IR 05000327/1993010
| ML20035D411 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 04/06/1993 |
| From: | Blake J, Gibson A NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML20035D401 | List: |
| References | |
| 50-327-93-10, 50-328-93-10, NUDOCS 9304130183 | |
| Download: ML20035D411 (33) | |
Text
.
...
.-
-.
',:
,
, - d * "' 8 %,
c UNITED STATES f,,
NUCLEAR REGULATORY COMMISSION
+,
o REGION ll
{
' $-
101 MARIETT A STREET.N.W.
g
,
%,4*++,8[
ATLANTA, GEORGIA 30323 o
i
?
Report Nos.:
50-327/93-10 and 50-328/93-10
.
Licensee: Tennessee Valley Authority i
3B Lookout Place 1101 Market Street i
Chattanooga, TN 37402-2801
Docket Nos.:
50-327 and 50-328 License Nos.:
DPR-77 and DPR-79 Facility Name: Sequoyah 1 and 2 Inspection Conducted: March 3 - 11, 1993 AIT Team Leader:
A M
'romeJ.yake, Date signed ectionCpiefMPS,DRS,RII AIT Team Members:
Billy Crowley, Engineering Inspector, DRS, RII
'
Dave LaBarge, LPM for Sequoyah, NRR Peter Vsang, Electr.;al Engineer, NRR Krzysztof Parczewski, ical Engineer, NRR, O
'
-
-
f b9h
.
Approved by:
1.
Date' signed Albert F. Gibson, Director Division of Reactor Safety i
9304130183 930407 PDR ADOCK 05000327 PDR O
s
-
.
.
.
'
!
i i
ABSTRACT
[
INSPECTION:
,
This report documents the findings of a special Augmented Inspection Team (AIT) that evaluated the March 1,1993, failure of a 10-inch diameter extraction steam line and the associated operational problems when the resulting steam engulfed the main generator voltage control cabinet and caused the main generator output voltage to increase.
Inspection activities included inspection of failed components, reviews of plant design and operational parameters, interviews with personnel, reviews and observations of testing being conducted to bound the limits of the hardware problems, and reviews of licensee programs and management systems which could have had an effect on the event or could have been a causal factor for the event. The inspectors also reviewed the preliminary results of plant trip reviews and investigations being conducted by the licensee.
RESULTS:
l The SNP erosion / corrosion program was not effective in monitoring the condition of steam-cycle piping.
Significant deficiencies were found in the implementation of the erosion / corrosion program.
Ineffective management oversite and failure to clearly assign responsibilities were contributing causes to performance deficiencies.
The Operations staff's decision to manually trip the unit was proper.
The overvoltage transient raised voltage levels on safety related equipment, but the transient was within the design tolerances of the equipment involved.
Inspections and reviews by TVA and the AIT revealed no hardware deficiencies caused by the pipe break or the resulting voltage excursion.
Thera were no indications of degradation of safety related portions of the feedwater and steam piping systems. The non-safety related condensate, feedwater, and steam piping were found to be degraded.
l l
_ _ - _ _ _ _ _ _ _ _
',
i
.
,
...
ii
,
t SUMMARY On March 1,1993, a 10-inch diameter extraction steam line break in the Sequoyah Unit 2, turbine building caused the main generator voltage regulator to operate erratically and ultimately resulted in a manual reactor trip. The primary sequence of events was as follows:
1) A 10-inch diameter extraction steam line feeding the No. B2 feedwater heater ruptured.
2) The voltage regulator instrumentation cabinet for the main generator was engulfed in steam, which caused the main generator voltage output to rise to about 19%
over the normal voltage output for approximately 3 to 31/2 minutes. 3) The reactor operators manually tripped the reactor when they realized that they could not control the electrical excursion.
A secondary sequence of events served as a trigger for the pipe failure event.
Just prior to the pipe rupture event, a member of the' operating crew attached a fuse pulling device to the fuse for the trip solenoid for the feedwater regulating valve for the Unit 2, No. 3 steam generator. This action apparently initiated a trip of the feedwater regulating valve. Operators attempted to correct the feedwater/ steam flow mismatch. The resulting pressure transient in the balance of plant produced the steam line break.
The equipment failures were limited to the failed 10-inch piping and the.
voltage regulator function during the event. Subsequent testing of the regulator components found no hardware deficiencies that were attributed to the event.
Inspections and reviews by TVA and the AIT did not reveal any hardware deficiencies which were caused by the initiating pipe break or the resulting voltage excursion.
However, subsequent inspections by the licensee revealed significant degradation in other non-safety related steam, feedwater and condensate piping.
In addition the CHECMATE model of the extraction steam system was flawed. The medel ignored six, small diameter, operating vent lines attached to the steam line. TVA personnel missed opportunities to discover the degraded conditions before the failure. These opportunities included the 1985 discovery that the 10-inch elbows connected to the heater were severely eroded. These elbows were replaced with stainless steel, but there is no evidence that the adjacent piping was evaluated at that time.
In September 1991, the wall thickness of an area of the 10" pipe to the 2C heater was measured as 0.140" (Compared to 0.375" nominal). During the Spring 1992 outage, with the 20-inch header open, there were no visual inspections of the inside surfaces of the piping, and a step in the work order which called for wall thickness measurements of the 10" piping at the direction of the system engineer, was overlooked.
The following deficiencies were found in the implementation of the erosion / corrosion program for Sequoyah: 1) Trending of throughwall leaks in small diameter piping was not being done; 2) The CHECMATE model for the plant was constructed by one person, without benefit of independent checking by personnel familiar with the systems in question; This contributed to the error in modeling of the extraction steam header that failed; and, 3) The management of the erosion / corrosion program was fragmented, with part of the responsibility in the corporate offices and part of the responsibility at the sit,
.
.
.
.-
. -.
4
.f
<
.
I
'
i
!
,
The SNP erosion / corrosion program was not effective in monitoring the
,
condition of. steam-cycle piping.
The program was not effective because:
1) The plant staff relied on corporate engineers to provide them with an erosion / corrosion monitoring program that could be implemented without l
question.
2)
Corporate engineers accepted the fact that there were no isometric drawings for field run piping and truncated the plant model at 4-inch diameter rather than 2-inch,-as recommended by EPRI. 3)
Responsibilities in the areas of monitoring material condition of the plant
were fragmented.
4) Site and Corporate management, at all levels, accepted
EPRI computer programs to provide state-of-the-art monitoring of material i
conditions in steam cycle piping systems, without an assessment of the-l personnel and equipment resources required to properly set up and maintain the
programs.
The Operations staff's decision to manually trip the unit was proper.
There were no indications of degradation of safety related portions of the feedwater
'
and steam piping systems. The overvoltage transient raised voltage levels on safety related equipment, but the transient was within the design tolerances of the equipment involved.
.
't U
E I
,
-- _ -
.,.
,..
.
.-
.
-..
- -.
.. _. _,, - -
__
-
.-.
_.
-.
'.,
.
,
!
'
!
?
I
!
!
1.
Event Description and Seauence of Events:
'
On March 1, 1993, Sequoyah Unit 2 was operating at 100% power when a'
,
Shift Operations Supervisor erroneously moved a fuse in the power supply
'
circuit for the No. 3 steam generator feedwater regulating valve. The e
valve partially closed and the resultant pressure transient caused a
'
degraded 10-inch diameter extraction steam line to rupture. This initiating event is described more fully in Appendix C.
