ML021900652
| ML021900652 | |
| Person / Time | |
|---|---|
| Site: | Cook  |
| Issue date: | 07/09/2002 |
| From: | Grant G Division Reactor Projects III |
| To: | Bakken A American Electric Power Co |
| References | |
| EA-01-286 IR-01-017 | |
| Download: ML021900652 (8) | |
See also: IR 05000315/2001017
Text
July 9, 2002
Mr. A. C. Bakken III
Senior Vice President
Nuclear Generation Group
American Electric Power Company
500 Circle Drive
Buchanan MI 49107
SUBJECT:
D. C. COOK NUCLEAR POWER PLANT, UNITS 1 AND 2
NRC SPECIAL INSPECTION REPORT 50-315/2001-17(DRP);
50-316/01-17(DRP); PRELIMINARY YELLOW FINDING, JUNE 10, 2002
Dear Mr. Bakken:
This provides our response to your letter dated June 24, 2002, regarding the subject inspection
report and preliminary Yellow finding. You requested additional details from the NRC
concerning certain conclusions and assumptions referenced in the inspection report in order for
your staff to prepare for the regulatory conference. I understand that the regulatory conference
has been scheduled for July 25, 2002, and you have requested receipt of the additional details
by July 9, 2002.
My staff has completed a review of your request and developed the response enclosed with this
letter. With the exception of your request for the SPAR model and SAPPHIRE engine used in
our risk analysis, the additional details you requested are consistent with information provided
previously to your staff. My staff has discussed your request for the SPAR model and the
SAPPHIRE engine used in our risk analysis with the appropriate NRC Headquarters staff.
Based upon these discussions, we will provide the generic SPAR model for the D.C. Cook site
under separate cover. In addition, appropriate NRC staff will be available at the regulatory
conference to discuss risk insights that we gained through our use of the SPAR model and
SAPPHIRE engine.
A. Bakken
-2-
If you have need of any additional details or have further questions, please contact
David Passehl, Acting Branch Chief, at 630-829-9872.
Sincerely,
/RA/
Geoffrey. E. Grant, Director
Division of Reactor Projects
Docket Nos. 50-315; 50-316
Enclosure:
NRC Response to Request for Additional Information
cc w/encl:
J. Pollock, Site Vice President
M. Finissi, Plant Manager
R. Whale, Michigan Public Service Commission
Michigan Department of Environmental Quality
Emergency Management Division
MI Department of State Police
D. Lochbaum, Union of Concerned Scientists
See Previous Concurrences
DOCUMENT NAME: C:\\ORPCheckout\\FileNET\\ML021900652.WPD
To receive a copy of this document, indicate in the box: "C" = Copy without attachment/enclosure "E" = Copy with attachment/enclosure "N" = No copy
OFFICE
RIII
E RIII
E RIII
NAME
DPasshel/trn
KOBrien
GGrant
DATE
07/08/02
07/08/02
07/09/02
OFFICIAL RECORD COPY
1
ENCLOSURE
NRC Response to Request for Additional Information
Letter from A.C. Bakken, III, Indiana Michigan Power [DC Cook Plant]
dated June 24, 2002 (AEP:NRC:2609)
Reference:
D. C. Cook Nuclear Power Plant, Units 1 and 2
NRC Special Inspection Report 50-315/2001-17(DRP); 50-316/01-17(DRP);
Preliminary Yellow Finding, June 10, 2002
RESPONSE
Request 1:
A description of the Significance Determination Process results for any other
sequences determined to be greater than Green by the Nuclear Regulatory
Commission (NRC).
Response 1:
As discussed with your staff during the inspection effort and at the exit meeting,
the NRC staff tentatively concluded that the risk associated with the finding was
dominated by a dual unit loss of offsite power (DLOOP) initiating event. For this
initiating event, the station blackout sequences were the dominate contributor to
the overall change in core damage frequency. The station blackout condition
was developed, in part, due to the presence of a failed essential service water
pump discharge strainer and your practice of cross-connecting the emergency
service water units and trains.
Our risk assessment of the DLOOP initiating event and related sequences was
based upon the conditions observed during the events of August 29, 2001, your
staffs representations regarding the reliability of plant equipment, the plant and
electrical switchyard equipment configuration, the plants operating history, and
generic industry information.
