discharge Instrument Volume High,
.
Water Level. These scrams as well as the initial manual scram positively interrupted power to the solenoids in the HCUs located in both the east and west division. The referenced drawing showed that this power
.-
- - _ - -
. -. -.. -. -...... -.
.
-...-.... - -.
.....
....
--
..
.
.
_
_
_
..
.
-8-
.
.
interruption of a given CRD group should occur simultaneously at both the east and west locations because a common scram contactor supplies power to both of the solenoids. The drawing also shows that 4 CRD electrical groups are dispersed evenly throughout the core in both the east and west divisions. This confirms that the scram pilot valves assigned to both divisions de-energized following the first manual scram.
-
The inspector verified from the sequence of event parameters'such as Discharge Volume High Water Level and Group 1 Scram Contactors, that RPS scram discharge volume vent and drain valves operated throughout the entire event, providing relief in the dich trge volume.
b.
Layout Inspection - The inspector examined, on July 2,1980, 8 fuse boxes (4 in each division) and 2 scram contactor cabinets located in the Auxiliary Instrument Room, for any physical degradation of compon-ents or layout modifications. There was no apparent sign of degrada-tions or alternations.
c.
Scram Relay Response Times - The inspector reviewed a test record of SMI 150, RPS Scram Solenoid Relay Response Times, conducted on 6-28-80, right after the event. This confirmed that the scram contactor relays were operated (de-engerized) within specified time limits.
s deviations or items of noncompliance were identified in this area.
11.
GE Tests General Electric Company conducted tests in San Jose in days following the event to try to simulate responses observed at Browns Ferry. Tests were witnessed by a Region V inspector.
The tests demonstrated on the one control rod prototype that 0.9 gallons were discharged from a scram stroke, 1.6 gallons per minute per drive continued into the SDV following a scram, and that a control rod would stop at an intermediate position if the space available in the SDV were limited.
Initial data on varying CRD leakage rates were inconclusive. These results appeared consistent with conditions observed during this event.
12.
Additional Information a.
Unit 2 tripped midmorning on June 28 due to loss of condenser vacuum.
Control rod response was normal.
b.
Three additional GE engineers arrived onsite on July 1 to assist TVA in the investigation.
c.
The NRC task force was provided a detailed briefing of th'e event,
,
investigations to date and status on July 2,1980.
..
.. _..
...
-_
.
- -
. _ _
=
_
.
.... -
..... -.. -.
_
.
' '
_
~T
.
_
-
.____ -.
~
e
-..
.
-9-
-
.
13.
Summary The most likely cause of this event was due to water accumulation in the scram discharge volume east header before the scram. The exact cause of this accumulation was not definitely established. There is good reason to think that the cause was a failure of the east scram discharge volume header to drain following the previous scram or from other inleakage that failed to drain. This appears to have been caused by a blockage of either the 2" drain line or the vent line to the clean rad waste systeme. There is no evidence of fuel damage as a result of this event. There wis no abnormal release of radioactivity during this event. The event served to point up several apparent design deficiencies in the SDV system at Br owns Ferry.
These items have been referred to IE Headquarters and NRR for ev.sluation.
Items identified included the following:
The east and west FW headers do not have positive redundant vents to a.
atmosphere. They are vented to the Reactor Building Equipment Drain Tank via a large closed volume of piping which receives water drainage from many sources throughout the plant. Thus venting of the SDVH may be impaired by the actions of other systems.
.
b.
The six level switches on the SDV system failed to detect water levels in the east SDVH which prevented rods on the east side of the core from fully inserting.
These switches are all located on a single instrument volume which is remote from the east SDVH. Improved moni-toring is necessary. An ultrasonic technique of monitoring the SDVH
-
reiiued during che inspectiou. Ver1Trcacion by this method that---
was Unit 1 and 2 SDVH were empty was first made on July 1,1980.
c.
Elect:ically the control rods are divided into four groups dispersed throughout the core. This event demonstrates that due to the hydraulic piping all the rods on one side of the core could be commonly affectew.
d.
During this event, the Rod Sequence Control System prevented the oper-ators from selecting a control rod to manually insert.
{
l
.
.
www e ese + eene en em me.se we, en eue.
w w.e - e *
.==a e
_
- = =
=
-*+e ee
++e==*we=*
e==ree= * a e = ee a
.
er*
e gg,,
,,
.go.
-
.e e m
- w.
p 4e ee e
e mi
e
mm.-
e-e w
,--, -
-, -, - - - - -, - -,, -,
mn,mo,_w
-
,,, - -
.. _ -,, - - - -
.-e--w
--
,w