The " time-line" l
sequence of events is presented in Appendix B of this report.
An Assistant Unit Operator (AV0) in the turbine building heard a loud noise that sounded like the start of a steam leak.
Upon investigation
-
he found a large steam leak in the vicinity of the No. 2 FW heaters for
'
,
Unit 2.
The steam was close to, but had not yet reached, the main
'
generator voltage regulator cabinet that is located close to the feedwater heaters.
(An ASOS had, within the previous minute, passed
-
through the same area as a part of his routine inspection tour but had not detected any abnormal indications.)
The AU0 reported the leak to the control room as a " major steam leak"
.
and returned to the area to investigate the condition further. Upon
'
returning to the area, he noted that the steam had reached the voltage
.
regulator cabinet.
It was later determined that a small circulating fan l
was drawing steam into the cabinet.
.
When the steam began to envelope the voltage regulator cabinet, several j
alarms associated with the main generator were received in the control
room, indicating problems with the generator exciter and the voltage
!
regulator. At the time that these alarms were received in the control
!
room, the Unit 2 operators were already busy trying to control a
feedwater transient in steam generator No. 3.
,
!
The generator voltage increased to 27 kilovolts (kv) (normally 22.5 kv)
l and the generator reactive load increased to 550 mvar. The voltage
'
regulator tripped to the manual position, but attempts to decrease the.
)
generator output voltage were unsuccessful.
-
As a result, Unit 2 was manually tripped from full power at 2:24 pm.
'
The voltage on the 6.9 kv shutdown boards increased to approximately
.
8200 v for approximately 3.5 minutes, until plant buses were shifted to
'
,
'
the offsite supply. All plant systems responded as expected to the
,
plant trip.
Unit I continued to operate at full power during these events and was not affected either by the steam leak or the voltage problems. Once the significance and impact of the steam leak were determined, Unit I was voluntarily shutdown to conduct an investigation for similar problems.
The unexpected interactions between secondary plant steam cycle piping
'
systems and the electrical control systems were the basis for the
!
determination that a special NRC AIT inspection of the event was i
warranted. The AIT charter is included as Attachment 1 of this report.
i i
,
... _..
,
-
-
,
, --.
..
-
-
. -
,
-
-
'.
l
.
,
i
!
2.
Eauipment Failures:
This area of the inspection was divided into two separate categories, immediate equipment failures and potential equipment failures. As a
,
sub-set of each of these categories, the team evaluated the licensee's
program for monitoring flow assisted corrosion (FAC), otherwise known as
erosion / corrosion (E/C).
(For the purposes of this report, FAC and E/C
{
will be used interchangeably.) The licensee's past activities in the
'
area of E/C are discussed in paragraph 2.a.(3), and the licensee's changes to the E/C program are discussed in paragraph 2.b.(3).
'
a.
Immediate Eauipment Failures
'
The immediate equipment failures, the primary cause of the event, include steam related electrical problems, such as the function of the main generator voltage regulator, and the FAC in the ruptured extraction
'
steam piping.
For the purpose of this inspection, the failure of the licensee's E/C program to predict the pipe rupture is included as an l
immediate equipment failure.
l (1)
Electrical l
Steam from the steam line rupture on the #2 extraction line to the l
B2 feedwater heater engulfed the main generator voltage regulator
'
cubicle with steam. The heat and moisture caused the voltage regulator to generate an erroneous signal which called for the generator voltage to ramp up to approximately 27 kv (rated 24 kv).
(The 161 kv switchyard voltage went from about 166 kv to about 181 kv per the Chickamauga Load Dispatcher.)
When the voltage increased, the 6.9 kv shutdown board (safeguard
!
bus) overvoltage alarms sounded.
The operations personnel noticed
.
the analog meters on the diesel generator panels registering
!
between 8100 and 8200 v (i.e.,19% above the normal 6.9 kv) while the digital meters on the main control room panels (range 6400 to 1'
7400 v) registered a flashing 8888 v (defaulted value), which indicated voltage in excess of 7437 v.
As a result of the high voltage on the system and the operators inability to control the generator voltage (even with manual control), they manually tripped the reactor and verified that the
.
!
plant stabilized after the trip.
The shutdown board voltage remained high for 30 seconds after the trip, at which time the
'
generator output breakers opened (this is per plant design) and the unit 2 power to Unit 2 transferred from the Unit 2 station
service transformer to the common station service transformer).
'
This action returned system voltage to normal. A review of plant
.
data indicated that the overvoltage condition lasted approximately l
3 minutes and 38 seconds.
._.
-
.
.
_.
__..
_
^
.
.
L l
Following the steam pipe rupture, the area in the vicinity was-
.
visually checked by the licensee to determine what equipment could have been affected by the steam or condensation.
The following i
items of electrical equipment were identified:
-
6.9 kv Unit Boards (non-safety bus)
-
480 v Turbine Building Vent Board 2B
-
Main Generator Voltage Regulator (Excitor) Cubicle
-
Local panels
The licensee performed inspections of the 6.9 kv Unit Boards, the 480 v Turbine Building Vent Board, and Local panels, while
Westinghouse was asked to assess the condition of the voltage
'
regulator.
t The team observed licensee and Westinghouse assessments of the electrical equipment in the vicinity of the pipe break.
'
The licensee's inspection of the 6.9 kv Unit Boards, local panels, and 480 v Turbine Building Vent board did not find any water
intrusion inside any of the compartments.
The System Engineer's report of the work done by Westinghouse on the voltage regulator j
indicated that there was no permanent damage to the unit because
of the steam and water intrusion. There were a few components changed during the maintenance check-out, but they were not components which would have been involved in the transient
.
(2)
Pipina Failure The rupture occurred in the No. 2 extraction steam line in a'10" diameter, schedule 40S, (nominal wall of 0.375",) carbon steel pipe branch connection between the 20" header and the 2-2B FW
heater. The rupture was approximately 6" long by 4" wide and was i
located about I foot from the 10"X20" tee connection.
,
,
The licensee performed a metallurgical analysis of the failed pipe which revealed:
-
At the rupture, the pipe was thinned essentially through wall. Wall thinning in the failed area involved
approximately 25% of the circumference of the pipe. A
!
second significantly thinned area was observed approximately
I foot downstream of the failure about 180 degrees around
,
the circumference of the pipe.
This thinned area
.
encompassed an area about 8"X8" and was thinned to a minimum
.
thickness of 0.189".
-
-
In addition to wall thickness measurements, the analysis included chemistry, hardness, and microstructure analysis.
,
All results were consistent with that expected for ASTM A
106 GR B pipe, the required material.
!
l
- -
-
..
.
.
.
.
_
__._ _ _.
_.
'.
'
.
.
,
-
The analysis concluded that the failure was caused by the pipe being thinned by flow-assisted corrosion.
In addition to visual observation of the failed pipe, the team
,
reviewed the metallurgical report and observed the condition of identical header and piping from Unit 1. (After the Unit 2 failure, this piping was inspected by ultrasonic test (UT), found to be thinned, and removed).
,
(3)
Erosion / Corrosion Program
,
'
The team reviewed the past history and the status of the licensee's E/C program. The following summarizes the findings
'
from this review:
E The E/C predictive program was initiated for SNP in 1983. At that-time the components for inspection were selected by plant personnel using engineering judgement and industry experience.