While the inspection report did not discuss and we did not identify any other
greater-than-Green initiating events, scenarios, or sequences, please be advised
that the risk assessment results described in the subject inspection report were
preliminary. The results were based upon our understanding, as of the date the
inspection report was issued, of those factors which could impact the risk
assessment. Consistent with NRC policy, we will continue to re-evaluate our
preliminary findings, using relevant new or different information, until the
regulatory conference is held. Additional information which may affect the
preliminary risk assessment results would include the results of subsequent
inspections, recent plant-specific or industry events, and our discussions at the
Request 2:
The basis for the assumption that the inrush of water expected to occur
immediately after a dual unit LOOP event has sufficient energy and flow
velocities to cause local eddies and vertical water velocities sufficient to entrain
2
debris located in the previous quiescent flow areas of the intake structure. Refer
to page 29 of the inspection report.
Response 2:
The NRC staffs assumption that an inrush of water into the intake structure
would have sufficient energy and flow velocities to cause local eddies and
vertical water velocities sufficient to entrain debris was based upon plant design
and operating data; simplified calculations; and engineering judgement.
The NRC staff noted that the intake structure would experience an inrush of
water following a DLOOP initiating event due to differences between the intake
structure and lake water levels. The staff estimated that approximately
1.5 million gallons of water would be required to equalize the water levels. This
inrush volume was based on an intake structure water level, prior to the DLOOP,
of minus 12.6 feet, compared to the lake level, and an intake structure free area
of approximately 200,000 square feet (204' x 100'). Approximately 20% of the
intake structure volume was assumed occupied by structures and equipment.
Given plant design data, the intake tunnel water velocity was estimated to be
approximately 8.3 feet per second, assuming the center intake was isolated and
all seven circulating water pumps were running. Based upon an intake tunnel
diameter of 16 feet and the calculated fluid velocity, the incoming water flow was
determined to be in the fully turbulent flow regime. Turbulent flow is
characterized by the generation of eddies which have a random velocity and the
destruction of laminar flow lines. Additionally, the staff noted that a dissipation of
energy within the intake structure water volume would occur due to frictional
interaction, further enhancing the turbulent flow conditions and generating flow
eddies.
Your staffs calculations regarding the intake structure indicated that, with the
center intake closed, the normal velocity of the water passing through the
traveling screens, at one foot above the intake structure floor, was approximately
5.0 ft/sec. Following a DLOOP, the circulating water pumps would stop. The
flow through the circulating water pumps would rapidly slow and likely reverse
due to the high frictional flow losses through the condenser and the sudden,
extreme drop in pressure at the discharge of the circulating water pumps.
However, water flow into the intake structure would initially be expected to
continue along the same streamlines because of the momentum of the flow
stream, the absence of barriers with high frictional flow losses, and the continued
presence of a strong driving force (i.e., the low intake structure water level
relative to lake level). The flow would then be redirected perpendicular to the
initial flow direction by the loss of the outlet flow path through the circulating
water pumps. Because of conservation of mass and momentum, the redirected
flow will have a local velocity comparable (on the same order of magnitude) of
the initial flow velocity just prior to the DLOOP. The energy associated with the
redirected flow will then be dissipated by frictional forces, creating additional
eddies and a potential to entrain debris. As demonstrated during the emergency
service water degradation event of August 29, 2001, debris was typically located
in quiescent areas adjacent to the normal flow streamlines such as at the base of
3
the traveling screens and in front of the emergency service water pumps and
would be available for entrainment by these redirected flows.
The NRC staff assumed that the intake structure inrush would occur over an
approximate 1 minute time frame. The one minute time frame was based on an
initial circulating water flow of approximately 1.6 million gallons/minute
(7 circulating water pumps operating at 230,000 gallons per minute per pump)
and the volume of about 1.6 million gallons necessary to equalize intake
structure water level with lake level. This volume of water would represent less
than one-quarter of the total water volume contained in the 16 foot intake tunnels
(based on a 2000 foot intake tunnel length with the center intake isolated). Due
to the momentum of the intake tunnel flow, the inrush transient was expected to
be a fairly dynamic event, resulting in an initial intake structure water level
overshoot and a dampening oscillation until an equilibrium level was established.
Given the approximate 1 minute time assumed for the intake structure and lake
water levels to equalize, the staff calculated that the intake structure bulk
average vertical velocity would be approximately 0.18 feet/second immediately
following a DLOOP. This level of bulk average vertical velocity was greater than
the licensee calculated vertical velocity necessary to entrain and sustain sand
particles in a fluid flow.
