Joint Subcommittees on Plant Operations and Fire Protection - June 28, 2001
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Subcommittees on Plant Operation and
Fire Protection Joint Meeting
Docket Number: (not applicable)
Location: Arlington, Texas
Date: Thursday, June 28, 2001
Work Order No.: NRC-298 Pages 1-265
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433 UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
JOINT MEETING OF THE ACRS
SUBCOMMITTEES ON PLANT OPERATIONS
AND FIRE PROTECTION
+ + + + +
THURSDAY, JUNE 28, 2001
+ + + + +
ARLINGTON, TEXAS
+ + + + +
The committee met at the Nuclear Regulatory
Commission, 611 Ryan Plaza Drive, at 8:30 a.m., Jack
Sieber, Chairman, presiding.
COMMITTEE MEMBERS PRESENT:
JACK SIEBER, Chairman
GEORGE APOSTOLAKIS, Member
DANA POWERS, Member
GRAHAM LEITCH, Member
ROBERT UHRIG, Member
ALSO PRESENT:
Dr. John Larkins, Executive Director, ACRS
Maggalean Weston, ACRS Staff
Howard Larson, ACRS Staff
Isabelle Schoenfeld, EDO Staff
Amarjit Singh
Pat Gwynn
Ken Brockman
Jeff Clark
Art Howell
Troy Pruett
Alberto Garcia, MIT
Eddie Horus Texas A&M University
Brandon Kennedy, Oklahoma Christian University
Brian Tindle, Oklahoma Christian University
Jeff Moreno
A-G-E-N-D-A
Opening Remarks. . . . . . . . . . . . . . . . . . 4
Region IV Organizational . . . . . . . . . . . . .17
Responsibilities/Accomplishments
Reactor Oversight Program Implementation . . . . .23
Senior Reactor Analyst Role in Risk. . . . . . . .92
Assessment Significance Determination
Process Implementation in Region IV
Plant Operations
Experience in IV . . . . . . . . . . . . . 133
Scam Trends. . . . . . . . . . . . . . . . 135
Callaway ALARA . . . . . . . . . . . . . . 146
Callaway Grid Experience . . . . . . . . . 173
Plant Experience in Region IV (Continued)
California Grid. . . . . . . . . . . . . . 193
Electrical Design and Operations . . . . . 194
Issues at Cooper
Fire Protection Experience in Region IV. . . . . 210
SONGS Electrical Fire
Region IV Responsibilities Under . . . . . . . 248
COOP/COG
Closing Remarks. . . . . . . . . . . . . . . . . 264
P-R-O-C-E-E-D-I-N-G-S
CHAIRMAN SIEBER: Good morning. This is
a public meeting of the ACRS and so we conduct it
under the rules published in the Federal Register, but
before we begin I'd like to thank Region IV
headquarters personnel for hosting this meeting.
These meetings are important to us, and
every year we try to go to at least once licensee and
one regional headquarters. This is intended to be a
two-way meeting, and we are very much interested in
your opinions, your candid opinions about how regional
operations are taking place, the problems that you
have, the successes that you're having, and what you
think the ACRS could or should do to help improve the
regulatory system not only at headquarters but also in
the regions.
So with that I would like to read our
formal statement to begin the meetings.
This is a meeting of the ACRS Joint
Subcommittees on plant operation and fire protection.
I'm Jack Sieber. I'm chairman of both subcommittees
for plant operations and fire protections at this
time. The ACRS members in attendance are George
Apostolakis, Dana Powers, Graham Leitch, and Robert
Uhrig. Also, Dr. Larkins, Maggalean Weston, and
Howard Larson from the ACRS and Isabelle Schoenfeld
from the EDO staff are present with us today.
The purpose of this meeting is for the
subcommittee to discuss Region IV activities and other
items of mutual interest, including significant
operating events and fire protection issues. The
subcommittee will gather information, analyze relevant
issues and facts, and formulate proposed positions and
actions as appropriate for deliberation by the full
committee.
Amarjit Singh is the Cognizant ACRS staff
engineer for this meeting. The rules for
participation in today's meeting have been announced
as part of the notice of this meeting previously
published in the Federal Register on June 11, 2001.
A transcript of this meeting is being kept and will be
made available as stated in the Federal Register
notice. It is requested that speakers first identify
themselves and speak with sufficient clarity and
volume so that they may be readily heard.
We have received no written comments or
requests for time to make oral statements from members
of the public, so we will now proceed with the
meeting. But before we do I'd like to have each of
the members and/or staff introduce themselves so you
get a feel as to who we are, what we have done, and
what our experience is.