In 1986, after an event at Surry, additional balance of plant piping and components were added to the program. Approximately 40
'
components in single-phase flow and 30 components in two-phase flow were included in the program.
In 1987 EPRI's CHEC program was introduced for single-phase systems.
In the two-phase systems
,
selections were made using Keller's equation and the methodology
-
described in NUREG-5007.
In 1991 EPRI's CHECMATE program was implemented and General Engineering Specification, G-97B, Erosion-Corrosion, was prepared.
This specification served as a basis for preparing the site specific Periodic Instructions (PIs) for controlling.the erosion-corrosion. The following PIs were issued:
1-PI-SXX-003-001.0, Unit 1 Wall Degradation Monitoring Program
-
for the Feedwater/ Condensate Piping, Turbine and Heater Drain Lines, Revision 3.
'
1-PI-SXX-005-001.1, Unit 1 Extraction Steam Pipe Wall Degradation Monitoring Program, Revision 3.
2-PI-SXX-003-001.0, Unit 2 Wall Degradation Monitoring Program for the Feedwater/ Condensate Piping, Turbine and Heater Drain Lines, Revision 2.
2-PI-SXX-005-001.1, Unit 2 Extraction Steam Pipe Wall
-
Degradation Monitoring Program, Revision 2.
The CHECMATE models for both units were developed at the corporate office.
With some exceptions, only components greater than 4" in
,
y
.
-
.
. _.
.. _ _ _
_
_
~--
.
. _.
_
..
. _ -
' '.
I c
.
!
!
diameter were modelled, since smaller diameter piping was field
run during construction, and configuration drawings were not i
available.
[
The corporate office also ran Pass 1 of the CHECMATE for Units 1
and 2.
The following systems were modelled for both Unit 1 and i
Unit 2:
Condensate
'
Heater Drains and Vents Main steam The Pass 1 did not include UT measured data and could be used only l
for ranking of the components according to the degree of their E/C susceptibility.
Subsequently, CHECMATE programs, including the system models, and the responsibility for running the code, were transferred to the site, and Pass 2 of the program was run by the site personnel for Unit 2 only. The responsibility was transferred to a 2-person Materials Engineering contingent whose other responsibilities included all welding and materials related concerns for the site, including problems with procurement and commercial grade
,
dedication. The responsibility was met by assigning a co-op
'
student to learn the CHECMATE program and operate it for the site.
The co-op student subsequently returned to school, leaving the
site with no one trained to operate the CHECMATE program.
- In order to run Pass 2 of the code, the model had to be validated-
!
by including the measured wall thicknesses for the selected components (TDAT).
Eighty-three (83) Toxys, obtained from UT
measurements, were used to validate the Code.
Because the transfer of the CHECMATE codes to the site apparently did not i
include a computer with adequate capacity to run the CHEC-NDE portion of the program, the licensee used hand calculations for i
determining T These calculations required determination of DAT.
the maximum measured wear rate for each individual component. The license calculated this value by subtracting minimum wall
'
thicknesses for a component, measured at two different i
inspections. This method differs from the EPRI recommended methods (which are included in the CHEC-NDE program) and in some l
instances can yield non-conservative results.
This is because
'
minimum wall thickness in a component may shif t from one location
'
to another, which would cause under-estimation of the E/C wear
rate.
In addition to a review of the licensee's specification, procedures, and program history, the team reviewed the input data and output results from the CHECMATE code for Unit 2 steam i
l
$
_.
, _ _ __
_ - _ -
.
_
-
-._
-
'.
.
>
-
.
extraction line containing the 10" piping to the B2.FW heater.
!
Inaccuracies in the input data to the CHECMATE code were noted, which probably caused the failure of the code to predict ~ the extensive E/C damage to the ruptured pipe. Actually, some of the i
adjacent components were predicted to have considerably higher i
degree of E/C damage than the ruptured pipe.
.
The input error which had the most significant effect on faulty predictions was a failure to model six, 3" vent lines upstream from the failed component, which injected two-phase fluid into the extraction line.
This omission of two-phase flow could
-
significantly change the modelled flow and thermodynamic
'
characteristics of the fluid in the steam extraction line and
>
result in incorrect CHECMATE predictions.
(This error had been identified by TVA during the investigation of the event, prior to this inspection.)
As a result of the review of the existing E/C program, the following weaknesses in the predictive methods were identified:
,
-
In developing the CHECMATE model, not all the pertinent
plant parameters were incorporated accurately. This could
!
be a source of significant inaccuracies and limit the predictive capabilities of the model. This was demonstrated i
by neglecting to include the input from the vent lines in the CHECMATE model for the ruptured steam extraction line.
-
There was no second party review of the input to the CHECMATE model.
Lack of this verification could lead to input errors, such as failure to include the vent lines in the extraction steam line model.
'
-
The licensee's G-97B Specification required that, for curved
pipes, straight, expanding and reducing elbows and tees, UT examination should extend 6 in. downstream from the examined component.
This specification does not follow EPRI's i
recommendation that UT examinations should extend 2 pipe diameters downstream from the examined component.
i
!
.
-
In determining minimum measured wall thickness (TDAT) for a component from UT measurement, the licensee did not follow
,
either the EPRI recommended band, or point-to-point,
methodologies. The method used by the licensee could yield non-conservative results.
,
-
Responsibility for the E/C program was fragmented with inadequate management attention and no real ownership.
-
The small-bore inspection program was very limited. During
'
Unit I cycle 5 outage only 9 components of 4" or below were
.
I
_
.
.
.
.-
-
..
.
.- -
. -- -
-
-
.- -
.
.
..
,
inspected.
During Unit 2 cycle 5 outage only 10 components j
4" or below were included in the inspection program.
j
-
Measured wall thickness (Tw) on only 83 components were used to validate the CHECMATE code, and these were not
'
evenly distributed among the systems modelled. This could result in inaccurate code predictions.
-
As reported in NRC Inspection Report 50-327,328/93-04, the-
,
small bore piping systems were apparently the assumed
'
responsibility of the maintenance department, who took it
!
upon themselves to contain leaks with furmanite " cans" until damaged components could be replaced. There was apparently no assigned or assumed responsibility to track or trend the i
number and type of failures for predictive purposes.
!
I b.
Potential Eauipment Failures.
Potential equipment failures involved the possibility of damage to safety related electrical equipment because of the overvoltage
condition, and general condition of carbon steel piping components in
'
the steam, condensate, and feedwater systems.
(1)
Electrical Prior to the opening of the generator output breakers, the
,
voltmeters associated with 6.9 kv shutdown board voltages were observed by the Control Room Operators to be approaching 8200 v and the Load Dispatcher observed that the 161 kv switchyard voltage went from 166 kv to about 181 kv. The observed voltage readings indicate that the overvoltage condition was between 12%
(181kv:161kv) and 19% (8.2kv:6.9kv) high. The overvoltage I
condition lasted approximately 3 minutes 38 seconds.
The licensee evaluated the potential adverse effects of the
,
overvoltage condition on various electrical components to verify
<
proper operation.
The electrical components considered in the i
evaluations were: circuit breakers, motor circuit protectors, transformers, relays (electromechanical and solid state), starters and contactors, switches, fuses, controllers for pressurizer heaters, battery chargers, inverters, pump motors, and MOVs.
,
)
The licensee had previously performed an engineering analysis for i
the susceptibility of 6.9 kv and 480 v Shutdown Board electrical devices to prolonged overvoltage conditions that may have existed
when Sequoyah was in an extended shutdown.
