ML20086G936
| ML20086G936 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 07/11/1995 |
| From: | Rosenblum R SOUTHERN CALIFORNIA EDISON CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9507170301 | |
| Download: ML20086G936 (6) | |
Text
-,
- _ - =.
2 fC 5" Southem Calitbmia Edison Ownpeny 23 PARKE R STREET (RVINE, CALIFORNIA 92748 I
f necwAno u. nosENotuu me mo=
i vie.
m.........
f July 11, 1995 i
4 i
U. S. Nuclear Regulatory Commission Document Control Desk i
Washington, D.C. 20555 Dear Sir-
Subject:
Docket Nos. 50-361 and 50-362 Reply to a Notice of Violation
..S.an. Onofre.Nucloar. C-enerating Station, Units-3 and-3
References:
(A) Letter, Mr. A. Bill Beach (USNRC) to Mr. Harold B. Ray (Edison), dated June 7, 1995 (B) Letter, Mr. R. P. Zimmerman (USNRC) to Mr. Kenneth P. Baskin (Edison), dated September 15, 1988 Reference A provided the results of the inspection (50-361/95-04 and 50-362/95-04) conducted from March 12 through April 22, 1995, I
at the San Onofre Nuclear Generating Station, Units 2 and 3.
The enclosure to reference A transmitted a Notice of Violation.
The Notice of Violation contained two violations: (A) failure to have adequate procedures, and (B) failure to take corrective actions to preclude continued failures of Agastat relays in safety-related components.
This letter provides Edison's reply to the subject Notice of Violation.
As discussed with Mr. Howard Wong, NRC Region IV, on June 14, 1995, the NOV Response due date was extended to July 14, 1995.
Violation A, the failure to have adequate procedures, contained two examples: (1) not including two of six available vent valves in the procedure used to vent the emergency core cooling (ECCS)
Train B suction header, and (2) inadequate leak verification for Chemical Volume and Control System (CVCS) filter replacement.
Both Violation A.1 and A.2 were caused by misjudgment by the respective procedure authors and reviewers.
i e ceq W
i fbU
/
g Ik\\
9507170301 950711
/
PDR ADOCK 05000361 l
Q PDR
1 y > kW L
~
Document Control Desk 2
For Violation A.1, the author and reviewers failed to recognize they needed to include two manual vent valves (MR029 and MR391) into the procedure attachment used to vent the ECCS Train B suction header.
The affected ECCS procedure was modified in 1990 to include an attachment for venting the ECCS suction headers based on past instances of incomplete pump venting.
However, the guidance provided to the author did not adequately state the procedure modification scope.
Thus, the modified procedure did not address all possible suction header vent points, but rather addressed the necessary venting sequence for the normal return to service of ECCS pumps (the vent valves mentioned above already existed in the pump return to service attachments).
In this case, in preparation for core reload during the Unit 2 Cycle 8 outage, the pump was initially returned to service, filled, vented, and successfully run from the refueling cavity (shutdown cooling suction line).
When the pump was later realigned for a pump J.ST us.ing the ECCS Train. B.suc.t. ion header. (the. r.efueling..
a...
water storage tank suction), it was not recognized that this flow path had not been properly vented by the ECCS suction header venting attachment.
For Violation A.2, the CVCS filter assembly was a new design feature installed in February 1995.
In developing the procedure for the new design filter replacement, the authors and reviewers did not recognize that the chosen leak dete tion method would not be able to detect gross leakage from the filter housing.
Their misjudgment was due, in part, to (1) the lack of experience in working with the new design filter replacement, and (2) focusing more on minimizing personnel exposure and the generation of radwaste.
[ Note:
Radiation fields in the vicinity of the filter housing could be in excess of 100 REM.)
As corrective actions, the personnel involved have been coached and are aware of their oversight, and, except for the ECCS venting procedure, the other applicable procedures were revised.
For Violation A.1, full compliance will be achieved upon issuance of an enhanced integrated ECCS venting procedure scheduled for completion by July 22, 1995.
