IR 05000369/1986004
| ML20198R872 | |
| Person / Time | |
|---|---|
| Site: | McGuire, Mcguire  |
| Issue date: | 03/17/1986 |
| From: | Brownlee V, William Orders NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML20198R831 | List: |
| References | |
| 50-369-86-04, 50-369-86-4, 50-370-86-04, 50-370-86-4, NUDOCS 8606100246 | |
| Download: ML20198R872 (9) | |
Text
7
. _ _ _.
_-
_ _ - _ _ _
____ _________ ___ ______________________ _ _______ - ____
.
n Kr?
UNITED STATES n
/
NUCLEAR REGULATORY COMMISSION o
[ -
REGION li
^
g'
j 101 M ARIETT A STREET, N.W.
- -
ATL ANTA, GEORGI A 30323
\\*.../
,
.-
Report Nos.:
50-369/86-04-and 50-3,70/86-04 Licensee:
Duke Power' Company 422 South Church Street-Charlotte, NC 28242
~
Docket Nos.:
50-369 and 50-370 License Nos.:
NPF-9 and NPF-17 Facility Name: McGuire 1 and 2 Inspection Conducted:
nuary 6 - February 28, 1986 Inspector:
(/~
Im 8 //
[
,
,
W. Ord r's', 5 lor sident y ector Odte 7i ned-Approved by: ! _ Lon.o YTILt %
3!/7 Ib
_.
Virgil (Ar6wnlee, Acting Section Chief Date Signed Division of Reactor Projects SUMMARY Scope:
This routine, unannounced inspection entailed 50 inspector-hours on site in the areas of operations safety verification, and event follow-up.
Results:
Of the areas inspection, one violation was identified.
,
P606100246 860602 PDR ADOCK 05000369
, _ - - _.
. _, _
___
_ _ - _
_
-
__
_. _
_
-
.
.
REPORT DETAILS 1.
Persons Contacted Licensee Employees
- T. McConnell, Plant Manager
- B. Travis, Superintendent of Operations D. Rains, Superintendent of Maintenance B. Hamilton, Superintendent of Technical Services i
L. Weaver, Superintendent of Administration
'
- M. Sample, Superintendent of Integrated Scheduling
- E. McCraw, License and Compliance Engineer
D. Mendezoff, License and Compliance Engineer D. Marquis, Performance Engineer Other licensee employees contacted included performance technicians, operators, and office personnel.
- Attended exit interview 2.
Exit Interview
.
The inspection scope and findings were summarized on January 31, 1986, with those persons indicated in paragraph 1 above. The licensee did not identify as proprietary any of the materials provided to or reviewed by the inspector during this inspection.
l 3.
Licensee Action on Previous Enforcement Matters Licensee action on previous enforcement issues was not inspected.
4.
Unresolved Items *
No unresolved items were identified during this inspection.
5.
Degraded ECCS Flow Path On the morning of November 2, 1985 both McGuire units were operating at 100%
power. At 6:40 a.m. that morning, the flexible discharge line of instrument air compressor B failed. The failed line resulted in the loss of instrument air on both units which in turn resulted in the main feedwater regulator
"An Unresolved item is a matter about which more information is required to determine whether it is acceptable or may involve a violation or deviation.
_-
__
. - _ - _.
.-.
-
.- -
.-
_
. _ _ _
- - -, - -
_. - -
.
'
.
valves closing on both units. As a result, reactor trips occurred on both units due to LO-LO Steam Generator Level, and a Safety Injection occurred on unit one. The Safety Injection on Unit I required the declaration of an Unusual Event which occurred at 7:25 a.m.
The resident inspector responded to the site to find both units being prepared for restart.
A review of the reactor Operators Logs, Shift Supervisors Logs, Technical Specification (TS) Logs, discussions with the operations staff, and a review of plant status did not reveal any items of concern.
,
It is important to note that no log entries relevant to valves 1NV-141, and 1NV-142 were made until 6.00 p.m.
that evening.
That log entry simply stated that the operator on valve INV-142 had failed, vas being replaced and the valve did not fall under any TS Limiting Condition for Operation (LCO).
A subsequent review of all required formal logs, revealed that there is no documented evidence that the operator on valve 1NV-141 had failed, nor that both valves were inoperable simultaneously.