Based on this previous
analysis, and other engineering considerations, the licensee concluded that no additional evaluation or testing was necessary for the following types of components.
!
~.
.. _.
-
-
-
-
.-
-
.
.
.
.
-
Testing of circuit breakers, motor circuit protectors, current transformers, solid state under/overvoltage and ground fault relays, starters and contactors, switches, fuse blocks, and meters was not necessary for those that.were I
rated for 125% of nominal voltage.
-
Fuses would not be adversely affected by the overvoltage but affected by the short duration overcurrent. Since no blown fuses were identified, the licensee concluded that the fuses l
were not affected.
-
Transformers (with exception of Main and Unit Station Service Transformers) are required by ANS C57 12.00 to
operate at 5% overvoltage fully loaded and 10 % overvoltage
,
unloaded, without exceeding guaranteed temperature rise.
Since these are liquid cooled transformers operating well j
below rated load, the licensee concluded that any increase in load due to overvoltage'would have been for such a short duration, that the heating effect would be negligible
'
compared to the rated temperature rating.
The licensee took or planned the following actions for other
~
equipment:
i
-
For those mechanical relays (W SG, CV-7, GE IAV54 and
.,'
RAV11B) which are rated for 10% maximum overvoltage, the
'
licensee will check the calibration of TS relays by performing regular surveillances.
These relays could have
been subjected to excessive torque due to the high voltage, which may have affected delicate. spring mechanisms.
If a pattern of calibration shifts is detected, the scope of
calibration check program would be enlarged to include other
,
IE and non-lE relays. This will also include verification of operation and control of pressurizer heaters and RPS l
undervoltage relay.
'
i
-
Inspect all battery chargers and inverters to verify proper operation.
-
Since no' motors tripped off line, the conclusion was reached f
that the overvoltage was not of sufficient duration to cause
[
significant overheating of the motors. However, to demonstrate that the voltage remained well below the motor
dielectric capability (i.e., no significant motor damage),
+
the licensee will megger a sample of the motors that were
running during the overvoltage event.
,
!
-
No reports of motor operated valve, (MOV) operational problems were reported. However, the licensee will confirm
,
that no damage was sustained by performance of M0 VATS
testing.
Comparisons of motor current traces from these new
^
t
.-,
.n.
.
...
,
-
--
- _.
...
--
-
.-
..
-
,
i
a L
t
tests with results from previous M0 VATS tests will be l
I performed to confirm that no damage was sustained.
-
Oil and gas from The Main and Unit Station Service i
Transformer will be sampled, and tested for heat related I
changes, to determine the effect of the overvoltage
excursion.
i The team reviewed the licensee's engineering evaluations and l
discussed these evaluations, as well as plans for confirmatory
!
testing, with site and corporate engineers and managers.
Several
,
of the tests were reviewed in process and/or the results of the
!
tests were independently reviewed by the team.
For battery chargers and inverters, the team observed the licensee l
performing 250 v battery charger 1, and 120 vac preferred inverter 1, measurements of the input and output voltages, visual
!
inspection of components, terminal blocks, resistors, capacitors
!
and panel wiring for any sign of heat related damage. There were
'
no problems found during the inspection of the battery chargers
'
and inverters.
The team observed meggering of 6.9 kv No.7 heater drain pump motor l
2B.
The results from meggering of other 6.9 kv and 480 v pump
motors were also reviewed.
The tests did not show any signs of l
degradation of dielectric capability due to the overvoltage.
i Motor current traces for the MOVs which operated during the event j
were compared with results from previous M0 VATS performance tests.
The comparison showed no appreciable differences.
l
,
'
The licensee's inspection of the selected samples of electrical equipment did not reveal any degradation which could be attributed
!
to the overvoltage excursion.
Since any degradation due to
!
overvoltage is usually caused by a combination of the heat
!
associated with the increased load current and the duration of the j
overvoltage experienced, the licensee and the team concluded that i
the level of overvoltage and the duration of the event which occurred at Sequoyah was not of sufficient level and duration to i
cause damage. The fact that normal operation of the electrical l
equipment was below rated design temperature and the ambient
temperature was low during the event also lessened the impact of
!
the event.
(2)
Pipe Wall Thinnina l
With both units shut down, the licensee management originally i
decided to concentrate the piping inspection and replacement
,
activities in an effort to return Unit I to service prior to the restoration of Unit 2.
The personnel and programs involved in the j
,
restart of Unit I are the same ones that would be involved with
'
the restoration of Unit 2.
For this reason, this part of the i
, - - _ - - - - _ _. - _ -
...
,
'.
l
,
.
i i
inspection concentrated on Unit 1 activities as well as activities related to Unit 2.
!
The team examined in-process activities relative to licensee's
!
current pipe wall thinning inspection effort, including the small-
-
bore inspections completed after the failure of a Unit 2, 3"
!
operating-vent, target-tee, in January, 1993.
Sample selection
criteria, inspection methods, and inspection results were examined i
as detailed below.
i (a)
Large Bore Inspection Criteria and Results (Unit 1)
i i
Based on failure of Unit 2, extraction steam #2, 10" diameter piping, the licensee measured the wall thickness of
identical piping, including the 10" piping and 20" header on j
the Unit 1, #2, extraction steam system.
The Unit 1 piping
)
was found to be thinned similar to the Unit 2, although not
.
as severe as Unit 2.
The Unit 1 and Unit 2 headers and 10" -
piping were cut out for replacement.
}
i In order to determine the extent of wall thinning in Unit 1 systems, the licensee developed a screening, UT thickness
,
measurement, program. This program was not meant to provide j
data suitable to feed the CHECMATE Erosion / Corrosion (E/C)
program.
For this screening program, Nuclear Engineering i
selected the inspection sample points based on the
'
following-
'
-
A Pass 1 CHECMATE model of Unit I systems susceptible to E/C had been completed. As a start, additional
'
inspection points (not already in the E/C program)
'
were selected for inspection. The components with the shortest life / highest wear rates from the Pass 1
CHECMATE model were selected.
f
-
Other points for inspection were selected based on
!
operating / leakage experience at SNP.
-
Industry experience (i.e. failures, other inspections l
at TVA facilities, and other utilities) was also used i
as input for the sample selection.
i
-
Input from personnel with extensive background in t
previous E/C issues at TVA was also used.
The inspection sample was increased based on the inspection
[
results.
After selection of inspection points, marked-up isometric
drawings were provided to the nondestructive examination
!
(NDE) personnel for UT thickness measurements.
Since this i
was only a screening process, E/C grid layout was not
!
!
!
,
,
_
_
-,
.,-
--
.-
-
-
.
=
-.
.
._
'
'
.
"
!
!
>
performed.
Instead, Nuclear Engineering provided NDE
'
personnel with instructions for the inspection scope for each inspection point.
In general, the following scanning
criteria was used:
!
q
-
For elbows, a continuous scan of a 6" band of the i
outer radius was used.
-
For piping adjacent to fittings and piping
intersections (typically 1 pipe diameter upstream and
+
2 pipe diameters downstream), continuous
circumferential 360 degree scans at 3" intervals were l
used.
!
,
-
For extended lengths of pipe, continuous
_
circumferential 360 degrees scans at I foot intervals were used.
i
-
For each interval, the lowest reading with a description of the low measurement was recarded.
j Based on the above selection criteria and scan guidelines, l
as of March 11, 1993, 216 inspection points had been
,
selected. Of the 216 points (included the #2 steam
extraction header and 10" piping noted above), preliminary
'
inspection had been completed on 153.