And as I said before, my name is Jack
Sieber. My background is basically with utilities in
the Navy. I worked at -- I've been in this field for
40 years and have retired twice. The third time is a
charm. Shipping port, Beaver Valley, Perry, Surry,
North Anna 1 and LaSalle are plants that I worked at,
and I've been two years on the ACRS.
George.
MEMBER APOSTOLAKIS: Thank you, Jack.
I'm George Apostolakis, chairman of the
committee. I'm a professor at MIT, and the area of
interest to me is probably risk assessment.
MEMBER POWERS: I'm Dana Powers. I guess
I'm the old man here. I have seven years on the ACRS.
I was formerly chairman of the power protection
subcommittee. Now my current focus of interest are in
the areas of fuel and human factors.
MEMBER LEITCH: I'm Graham Leitch. I've
been on the ACRS for about six months, and my
background is primarily nuclear power plant
operations. I was the site vice president of Limerick
during the startup period, and later the vice
president at Nang Yaki.
MEMBER UHRIG: I'm Bob Uhrig. I'm a
professor at the University of Tennessee and also work
at Oak Ridge National Laboratory. Previously I spent
13 years with Florida Power and Light, where I was
vice president for advance systems and technology.
MR. LARKINS: I'm John Larkins, the
executive director for the Advisory Committee on
Reactor Safeguards and the Advisory Committee on
Nuclear Waste. My responsibility is to provide
administrative and technical support to the committee
in addition to a bunch of other things.
I know some of you -- I started as the
project director for Region IV in NRR, so somewhat
familiar with what you do. I've been with the agency
for 30 plus years and been in research, NRR,
chairman's office, OP, so I've been around for a
while.
I'd like to add to Jack's opening comments
our appreciation for Region IV hosting this meeting.
I realize it takes -- it does have a resource impact
and takes time to get prepared for these meetings, so
we certainly appreciate it, but it is a valuable part
of the committee's information gathering activities.
We hear a lot about programs being implemented in NRR
and other parts of the agency, and it's important for
the committee to see how these activities are actually
being carried out in the regions and other areas.
One of the key requests from the
commission this year is an assessment of the revised
reaction oversight program, so it will be useful for
us to hear your candid insights on that program and
other activities. And again, we appreciate your
hosting us here today.
MR. LARSON: I'm Howard Larson. I work
for John Larkins so that's why I was glad he talked
first. I'm special assistant for the ACRS and the
ACNW, so I work with both committees.
MS. SCHOENFELD: I'm Isabelle Schoenfeld,
16 years with NRC, four years with NRR, and 12 years
with research, and currently I'm working as a
coordinator -- the EDO's coordinator with ACRS and
ACNW and the Office of Research.
MR. SINGH: My name is Amarjit Singh. I'm
with the ACRS for the last seven years. Prior to that
I was NRR inspector here with Region IV.
MR. GWYNN: We're proud of the fact that
Jit helped us for quite some time in very important
areas, including fire protection, and he continues to
help the committee in outstanding fashion.
MEMBER POWERS: If you're responsible for
any of this training you're doing good.
MR. SINGH: Thank you, Pat.
MS. WESTON: I'm Maggalean Weston, senior
staff engineer for ACRS and responsible for the plant
operations subcommittee where I have South Texas
Project and the reactor oversight process. I'm
formerly with the tech specs branch and technical
assistance to the director of NRR.
MR. GWYNN: Chairman Sieber, would you
desire for us to provide background information about
our employees that are going to present? They are
just introductions.
CHAIRMAN SIEBER: I think it would be
helpful if we had a little bit of background.
MR. GWYNN: My name is Pat Gwynn. I'm the
deputy regional administrator for NRC Region IV, and
I'd like to welcome the committee to our offices.
We're pleased to have you back again.
I began my career in the nuclear arena in
1969 when I joined the United States Navy. I was a
reactor operator and electronics technician until I
went to Purdue University, got my bachelor's degree in
nuclear engineering and joined the Bettis Atomic Power
Laboratory where I worked for a period of time as a
Bettis physicist and test engineer.
After that I joined the Nuclear Regulatory
Commission in 1980. I was a resident and senior
resident inspector in Region III at Zimmer and at the
Clinton Power Stations. I joined the staff of
Chairman Lando Zech in 1987, where I served until
1989. During that period I had the distinct pleasure
of accompanying him and a group of 19 nuclear safety
government professionals who went to the former Soviet
Union and established a joint coordinating committee
on nuclear reactor safety. During that time I also
had the pleasure of working with John Larkins, and I'm
pleased to have John here with us today.