Later that day, Unit I was restarted, entered Mode 2 and returned to Mode 3 to repair a leaking instrument fitting and a feedwater check valve.
The unit was ultimately placed on line on the morning of November 9.
Following trip recovery, Unit 2 was returned to power, reaching full power the following afternoon.
Subsequently, during the week of January 6,1986, a detailed review was performed of the circumstances surrounding the Safety Injection. Initiation of that review was prompted by a review of the Licensee Event Report (LER)
associated with the event and what appeared to be an inadequate resolution
'
of an alarm received during the Safety Injection.
Specifically, an alarm indicated that one group of heaters in pressurizer heater bank A was out of service for approximately 1.5 minutes during the transient. The inspector, in scrutinizing this area, found that no action had been taken to resolve the alarm. This was indicative, among other things, of an inadequate post
'
trip review. Details of the actions taken with respect to the heaters, are entailed in report 369/85-45.
During that review, the inspector determined that Unit I had apparently been operated.in specific modes with valves 1NV-141 and 1NV-142 inoperable and that during this period of operation these valves were required to be operable for isolation of the Volume Control Tank (VCT) in the event of a Safety Injection Signal.
This determination resulted from the review of associated work requests, shift engineers logs and discussions with personnel of several disciplines in an attempt to resolve the pressurizer heater issue. The inspector found in an operating engineers log (this is a non-required informal log) that both valves 1NV-141 and INV-142 had failed operator motors. This prompted the inspector to review the TSs, TS bases, system description, and the FSAR among other documents to determine the opcrability requirements for the valves.
The TSs do not clearly specify operability requirements.
Only
'
.
through understanding the design function of the valves, coupled with the requirements of TS 4.5.2(e) (which requires that every 18 months each automatic valve in the ECCS flow path be verified to actuate to its Safety Injection position on the appropriate signal) was the inspector able to conclude that the valves had been inoperable when they were required to be operable.
Following is the event scenario which was pieced together using information from several sources:
Event Scenario I
At 6:40 a.m. on November 2,1985, both valves INV-141 and INV-142 closed during the Safety Injection as designed.
When terminating the Safety Injection, operators were unable to open valves 1NV-141 and 1NV-142 from the control room, but succeeded in manually opening the valves locally, re-establishing the normal make-up flow path from the VCT.
By approximately 4:00 p.m. that afternoon, it was ascertained that the motor operator for INV-142 had failed.
The licensee subsequently stated that since the associated TSs did not list 1NV-142 specifically, the tie to TS 3.5.2(e) (Operable Flow path) was not made. Therefore, it was decided that the failure of INV-142 motor operator would not prevent restart.
By approximately 9:00 p.m. that evening, it was known that the operator on INV-141 had also failed.
The licensee later stated that based on the precedence from the 1NV-142 decision, it was erroneously determined that the INV-141 motor operator failure would not prevent restart.
At 6:15 a.m. on the morning of November 3, the unit was restarted. Reactor thermal power was maintained below 5% until 12:55 p.m.
awaiting the completion of ongoing maintenance. At that time, the unit was placed back in Mode 3 to facilitate secondary system leak repairs. The unit remained in Mode 3 with both valves inoperable until approximately 7:30 p.m.
on November 4, when the valves were returned to service.
Analysis Valves 1NV-141 and 1NV-142 are the isolation valves off the VCT supplying a common suction to the charging pumps of the Chemical and Volume Control (NV)
System. These valves receive a signal to close during a Safety Injection which isolates the VCT from the suctions of the NV pumps. The NV pumps also serve as the high head Safety Injection pumps.
This prevents the introduction of the VCT contents, which during operation is at much less than 2000 ppm boron, into the Safety Injection flow which is highly borated.
Further, with 1NV-141 and INV-142 incapable of closing, it is possible that the VCT would be depleted during a Safety Injection resulting in the introduction of hydrogen into the Safety Injection flow. Hydrogen, used as l
-
.
the pressure control medium in the VCT, if introduced into the NV pump suction could result in pump cavitation and possibly gas binding.
Preliminary calculations performed by the inspector relative to the static head available to the NV pumps suction are as follows:
Assumptions:
a.
INV-141 and 1NV-142 are open and electrically inoperable b.