For Violation A.2, full compliance was achieved on April 17, 1995, when the Operations procedure for returning F020 to service was modified to incorporate an improved method of identifying proper filter assembly integrity following filter replacement.
Violation B, is not considered by Edison to be a failure to take corrective actions related to the inability of Agastat relays to reliably meet Technical Specification (TS) setpoints.
As identified below, Edison has had an evolving action plan to address the TS requirements.
This plan was reviewed and accepted
{
by the NRC in 1988.
Edison's action plan to resolve this issue L
_ ~. _..
L Y*
,/
l Document Control Desk 3
i was pursued consistently, but with a relatively lower priority than other issues with greater safety significance.
The sequence of events in this evolution is recounted below:
The original 1970's plant design for the emergency diesel generator (EDG) load sequencing used Agastat time delay l
relays which have a tolerance of "+/- 10% of setting."
The original 1976 Unit 2/3 combined FSAR and TS submittal identified the load sequencing timersLas "+/- 10% of setting L
devices."
This was revised in Amendment 26 to the FSAR in l
1981 to read +/- 10% of its design interval.
The 1982 NRC l
l approved TS contained the more restrictive "+/- 10% of l
design interval" requirement (more consistent with standard l
TS).
l Basad.cn 4992-1983-Agastat surveillence-data, Edison 4
implemented an enhanced maintenance program to minimize out-of-tolerance test failures which resulted in limited improvement.
Because a more accurate time delay relay was not available, Edison concluded that the TS was inappropriately written and a TS change was necessary to
(
modify the existing time interval between each load block l
from "+/- 10% of design interval" to a tolerance which would meet the relay manufacturer's design specification acceptance criteria.
In 1983, Edison drafted and initiated internal review on a proposed TS change (PCN NF-10-95).
The proposed change initially addressed a "+/- 5% of design value" which could be supported by analyses.
During the internal review, it was determined that "+/- 10% of design setting" was required l
to ensure any potential load sequence overlap would be L
acceptable.
However, at that time, there were no dynamic
~
l analysis codes to support this change.
Therefore, PCN-95 l
was canceled.
1 The acceptability of Agastat time delay relay tolerances was subsequently raised as a concern in NRC Inspection Report (IR) 50-362/87-05, and was tracked by the NRC as Inspector Follow-up Item 50-362/87-05-01.
In response to the NRC's concern, Edison performed a formal evaluation to determine the safety significance and operability impact.
It was j
concluded that no safety issue existed since TS Engineered l ~:
Safety Feature (ESF) time response requirements remained within design limits.
'i 4
y-3 g-,
w e.
3
..,yy
,.om
. n.m y.
>m
,,,_..,r.-.we.#,m,-.,
,,,,,.,.%.,,,.-_.,-,,.-.,..,,...,,.y-
,-.+
,.-..,v.
?
Document Control Desk 4
j NRC IR 50-362/88-20, Reference B, reviewed the Agastat issue
=
in detail (NRC Follow-up Item 87-05-01), consulted with NRR, and concluded Edison's actions to be appropriate.
The NRC staff then closed the issue, stating:
"The inspector previously identified that several Agastat relay timers did not function within the required time interval during integrated ESF testing.
The licensee evaluated this condition and concluded that in certain instances the design tolerances for Agastat relay response time would allow the Agastat relay to function slightly outside of the Technical Specification required time interval.
The licensee currently readjusts Agastat relays that do not function within the Technical Specification required time interval, and action was being taken to identify an acceptable replacement relay for these applications.
Based on discussions with-NRR,- the inspector -
~
determined that the licensee's actions were appropriate.
This item is closed."
In 1989 ASEA and Agastat developed digital timing relays, but they did not meet Edison's requirements.
ASEA's relay was not 1-E qualified, and the Agastat relay had 25VDC contacts, not 125VDC as required at SONGS.