Fourteen (14) points
(components) were rejected because of wall thinning below minimum required thickness (minimum based on hoop and
,
!
bending stresses plus wear allowance). A similar inspection program was planned for Unit 2 large bore piping.
!
On March 9, 1993, TVA announced the decision to not restart Unit 1 prior to the planned April refuelling outage. The I
emphasis for screening UT measurements shifted from Unit I
to Unit 2. and an inspection program similar to the Unit 1
program was initiated.
(b)
Small Bore Inspection Programs and Results (Units 1 &
E 2)
l i
After the January, 1993, 3" diameter target tee failure in a Unit 2 operating vent line, TVA developed an inspection
program for inspection and repair of small bore piping. The
following summarizes the program for Unit I-
.
-
The scope of the program included 30 piping runs in l
Moisture Separator Reheater (MSR) high pressure and
!
low pressure vents, MSR high pressure and low pressure
drains downstream.of level control valves, low
!
pressure extraction drains, feedwater heater drains to l
- 3 heater drain tank (actually large bore component -
6"X10" reducer), and feedwater heater relief tailpipe.
i
!
i
.
i
i
,.
.
!
-
The wall thickness mear.urements started at the tie-in
to the low pressure piping and all upstream
,
components, including piping, were inspected until two
'
consecutive acceptable fittings were found.
Inspection of fittings included adjacent piping at the
,
toe of the fitting to pipe weld and at 6" intervals i
along the pipe. A minimum of 5 UT readings in a diamond pattern on the outside radius of the ell and 4
>
90 degrees UT measurements at each interval on the i
piping were taken. These were minimum requirements and in many cases, UT scanning was much more extensive. This process resulted in inspection of the
-
following approximate quantities of piping and piping
components:
Feet of Pipe
-
182 Tees
-
-
Elbows
-
-
1 Reducers
-
26-
!
-
As a result of the inspections,14 runs of piping
.
required some material replacement, including 5 tees,
!
10 elbows, 6 flanges, 26 feet of 1-1/2" diameter pipe, and 6 feet of 4" diameter pipe.
In addition, 2' scab i
plates (at localized thinned areas - to be replaced at
next outage) and I weld repair were installed.
t A similar inspection program has been issued for Unit 2 and
,
was in pro::ess during the inspection. As of March 8, 1993, 58 runs of piping had been identified to be evaluated.
Approximately 30 runs had been evaluated and 113 inspection
points in vent and drain small-bore piping had been selected i
for inspection, F:fty-five (55) points had been inspected with 10 of the 55 iound to be below minimum thickness. As a result, seven (7) of the 30 runs require some material i
repair / replacement.
J (c)
Observation of Wall Thickness Measurements
[
.
To verify the wall thickness measurement techniques, the team reviewed procedures and observed in-process UT-inspections, and reviewed radiographic (RT) film (small-bore piping) as follows:
UT - The applicable UT procedure is N-UT-26, Revision 12, TC 93-10, Ultrasonic Examination for the Detection of
ID Pitting, Erosion, and Corrosion.
The team reviewed procedure N-UT-26 and observed UT wall
thickness measurements for the following grids:
l
!
!
l l
.
.
.
.
-
t I
,
-
3G - #3 Extraction Steam - During this inspection, the examiners identified a thinned area on the 16" bare of the 28"X16" Tee. The minimum wall measured was 0.160", which was
below the required nominal wall thickness. The licensee was considering a weld buildup of the thin areas as a preliminary. fix.
- i
-
3B - #3 Extraction Steam - This component was a F
28"X10" tee. The inspection results were
acceptable on this component.
l
-
61 - Feedwater - 20"X32" Tee - Wall. thickness appeared to be uniform at the time of f
observation.
-
-
6H - Feedwater - 20" Pipe and Elbow - At the time of the observation, wall thickness measurements showed uniform thicknesses.
I
-
1052E032 - #2 Extraction Steam 20" Elbow - This
,
elbow had a full grid inspection to feed the new CHECMATE model.
In addition to observation of UT of this grid the team also observed grid placement activities and reviewed UT, including i
DATAMATE III output, data for grids 1052E048 and
,
1052T041.
Layout of grids was being accomplished in accordance with Attachment I to Work Order (WO) 93-00438-00.
UT examiner certification records were reviewed for 1 level III, 3 level II, and 2 level I examiners. The j
level 1 examiners were TVA employees and the other
examiners were contractor (Asea Brown Boveri)
'
employees.
!
Equipment Certification records for the following UT
[
equipment and materials were reviewed:
UT Couplant
- Batch #093001
'
UT Coup 1 ant
- Batch #9008
UT Couplant
- Batch #092008 i
UT Instrument
- Serial #489414 l
UT Instrument
- Serial #522008 UT Instrument
- Serial'#497619 j
UT Instrument
- Serial #489560 l
UT Transducer
- Serial #E04702
!
UT Transducer
- Serial #D20724 UT Transducer
- Serial #D20729 l
UT Transducer
- Serial #M18644 j
.
_
.
,
i
,
'
RT -
For small-bore piping inspections in Unit 2, the licensee is using RT to measure wall thickness in many cases.
The applicable procedure is N-RT-2, Revision
,
3, TC 93-09, Radiography Examination of Structures,
'
Systems, and Components (Non-Mandatory). The team
!
reviewed TC 93-09 and reviewed RT film for the-l'
following inspection locations: 4B, 6A, 6B, 5B, SA, 4A, 2C2B, 282B, 2C2A, 201B, 2B2A, 2Al HP, 281 HP, 2C1 HP, 2A2 HP, 2B2 HP, 2C2 HP, and demonstration of the
'
RT technique on a Target Tee.
The RT technique appears to be providing good results
and many of the films reviewed exhibited evidence of
$
wall thinning.
,
The team concluded that the licensee had a comprehensive inspection effort in progress to bound the wall thinning problem.
By the conclusion of the inspection, the Unit.1 large-bore sample a
size had been increased to 216 inspection points.
The sample size i
for the Unit 1 small-bore inspection effort was 149 components
plus 182 feet of pipe.
Similar sample sizes for Unit 2 were being
,
planned. The methods being used for wall thickness measurement appeared to be well controlled. Approved procedures and qualified personnel were being used.
l (3)
Erosion / Corrosion Proaram (
The team evaluated the licensae's planned improvements to the E/C program. One improvement, obtaining outside help to upgrade the E/C and CHECMATE programs, was in process at the time of this inspection.
The licensee had contracted ABB Impell Corporation
,
for developing future predictive E/C program. The scope of the i
work for the contractor consisted of.
,
(a) Upgrading the existing predictive program, based on CHECMATE code by revising input parameters and including additional UT data, (b)
Expanding the program to include additional systems, as
recommended by EPRI, and
(c)
Implementing the CHEC-NDE code for managing UT results and calculating minimum measured wall thickness (TDAT), to i
be used in subsequent CHECMATE passes.
,
The contractor's work was divided into three phases.
Phase one i
contained the systems which had to be evaluated before the restart of Unit 1.
(As stated above, until March 9, 1993, Unit I had the l
highest priority.) The systems included in this phase were:
i Extraction Steam l
,
>
b s
.
- -
..
. _.
-
-
- =. -
--
--
i
~
..