In addition to the bulk average vertical velocity of the water, the staff considered
the potential for localized vertical velocities. Specifically, the staff considered
changes in the intake structure water velocity that would initially occur in the
vicinity of the circulating water pumps. These localized disturbances in the flow
profiles were similar to profiles observed when a fluid traveling with a horizontal
velocity enters an enclosure and is forced to change directions such as water
entering a lock and dam structure or a large tank from a pipe. Based on the
momentum of the incoming flow streams, the staff assumed that the initial
disturbances in the flow patterns would be localized to the east side of the
traveling screens. Perturbations in the water velocity profiles on the west side of
the intake structure were not considered reasonable until after flow changes,
originating near the circulating water pumps, following a stopping of the pumps,
worked back to the west side of the intake structure. Consequently, because the
volume east of the traveling screens represented approximately one third of the
intake structure water volume, the staff concluded that the average bulk vertical
velocity in the volume east of the traveling screens would be initially about three
times the overall average bulk vertical velocity. Therefore, in addition to
localized high velocity eddies, a bulk vertical velocity between the traveling
screens and the east wall of the intake structure of up to 0.54 feet per second
was considered possible. Licensee calculations indicated that a vertical
velocities of 0.30 feet per second were sufficient to entrain and transport sand
and shells.
Request 3:
The details from recent NRC studies indicating that the conditional probability of
large early release, given core damage, is approximately 0.82. Refer to page 27
of the inspection report.
4
Response:
The recent NRC studies, referenced in the inspection report and discussed with
your staff during the inspection and several associated meetings, were
documented in NUREG/CR-6427 [SAND99-2553], Assessment of the DCH
[Direct Containment Heating] Issue for Plants with Ice Condenser
Containments, April 2000. Table 4.21, Recommended DCH Containment Over
pressure Failure Probabilities For Extrapolation Evaluations Assuming A DCH
Event Occurs, on page 67, recommends a value of 0.82 for the conditional
probability of a large early release at the D.C. Cook plant.
Request 4:
The basis for using a large early release frequency value of 0.4, since the value
appears to exceed the maximum conditional probability value provided in
NUREG/CR-6595, An Approach for Estimating the Frequencies of Various
Containment Failure Modes and Bypass Events, dated January 1999. Refer to
page 27 of the inspection report.
Response 4:
The staff utilized engineering judgement in the development of the large early
release frequency value of 0.4 employed in our risk assessment. The value was
developed using guidance provided in NUREG/CR 6427, as referenced above,
and the staffs review of information supplied by your staff relative to their
development of a similar factor using the guidance of NUREG/CR-6595. Your
staff should be prepared to justify their basis for a large early release frequency
value other than 0.4.
Request 5:
The basis for and method used to combine the individual block evaluations into
a D/G common cause failure factor, including the final value reached. Also,
please provide a description of how the SPAR model was modified to account for
this factor. What failure modes were considered, a failure of individual
emergency diesel generators (EDG) in any combination or failure of the 4 EDGs
as a set? Refer to Page 27 of the inspection report.
Response 5:
The emergency diesel generator (D/G) common cause failure factor was
developed based upon NRC staff review of information provided in your staffs
analysis entitled, Debris Intrusion Into the Essential Service Water System
Probabilistic Evaluation, April 2002, and other related data. The common cause
failure factor was used as a direct adjustment to the SPAR model core damage
frequencies. Therefore, the SPAR model was not specifically modified to
account for this factor. The common cause failure factor was based upon a
common cause failure of all four D/Gs; therefore, individual or combinations of
individual D/G failures were not addressed. The SPAR model outputs were
focused on station blackout sequences; a condition that could only occur given a
failure of all four D/Gs.
The D/G common cause failure factor included inputs related to Blocks 2, 3, 4, 7,
and 8 of your staffs analysis. The NRC staff developed probabilities for these
blocks; specifically, 0.5, 1.0, 0.77, 0.25, and 0.25, respectively. The D/G
common cause failure factor for a DLOOP was 0.024, as discussed in the
inspection report. The common cause failure factor did not include inputs for
Blocks 1 and 9 of your staffs analysis due to these items having been previously
accounted for in the SPAR model results. Your staffs assumptions, used to
5
develop information associated with Blocks 5 and 6, could not be confirmed or
supported; therefore, the NRC staff did not include these items in development
of the D/G common cause failure factor.
Request 6:
The SPAR model and SAPPHIRE engine used to perform the risk analysis.
Response 6:
My staff has discussed your request for the SPAR model and the SAPPHIRE
engine used in our risk analysis with the appropriate NRC Headquarters staff.
Based upon these discussions, we will provide the generic SPAR model for the
DC Cook site under separate cover. In addition, appropriate staff will be
available at the regulatory conference to discuss risk insights that we gained
through our use of the SPAR model and SAPPHIRE engine.