Edison was also concerned with introducing a common mode failure mechanism not previously analyzed with the use of digital relays, and therefore felt the cost and concerns did not warrant replacing the existing relays.
This common mode failure concern was later found to be valid as noted in NRC Information Notice 94-20, in which case Beaver Valley's EDGs were found to be inoperable because of the failure of their digital Agastat timer relays due to voltage spiking.
SONGS existing Agastat time delay relays are not susceptible to failure from the voltage spiking referenced in IN 94-20.
During the early 1990s, Edison focused on either designing
+
and building qualified replacement relays, or obtaining qualified relays that would meet Edison's design requirements.
Although these activities were pursued for several years, they were eventually abandoned due to business and legal considerations, with no acceptable relays being available at that time.
)
In November 1992, the NRC issued Information Notice 92-77.
To document Edison's evaluation of this notice and provide an operability assessment, Nonconformance Report (NCR) 93070031 was issued in July 1993.
Engineering performed analyses and electrical calculations with newly available PC based dynamic modeling software to demonstrate that no
Document Control Desk 5
undesirable interactions occur due to overlapping load groups.
These analyses demonstrated that the existing relays are adequate and do not require replacement.
The final disposition step of NCR 93070031 requires a TS change be submitted.
Note that this disposition step differs from Edison's previous 1987 evaluation which recommended that suitable replacement relays be evaluated when available.
It was not until recently that Edison was able to complete the needed Engineering analytical analyses to support a TS change instead of replacing the relays.
NCR 93070031 documented that the EDGs and ESF systems were not adversely affected by the time response, and that there was minimal safety significance.
At the time this issue was raised in March of 1995, the NCR disposition step to modify the TS had not been completed.
Based on the above, Edison believes the cover letter statement in Reference A which states, in part, "The second violation (Agastat relays),
is of concern because of Engineering's failure to resolve the matter resulting in the repeated failures to meet Technical Specification acceptance criteria....," is a mischaracterization of the circumstances.
Edison believes that the engineering response to this issue was technically correct and appropriate, and that the resources applied in this case were commensurate with the safety significance of the issue.
The significance of the Agastat issue was clearly established in 1987, acknowledged by both the Region and NRR staff, and the NRC staff concurred with Edison actions.
Determination that the same situation now constitutes a violation appears to be the imposition of a regulatory staff position that is either new or different from the previous (IR 50-362/88-20) staff position.
Since 1987 Edison has continued evaluating alternate relays, but no qualified relays were available until 1992.
Modifications required to use these alternate relays were not justified on either a safety or economic basis, and therefore Edison decided to change the TS.
Since 1993, Edison has had an effort in progress to develop a new TS to eliminate the conflict between the existing TS timing requirement and the actual capability of the existing time delay relays on a priority commensurate with other engineering and TS activities.
l Edison believes this apparent violation, as a minimum, meets the requirements of 10 CFR 2, Appendix C, to be noncited.
Edison requests Violation B to either be reclassified, as a noncited violation, or withdrawn.
While a somewhat more timely TS submittal could have been possible, the situation was identified by Edison, resolution efforts remained consistent with the NRC's
ymW Document Control Desk 6
earlier understanding-(Reference B, IR 50-362/88-20), there is no safety significance, and, as noted in Reference A, Edison will submit a TS change.
Edison plans to submit this change prior to the Unit 2 Cycle 9 refueling outage.
If you have any further questions, please-contact me.
Sincerely, f
FA i
'L."J.' Callan, Regiona1" dm'inist'~rator', NRC'Rigion IV
~
cc:
J. E. Dyer, Director, Division of Reactor Projects, NRC Region IV i
K. E. Perkins, Jr., Director, Walnut Creek Field Office, NRC Region IV J. A. Sloan, NRC Senior Resident Inspector, San Onofre Units 2& 3 M. B. Fields, NRC Project Manager, San Onofre Units 2 and 3 i
h i
i i
i l
i t
t
' ~~
1 i
k-l
.. ~..
_