Since Chairman Zech's term expired I've
been assigned here in Region IV, first as a deputy
director of the Division of Reactor Projects and then
as director, Division of Reactor Safety, director
Division of Reactor Projects, and now as deputy
regional administrator.
I have with me today Ken Brockman, who's
the director of our Division of Reactor Projects, and
Ken is uniquely positioned to provide you insights
about the initial implementation of the NRC's Reactor
Oversight Program given that not only has he been
leading that program here in Region IV but he was also
an important member and contributor to the agency's
PACA panel, the IIEP that provided advice and
recommendations to the agency on that program.
Ken, would you like to give a little
background about yourself?
MR. BROCKMAN: Probably even more unique
about me is I'm not Navy. I'm a graduate of the
military academy at West Point, which puts me very
much in the club because I'm so much out of the club,
but I was eleven years in the military duty there, the
last part spent with Armor H Airborne in research and
development activities for weapons systems. When I
left the Army I went to work for Westinghouse, so not
only am I an Army person I'm Navy qualified on
reactors by working for Bettis Atomic Power
Laboratories.
I've got experience in the utilities side.
I worked for Detroit Edison Company during their final
stages of construction and initial startup as a member
of their management team, their training department
out there. I've been with the agency since 1984 at
Region II as a license examiner and as an inspector
out of that regional office. I was up at headquarters
for about five years, worked on the staff of EDO, was
a technical assistant for Chairman Selling.
I was also in charge of the incident
response organization up there now at the time they
built out the new facility, made the transfer, had the
opportunity to work with the Russian Federation and
the Ukranian Republic as part of our USA IDG7
initiatives in establishing emergency response
capability in those two countries, which many people
don't know that they had absolutely no nuclear
emergency response capability at all.
Then in Region IV now for six years in the
Division of Reactor Safety, and now as a director in
the Division of Reactor Projects.
MR. GWYNN: And to his right we have Jeff
Clark, who's our senior resident inspector at the
Cooper Nuclear Station. Jeff, would you like to give
a little background about yourself?
MR. CLARK: Sure. Good morning. I
started out my nuclear career -- nuclear Navy. I had
nine years active duty in the Nuclear Navy Program.
Subsequent to that I worked for 14 years for the
Baltimore Gas and Electric Company. There I was
maintenance supervision, planning and scheduling, and
my last functions at Baltimore Gas and Electric was as
a senior project engineer in capital improvements
area.
After that I joined the NRC in 1996. I
was in Region III. After a short period of time in
the Division of Reactor Safety I was the resident at
Perry, and I moved on from resident at Perry to the
senior resident at Cooper Nuclear Station in 1999. I
came on board there just about the same time that the
Revised Reactor Oversight Process was beginning, the
pilot process at Cooper, so what I'm planning to do
today is share some of those insights and dialog with
you on what those insights are from that perspective
of a pilot plant and going into the Revised Reactor
Oversight.
MR. GWYNN: To Jeff's right is Art Howell,
director of reactor safety in Region IV.
MR. HOWELL: Good morning. I also started
my career in the Nuclear Navy. I spent five years on
active duty nuclear powered submarine on the West
Coast, worked briefly at Rancho Seco Nuclear
Generating Station, which is near Sacramento,
California before it was permanently shut down.
Joined the NRC in 1985 in the former office of
inspection and enforcement, spent my time primarily
conducting safety system functional inspections, and
then also in the former office of AAOD performing
diagnostic evaluations before coming to the region in
1988.
And since that time I was a senior project
engineer, resident inspector at Comanche Peak Unit 1
during the startup testing of that unit, section chief
in the Division of Reactor Projects for South Texas
Project in Wolf Creek, and also the deputy directors
of both the divisions of reactor safety and projects,
and then for the last four years the Division of
Reactor Safety.
I too, like Ken, have spent a lot of time
working with the Russians and Ukrainians with respect
to the Lisbon Nuclear Safety Initiative. I was a co-
team leader with some Russian counterparts at a fairly
extensive team inspection at the Balakovo Nuclear
Power Plant in 1995, and we've done a lot of work in
hosting Russian and Ukranian regulators in this region
over the years in both divisions, and I'm going to be
sharing with you our experiences with respect to the
new fire protection inspection program as well as some
risk insights and how we incorporate risk into day to
day regional operations.
Thank you.
MR. GWYNN: On my left is Mr. Troy Pruett,
who is one of our senior reactor analysts here in
Region IV.
Troy.
MR. PRUETT: Good morning. My name's Troy
Pruett. I'm a senior reactor analyst.
I started off in the Nuclear Navy as well.
I was an enlisted plant operator and staff instructor
at the New York prototypes. After leaving the Navy I
went to work at D.C. Cook as an instructor in their
training department, and then joined the NRC in 1992
as a health physicist inspector in Region V in the
materials group.