.
,
.k
)
.L
j Reheater Drain Moisture Separator i
Feedwater Recirculation l
The other two phases contained the systems which form'a part of the long term erosion-corrosion program to be implemented in the
?
Sequoyah plant.
l t
The contractor, in close cooperation with the~ plant technical
!
personnel, was preparing inputs to the CHECMATE code for the phase
.t one systems. The plant's design, operational and chemistry
parameters were being verified before their inclusion into the i
systems modelled.
Pass 1 for some of the systems in Unit I had been completed and
,
the team reviewed the output data for the extraction steam lines.
l The results indicated that the 10" pipe-to'-feedwater heater, i
analogous to the one which failed in Unit 2, was predicted to have
+
the highest E/C damage.
!
P'
The contractor had started to prepare input for pass 2.
Pass 2 requires validation of the CHECMATE code by including measured wall thicknesses (T,7) for a few components. The licensee o
,
'
intended to include approximately 5.T s per system.
In order to DAT obtain these DATs the UT measurements were to be performed on
-
selected components and the results entered into and processed by
!
the CHEC-NDE code.
i The licensee indicated that the above new CHECMATE analysis, will
!
be developed for large-bore piping.
For this effort, the i
definition of large-bore will be-2-1/2" and larger. Where
-!
necessary, field walkdowns will be performed to determine actual j
piping configurations. This effort will be part of any restart
.
plan.
i 3.
Human f actors / Procedural Deficiencies
{
There were several human factors / procedural deficiencies which directly
'
contributed to the event.
i
!
a.
CHECMATE model
!
!
The model of the extraction steam system put into the CHECMATE computer i
program, designed to predict the type of FAC that caused the failure, was flawed.
The model ignored six, small diameter, operating vent lines attached to the steam _line. These vent lines added enough moisture to l
the steam to significantly. change the FAC rate of the downstream piping.
!
!
!
!
I
!
!
t
-
.
.
.
-
I
m
.,
-,
-
_... - -.. _.. _. - -
. - _
,_,
a m
..a_#.a
_m_,-
ma_-
_
s.
-
'
,
.
.
'
i
?
!
!
b.
Plant material monitorinq
!
Plant and Corporate personnel involved in the monitoring of the FAC
!
program for the site overlooked several material condition warning signs and therefore missed opportunities to discover the severely degraded
conditions before the failure. These warnings included the following:
'
l (1)
In 1985 during replacement of the 2A, 2B, and 2C feedwater
-
heaters, the 10-inch elbows at each of the heaters were found to be severely eroded and replaced with stainless steel.
(2)
In Sept 1991, there was a steam leak from the weep-hole in the doubler plate at the junction of the 10" and 20" line for the
.
2C Heater. During the Furmanite injection to stop the steam leak,
-
the wall thickness of an area of the 10" pipe was measured as
0.140".
An engineering error led to the return of the system to i
service. The minimum wall required for return to service should
have been 0.250", whereas the work order stated that the required
'
minimum wall was 0.107".
i i
(3)
A work order to replace the carbon steel header with l
stainless steel, during the Spring 1992 outage was changed to a
!
work order to open up the 20" header and to weld repair the
!
junctions of the 10" to 20" piping from the inside.
During the
weld repair evolution, there were no visual inspections of the i
inside surfaces of the piping, and a step in the work order which called for wall thickness measurements of the 10" piping at the direction of the system engineer, was not performed.
'
c.
Fuse Location The fuse for the feedwater regulating valve, trip system, solenoid valve l
was located about midway in a vertical strip of fuses that ranged from
,
just above floor level to about six feet off the floor. The location of
i fuse 10 tags adjacent to the fuses required that operators bend down to (
eliminate parallax when confirming that they were pulling the right i
i fuses.
j 4.
Probable Contributing Causes of the Event:
!
The AIT found that the primary contributing causes of the event were:
a.
The plant relied on corporate engineers to provide them with an
!
erosion / corrosion monitoring program that could be implemented a
without question.
b.
Corporate engineering accepted the fact that there were no isometric drawings for field run piping and therefore truncated the plant model at 4" diameter piping rather that 2" diameter as recommended by EPRI.
'
!
'
,
,.
_
-
..
-
-
- _.
-
- _. -
-
.
~
j
.
.
e
c.
Responsibilities for monitoring material condition of balance of plant piping were not clearly assigned.
In the area of the i
turbine building piping vulnerable to FAC, personnel involved included maintenance supervisors, system engineers, site materials engineers, and corporate materials engineers. There were no clear l
guidelines as to who really had responsibility.
For at least the
j last year, all four groups had sufficient information that any one i
of them could have taken the initiative and determined that there were serious problems in the extraction steam systems.
d.
Site and corporate engineering accepted that a computer capable of
~
handling the CHEC-NDE program was not available to them; so they made do with manual methods of analyzing the NDE thickness readings. The manual method compares the thinnest readings from
-
two inspections, regardless of location on the fitting, to determine wear rate; while CHEC-NDE uses a point-to-point thickness comparison to establish wear rates at a number of locations and then records the largest change as the wear rate.
The manual method could have led to possible non-conservative data j
comparisons.
t 5.
Findings and Conclusions:
,
i
-
a.
The AIT findings were as follows:
'
(1)
The operating staff's decision to manually trip the unit as soon as they realized that they could not readily control the observed conditions was proper.
t (2)
The engineering staffs (systems and materials engineering)
involved with monitoring the condition of the extraction steam lines were not trending material condition indicators
,
that should have alerted them to the potential problems
,
prior to the failure.
(3)
The erosion / corrosion program was not representative of the actual plant systems in that small diameter lines which
-
provided significant moisture inventory input to the I
extraction steam system were not included.
The program also did not meet EPRI guidelines, in that it only included pipe sizes above 4-inch diameter rather than above 2-inch diameter.
(4)
The design voltage ratings for a majority of electrical
'
equipment at SNP were 125% of the nominal operating voltage.
b.
AIT conclusions were as follows:
i (1)
The SNP erosion / corrosion program was not effective in i
monitoring the condition of steam-cycle piping.
Ineffective
!
management oversite and failure to clearly assign i
t
[
,
- - - - - - - -.
. - -
- -
.
- - - - - -
- - - - - - - -
..u-v
-
_.
__
_
_ _. _ _ _.
.. _
_.
.
l
.
c
)
responsibility for management of the over-all E/C program.
.
They accepted the premise that the introduction of a family I
of computer modelling programs called CHEC, CHECMATE, and
'
CHEC-NDE would provide state-of-the-art control and i
monitoring of material conditions in steam cycle piping
systems without an assessment of the resources required to
properly set up and maintain the programs.
Decisions about i
staffing and equipment requirements seem to have been made
!
at the lowest supervisory levels.
!
l (2)
The condition of safety related feedwater and steam piping were satisfactory.
The non-safety related condensate,
,
feedwater, and steam piping were found to be significantly
,
degraded.
'
(3)
The overvoltage transient raised voltage levels on safety related equipment, but the transient was within the design tolerances of the equipment involved.
j
!
i i
i
!
i f:
t i
,
i
l I
-
f
,
ll
,
!
.,
I b
i
>
>
i
_.m.
-
.~-
,
.
--
.
-
.-
-
_
,'
,-
4'
t k
APPENDIX A - PERSONNEL CONTACTED
Licensee Personnel:
- D. Ashley, Acting Manager, SQN Training
!