VCT at normal operating pressure and level Pressure = 2
'sig
= approx. 4 ft.
Level
=
c.
VCT is a vertical tank approximately 8 ft. high sitting approximately 2 feet off the floor or the 733' elevation of the auxiliary building.
d.
RWST full Level = 40'
RWST elevation is 760'
e.
NV pumps both operable NV pump location is 716' elevation in the auxiliary building f.
Ignore all line losses Analysis:
RWST:
760' elevation 40' water in tank 800' head total-716' pump elevation T
84 pump head 84' x 0.434 psi = 36.46 psi pump head foot
VCT:
733' tank elevation
+
4' water
+
2' tank off floor 740' water head
- 716' pump elevation
'
24' water head 24' x 0.434 gsi = 10.42 psi foot 10.42 psi + 30 psi tank pressure = 40.42 psi l
Therefore, VCT head = 40.42 psi RWST head = 36.46 psi Difference = 3.96 psi Conclusion
- _
. -
.. _
_.
_.
__
... _. _ _ _ _ _. _ _, - - -
_
.
__
-
_
.
l It appears, at least initially, that the VCT would be the primary suction source at the outset of the event.
Subsequent licensee calculations confirmed those of the inspectors and I
indicated that the VCT would provide a pump suction head 3.77 psig above that provided by the Refueling Water Storage Tank (RWST). This confirmed that the VCT would have indeed supplied initial section to the NV pumps had there been another Safety Injection while valves INV-141 and 1NV-142 were inoperable.
The licensee subsequently performed a safety analysis to determine if operation with the two subject valves open during a small loss-of-coolant accident could render the NV pumps inoperable as the result of the VCT being depleted resulting ia gas binding. This analysis determined that hydrogen binding could occur 18 1/4 minutes after actuation of Safety Injection. The analysis was based on an initial VCT level of 50% and an initial pressure of 30 psi.
The licensee further stated that if hydrogen would enter the NV system and bind the NV pumps, the reactor could be shut down safely with the control rods.
However, there would be no make-up to the Reactor Coolant System until a lower reactor coolant pressure was reached to initiate the Safety Injection pumps (approximately 1500 psig discharge pressure).
SAFETY SIGNIFICANCE The safety significance of this event as previously stated, is the possibility of diluting Safety Injection flow with VCT contents and introducing hydrogen into the suctions of the NV pumps leading to cavitation and gas binding.
Herein lies the basis for the isolation signal sent to INV-141 and 1NV-142 on a Safety Injection, the requirement for operable ECCS and Boration flow paths and flow path valve operability.
TS 3.1.2.2 requires in modes 1, 2, 3 and 4, that at least two of the following three boron injection flow paths shall be OPERABLE; a.
The flow path from a boric acid tank via a boric acid transfer pump and a charging pump to the Reactor Coolant System, and b.
Two flow paths from the refueling water storage tank via charging pumps to the Reactor Coolant System.
The safety implications here concern the inability to initiate alternate emergency boration which requires the isolation of valves INV-141 and 1NV-142. Emergency boration may be required if the following symptoms were to exist:
Excessive control rod insertion Failure of two or more control rods to drop following a Reactor trip Unexplained or uncontrolled reactivity increase Inadequate shutdown margin
!
.
TS 3.5.2 requires in modes 1, 2, and 3 that two independent Emergency Core Cooling System (ECCS) subsystems shall be OPERABLE with each subsystem comprised of:
a.
One OPERABLE centrifugal charging pump, b.
One OPERABLE Safety Injection pump, c.
One OPERABLE RHR heat exchanger, d.
An OPERABLE flow path capable of taking suction from the Refueling Water Storage Tank on a Safety Injection signal and automatically transferring suction to the containment sump during the recirculation phase of operation.
Withoet restating, the concerns here deal with the inoperable flow path as described earlier.
TS 3.0.4 requires that entry into an OPERATIONAL MODE or other specified condition shall not be made unless the conditions for the LCO are met without reliance on provisions contained in the ACTION requirements.
TS 3.0.3 requires that when a LCO is not met, except as provided in the associated ACTION requirements, within I hour action shall be initiated to place the unit in a MODE in which the specification does not apply.