E. Barels, Systems Engineer
- J. Baumstark, Operations Manager
- R. Beecken, Plant Manager R. Bentley, NDE Level III
,
E. Booker, Maintenance Superintendent
- L. Bryant, Maintenance Manager
- L. Butcher, Nuclear Engineering, Lead Electrical Engineer
- J. Bynum, Vice President, Nuclear Operations P. Caldwell, Nuclear Engineering, Electrical Engineer
- M. Cooper, Site Licensing Manager
- R. Drake, Manager, Project Management / Controls
- B. Durst, Nuclear Engineering, Lead Mechanical Engineer
- M. Fecht, Manager NER/ISE i
- R. Fenech, Vice President, Sequoyah
'
- T. Flippo, Site Quality Manager l
W. Goins, Nuclear Engineering
'l
K. Greene, Senior Electrical Equipment Specialist
'
J. Hamilton, Quality Audit & Assessment Manager
J. Hanson, INP0
'
M. Heatherly, Principle Mechanical Engineer r
I
B. Kimsey, Project Principle Electrical Engineer
- 0. Kingsley, President, Generation Group J. Long, Systems Engineer
!
D. Lundy, Nuclect Engineering
'
R. Mages, Nuclear Engineering, Electrical Engineer i
- T. McGrath, Chairman, NSRB
'
J. Mills, Nuclear Engineering, Mechanical Engineer R. Mooney, Mechanical Systems Engineer - BOP i
J. Pitchford, ASOS
,
- R. Proffitt, Compliance Licensing Engineer i
- R. Rausch, Modification Manager
- J. Robertson, ISEG Manager
- H. Rogers, Acting Technical Support Manager
- M. Shepherd, SQN Training Manager i
- M. Skarzinski, Reactor Engineering Supervisor
'
G. Strickland, Corporate Materials Engineer
- R. Thompson, Compliance Licensing Manager
- P. Trudel, Site Engineering Manager
)
- M. Turnbow, Manager Inspection Services
G. Wade, ISI/NDE Supervisor
- J. Ward, Manager, Engineering / Modifications
- N. Welch, Operations Superintendent
- K. Whittenburg, Public Relations Manager D. Willis, Systems Engineer T. Woods, Chief Materials & Inspection Department
Other licensee employees contacted during this inspection included technical support, maintenance and modifications, inspection, and administrative personnel.
Attended Entrance Meeting - March 4, 1993
- Attended Exit Meeting - March II, 1993
~. -
. _ _
_
__
_
.
.'
.
.
A-2 APPENDIX A - PERSONNEL CONTACTED (Continued)-
NRC Personnel.:
- J. Blake, AIT Team Leader, (Eng. Section Chief, RII)
- B. Crowley, AIT Team Member, (Engineering Inspector, RII)
!
- S. Ebneter Regional Administrator, Region II
- B. Holland Senior Resident Inspector, Sequoyah
';
- G. Lainas-Assistant Director for RII Reactors, NRR-
.K. Parczewski, AIT Team Member, (Chemical Engineer, NRR)
'
- S. Shaeffer, Resident Inspector, Sequoyah
- S. Sparks Project Engineer for Sequoyah, RII
,
!
,
!
!
!
!
!
I
!
,
.
i
,
- Attended Entrance Meeting'- March 4, 1993 l
- Attended Exit Meeting - March II, 1993
l
!
.
w
-
-
-
.
.
.
.'
.
APPENDIX B - SEQUENCE OF EVENTS The " time-line" sequence of events was as follows:
a 1400 Actions were initiated to remove power to a Reactor Building Floor and Equipment sump pump containment isolation valve when its series valve was declared inoperable.
The valve was shut and its solenoid valve was being de-energized by removing the control power fuses.
This would cause the valve to remain in the shut position.
m 1410 An ASOS who had just completed his switchyard inspection, passed directly above the B2 heater on his way back to the main control room (normal pathway).
He was in visual ~ sight of the area that subsequently ruptured and observed no steam at that time.
m 1419 An ASOS was assigned to pull the control fuse for the solenoid valve for the containment isolation valve.
He placed the fuse pullers on a fuse and moved it sideways to ensure he had a firm grip on it. He then re-checked the fuse number, realized he had placed his fuse pullers on the r
wrong fuse, slid the original fuse back to its original position in the fuse clip (the fuse did not lose contact with the fuse clip), and proceeded to remove the correct fuse.
The fuse panel that the ASOS was working with contains a vertical strip of horizontal fuses. The strip runs from
-
near the floor of the cabinet to about six feet off the floor.
The fuses are located so close to each other that plastic inserts have been installed to prevent contact between the fuse holders when they are displaced during insertion or removal of fuses.
The fuses involved during this incident are located about three feet off the floor and the ASOS apparently attached his fuse puller to what he thought was the correct fuse while standing erect, and when he bent his knees to view the fuse at eye level he determined that he was prepared to pull the wrong fuse.
Rather than attaching the fuse puller to the correct fuse, the ASOS had attached it to the next lower fuse; the power supply fuse for the No. 3 steam generator feedwater regulating valve, solenoid trip valve.
m 1419 Alarms that were received in the control room indicated a steam flow / feed flow deviation and low level in steam generator (S/G) #3.
The reactor operator took manual control of the S/G level regulating controller and increased flow to raise the S/G level back to normal.
,
,
.
.
.
.
B-2 Subsequent investigation revealed that the #3 S/G feed flow had dropped down close to zero at the start of the event and was the cause of the deviation annunciation.
S/G level dropped as low as 22 percent, but rapidly returned to normal due to the operator action.
.
An initial attempt to place the level controller back to AUTO was unsuccessful and the level controller was placed back to MANUAL.
-
The Unit 2 AVO heard a loud noise that sounded like the start of a large steam leak.
He immediately went to the i
area of the noise to determine the source and found a " major steam leak" in the vicinity of the #2 feedwater heaters and reported it to the Unit 2 control room.
The steam was blowing up through the grating, around the #3 heaters, and was very close to the generator exciter cubicle, but had not yet reached it. When the AVO returned to the area, the steam plume had reached and totally covered the exciter cubicle, and moisture was seen on the external surface of this cubicle.
The AVO then proceeded to the #2 heaters and was able to verify that the steam leak was from an extraction line between the common header and the valve leading into the B2 heater.
When the control room learned of the " major steam leak" they were still concentrating on regaining control of S/G #3 l evel. They acknowledged the " major steam leak" in the turbine building and began to increase their monitoring of other primary and secondary parameters.
Their immediate
'
concern was still focused on regaining level and control of the #3 S/G.
By this time the Shift Operations Supervisor (SOS) was also assisting with the transient.
m 1420 Unit 2 received the following alarms associated with the main generator: exciter rectifier power loss or failure, Unit 2 exciter insulation resistance low, generator exciter field overcurrent.
m 1421 Additional alarms annunciated: generator volts per cycle high, generator voltage regulator trip, 6.9 kv Board 2B-B and 2A-A overvoltage.
The operators noted that generator output voltage was approximately 27 kv, generator reactive load had increased to approximately 550 MVAR, and shutdown board voltage was approximately 8200 volts (as observed on the common diesel generator panel - the shutdown board digital voltage indicator had exceeded its range and was flashing "8888").
When the voltage regulator tripped to MANUAL, the R0 tried I
to decrease voltage using the Base Adjust Controller Sut was unsuccessful.
At this point the SOS (who was in i
l i
,
-
.
.
.
.
.
B-3 communication with the Operations Superintendent) and the AS0S made the decision to trip the reactor due to the uncontrollable generator voltage condition.
m 1424 Unit 2 was manually tripped and the emergency procedure entered.
m 1425 The shutdown board voltage remained high for 30 seconds after the trip, at which time the generator output breakers opened and the unit power transferred to offsite power (per plant design). A review of plant data indicated that the overvoltage condition on the plant boards lasted for approximately three and one half minutes.
After the trip the operators shut down unnecessary secondary equipment and dispatched personnel to verify the steam source had stopped once the turbine was tripped.
m 1638 A four hour phone call was made to the NRC pursuant to 10 CFR 50.72.b.2.ii as a result of the manual reactor trip.
,
&
_
.
.'
\\
'
APPENDIX C - POSTULATED INITIATING EVENT The most likely initiating event for the line break event was as follows:
(1)
Just prior to the time of the pipe failure, a routine surveillance test of the reactor building floor and equipment sump pump, containment isolation valve, 2-FCV-77-127, determined that it was inoperable due to its stroke time exceeding the technical specification requirements.
In accordance with the technical specifications (LCO 3.6.3) this condition required that the valve r
in series with it be shut and rendered inoperable.
,
(2)
An Assistant Shift Operations Supervisor (AS05), who is also a licensed reactor operator, was assigneo to remove the fuse (to be followed up with an independent verification by another operator)
to the solenoid valve.
The fuse is located in a DC cabinet in a room adjacent to the control room. The ASOS experienced some difficulty positively identifying the proper fuse since the cabinet is rather small and the proper fuse in one of many in a vertical string of fuses. He attached the fuse puller to what he believed was the proper fuse and moved it sideways slightly. This
fuse was, in fact, the power supply fuse for the No. 3 steam generator feedwater regulating valve trip system solenoid valve.
He then rechecked the fuse number, realized he had placed his fuse
-
pullers on the wrong fuse, slid the fuse back to its original position in the fuse clip (the fuse did not lose contact with the
fuse clip), and proceeded to remove the correct fuse.
,
(3)
Apparently as a result of the slight movement of the fuse, the No.
i 3 steam generator (S/G) feedwater regulating valve started to shut
,
resulting in a steam flow / feed flow mismatch and low level in S/G
!
No. 3 (The valve is designed to shut in 3 seconds as a result of a
'
trip signal.) A reactor operator took manual control of the S/G
,
level regulating controller and increased flow to correct the steam flow / feed flow mismatch.
(4)
A pressure spike of undetermined magnitude was introduced to the extraction steam line through the following sequence of events:
'
The partial interruption of feedwater flow, No 3 feedwater
,
regulator valve trip, caused a back-up in the No. 3 heater t
drain tank, raising the level of the tank above the level of the inlet lines, and forming a solid water column from the feedwater system to the secondary side of the feedwater heater.
(The 10" diameter inlets for the tank are near the mid-point of the tank, El 697'0"; and normal control level for the tank is between El 696'9" and 697'4".)
i
The operator's action to restore S/G level by manual control
.!
of the FW regulation system caused the feedwater regulator valve to open a sudden drop in pressure in the feedwater
line, resulting in void formation in the feedwater followed by void collapse.
,
,,.,
,, -,
-
--
- - -,
,
- - - - - - - -
.-...- - -
- -
-
.
.
....
.
'
'
.
'
.
.
..
l l
C-2
The resulting " water hammer" was not.large enough to cause
damage to a normal system, but was apparently a large enough
!
perturbation of the steam /feedwater system to cause a
'
pressure spike in the extraction steam line to the number 2 feedwater heaters, large enough to initiate the steam line failure.
i i
i
<
,
?
s
!
,
I
,
l
!
i e
h
--a------
--.. -
.
-
.-
-
- -
,,,
_
..
..
.
.-
.
-..-
~
-
-
' ~
..
..
l APPENDIX D - ACRONYMS AND INITIALISMS
i ABB Asea Brown Boveri
,
- AIT Augmented Inspection Team
,
ANS American Nuclear Standard
ASOS Assistant Shift Operations Supervisor
)
ASTM American Society for Testing of Materials l
AVO Assistant Unit Operator j
CHECMATE EPRI computer code for erosion / corrosion i
E/C erosion / corrosion El elevation
!
EPRI Electric Power Research Institute i
FAC flow assisted corrosion FSAR Final Safety Analysis Report
!
ISI Inservice Inspection l
IST Inservice Testing
LC0 Limiting Condition for Operation j
LPM Licensing Project Manager
-
kv kilo (thousand) volts
,
lbs/hr pounds per hour MOV motor operated valve
,
MOVATS MOV automated testing system i
MSR Moisture Separator Reheater
,
mvar milli var (var is an electrical unit of reactive power)
NDE nondestructive examination PI Periodic Instruction
psi pounds per. square inch
-!
RT radiographic test
S/G steam generator i
SNP Sequoyah Nuclear Plant
.!
SOS Shift Operations Supervisor
TS Technical Specifications i
TVA Tennessee Valley Authority
!
UT ultrasonic test v
volts i
{
WO Work Order
!
,
I
!
i i
<
$
r
[
l
+
,
_
.
.
. _. -
-
.
.
e
ATTAR.HMP4T 1
_
, #g#
g UNITED STATES NUCLEAR REGULATORY COMMisslON
.,
o REGION ll y
,. $
101 MARIETTA STREET N.W.
- t AT L ANT A, GEORGI A 30323
,~
%,.....#
!1 arch 3,1993 MEMORANDUM FOR:
Jerome J. Blake, Team Leader Sequoyah Nuclear Plant Augmented Inspection Team (AIT)
FROM:
Stewart D. Ebneter, Regional Administrator SUBJECT:
AIT Charter - Sequoyah Nuclear Plant Enclosed for your information is the AIT Charter which reflects instructions that you were given for the inspection of events associated with the Sequoyah Nuclear Plant Erosion / Corrosion Program and the electrical and secoadary plant transient which occurred on March 1, 1993. This Charter, prepared in accordance with NRC Incident Investigation Manual (NUREG 1303), reflects the needs of Regions II and NRR management. The objectives of the AIT are to communicate the facts surrounding these programs and events to Regional and Headquarters management, and to document the findings and conclusions of the onsite inspection.
If you have any questions regarding these objectives of the enclosed Charter, please do not hesitate to contact me or Albert F. Gibson of my staff.
g tz Stewart D. Ebneter Enclosure:
AIT Charter cc w/ encl:
J. Sniezek, EDO T. Murley, NRR S. Varga, NRR B. Grimes, NRR G. Lainas, NRR E. Jordan, AE0D L. Plisco, EDO l
9304130189 930407 PDR ADOCK 05000327
0 PDR
..
'
ATTACHMENT 1
.,
.$.
ENCLOSURE AUGMENTED INSPECTION TEAM CHARTER e
Develop a sequence of events associated with the electrical and secondary plant transients of March 1,1993 at the Sequoyah Nuclear Plant (SNP).
- Assess the efficacy of the SNP Erosion / Corrosion (E/C) program.
Ascertain the condition of SNP safety and non-safety related
piping susceptible to E/C.
- Evaluate the effect of the over-voltage transient on safety related equipment.
e Prepare a special inspection report documenting the results of the above activities within 30 days of the inspection completiion.
l