Contrary to the above requirements, McGuire Unit I was:
1)
Operated in modes 2 and 3 from 9:00 p.m., N. smber 2, until 7:30 p.m.
November 4, with 2 of the 3 boron injection flow paths required by TS 3.1.2.2 inoperable.
2)
Operated in modes 2 and 3 from 9:00 p.m. November 2, until 7:30 p.m.
November 4, with both ECCS subsystems required by TS 3.5.2 inoperable.
3)
Escalated into an operational mode in violation of the requirements of TS 3.0.4 when the LCO associated with TS 3.1.2.2 and TS 3.5.2 were not satisfied.
4)
Not placed in an operational mode where the requirements of TS 3.1.2.2 and TS 3.5.2 were not required, as required by TS 3.0.3.
The above described events singularly, and collectively, constitute significant violations of the referenced requirements.
It is considered that the NV system was seriously degraded, perhaps to the point that it would not have performed its intended safety function.
/
.
,
6.
Enforcement Conference An Enforcement Conference was held at Region II's request in the NRC Region II Office on February 28, 1986, to discuss the inoperability of the VCT isolation valves and the effect on the Chemical and Volume Control (NV)
System at McGuire Unit 1.
The following personnel were in attendance, a.
Duke Power Company G. E. Vaughn, General Manager Nuclear Stations T. L. McConnell, Station Manager, McGuire N. A. Rutherford, System Engineer, Licensing E. O. McCraw, Compliance Engineer, McGuire D. J. Bumgardner, Operating Engineer, McGuire R. B. Travis, Supt. of Operations, McGuire P. A. Goodwin, Engineer Associate, Safety Analysis Nuclear Regulatory Commission R. D. Walker, Acting Deputy Regional Administrator G. R. Jenkins, Director, Enforcement / Investigation Coordinator, Staff V. L. Brownlee, Chief, Reactor Projects Branch 3, DRP C. A. Julian, Chief, Operations Branch, DRS W. T. Orders, Senior Resident Inspector, McGuire P. A. Taylor, Reactor Inspector, DRS i
C. W. Burger, Project Engineer, DRP b.
Event Discussion The NRC Staff opened the discussion concerning the inoperability of the VCT isolation valves (1NV-141 and INV-142) and the effect on the NV system with the Region II perception of the event.
The NRC also expressed concern that the inoperability of these valves were not taken seriously enough to prevent restart of the Unit.
Duke Power Company (DPC) provided a description of the sequence of events, corrective i
action and a safety analysis. The meeting summary notes are described below.
The event details are discussed in item 5 of this report and the l
meeting handouts are attached.
(1) Sequence of Events During a November 2.
1985 event (loss of instrument air and resulting safety injection) the two VCT isolation valves were rendered inoperable.
Because these valves are not specifically addressed in the TSs the licensee did not consider their inoperable status prior to restart.
The valves were manually opened which is their normal operating position and the unit was taken to Mode 2 before a secondary leak inside containment brought the unit back to Mode 3 to repair the leak.
Both of the VCT isolation valves were repaired while the unit was in Mode 3.
!
.
---
,w.,,1--i-
.
g.
q q,a-m
>m vi
= - * - - - -
m
,.r--
-p.-,-
.. - - - - -
e--m----q-
-
v- - -+- --
--
--
,
I
"
.
.
(2) DPC Corrective Action The three items of corrective action taken by DPC are explained in the attached handout.
(3) Safety Significance In the event of a Safety Injection Signal these valves must close.
The closing of these valves isolates the VCT from the normal Safety Injection flow path which is to obtain suction from the RWST. The problem with obtaining suction from both the VCT and RWST is that if the VCT is depleted of water, the hydrogen cover gas could enter the system and cause cavitation of the pumps, thus the NV Safety Injection capability may subsequently be lost.
(4) Summary and Comments The NRC expressed concern that a total system interaction regarding these inoperable valves had not been performed prior to restart, that the total circumstances of this event had not been detected by DPC until brought to their attention by the NRC inspector and that the DPC training had not been indepth enough for this situation.
DPC stated that their training will be updated to cover this event, that DPC will redraft their corrective action to include
.
the comments heard at this Enforcement Conference and that DPC has learned a lot from this incident and have changed their philosophy.
.