With the consolidation of Region V and IV
I took a slot as a resident inspector at Waterford,
spent three years down there, took a senior resident
slot at the Clinton Power Plant in Illinois, and once
we got them back on line I decided I needed to go back
to a warmer climate and took the senior resident slot
at the River Bend Station, and I was done there for
about two years and I'm currently filling the senior
reactor analyst slot now.
MR. GWYNN: Thank you, Troy.
We have a number of other staff members
that will be making presentations throughout the day,
and I think that we need to move forward with our
presentation. However, I would like to recognize five
special people that we have in the room today.
Alberto Garcia is with us from the Massachusetts
Institute of Technology, Eddie Horus from Texas A&M
University, Brandon Kennedy and Brian Tindle, both
from Oklahoma Christian University, and Jeff Moreno
from Oklahoma State University. They are five
engineering associates who are working in our offices
this summer and learning about the NRC, and they're
here for training purposes.
Welcome, this morning.
I also wanted to express the regrets of
our regional administrator, Mr. Merschoff. He
unfortunately was unable to be here today. I'm sure
you're aware that the agency's first meeting of the
agency action review is being undertaken right now in
Atlanta, Georgia, and for that reason he was unable to
be here. He recalls that the last time you were here
that was his first year in Region IV, and he also was
unable to attend, and --
MEMBER POWERS: I hope that everyone
congratulates him on his presidential award for
meritorious service to the agency.
MR. GWYNN: Thank you. I'll pass that
along to him.
I believe we have an interesting agenda
today, and in addition we have arranged for some of
the best Texas barbecue to be served at lunch, and
that will give us an opportunity perhaps to have some
more informal discussions, and we've asked additional
members of the Region IV management team and the staff
to come and join us for that luncheon.
Does everybody have a copy of my handout,
because you can see the colors from the handout, and
I'll be referring to the colors.
The Region IV organization is consistent
with the organizational structure found in the other
three regional offices of the Nuclear Regulatory
Commission. The only major differences are the lack
of deputy division directors in two of the three
technical divisions, and that difference exists
because of our relatively small size.
At the top of the organization chart
you'll see Mr. Merschoff and myself, the regional
administrator and his deputy. We're responsible for
the day to day operation of the region, which includes
this office, 14 resident inspector offices,
approximately 160 staff members, and a budget of about
$4.3 million this year. The majority of our budget
goes to office rent and travel expenses, but this year
there's a substantial additional amount in our budget
to provide for the upgrading of our incident response
center for continuity of operations and continuity of
government functions, and Mr. Andrews, our emergency
response coordinator, will talk a little bit more
about that this afternoon.
To the left of Mr. Merschoff is a dotted
line going to Mr. Lynn Williamson, who's the director
of the Office of Investigation field office that's co-
located with us here in Arlington, Texas. The Office
of Investigation's field office is responsible for
investigating allegations of wrongdoing by NRC
licensed entities and their contractors.
The gray boxes below myself and Mr.
Merschoff are the regional administrator staff
including our allegation coordination and enforcement
staff, our emergency response coordinator, our state
liaison officer, our regional counsel, and our public
affairs officer, who actually reports to the Office of
Public Affairs in headquarters, Mr. Bill Beeacher.
From time to time some of the regional
administrator staff members will be joining us today,
and right now Mr. Charles Hackney, our state liaison
officer, is sitting behind you, and Mr. Breck
Henderson, who's our public affairs officer, is also
here in the room.
We have three technical safety divisions
represented by the blue, green, and yellow boxes that
you see below the regional administrator's staff. Two
of these divisions, the Division of Reactor Projects
and the Division of Reactor Safety, are involved in
the implementation of NRC's power reactor inspection
program. The Division of Reactor Projects or DRP is
composed of the resident inspector's staff, their
supervisors, and regional support functions. They are
the eyes and ears of the NRC at every operating
nuclear reactor in the region.
The resident inspectors are generalists
who live in the vicinity of their assigned plants.
They monitor the overall safe operation of their
assigned facilities. They're the first to respond to
events at the plant, and they are the primary NRC
spokesman for the NRC in the local community.
The Division of Reactor Safety or DRS is
composed of specialists, inspectors, and reactor
operator license examiners that are all based here in
Arlington. They include specialists in plant
operations, maintenance, physical security, radiation
protection, emergency preparedness, and engineering
disciplines to name a few. These inspectors travel to
all of the power reactors in the region performing
scheduled inspections in their areas of expertise.
Mr. Brockman will talk more about the
implementation of our power reactor inspection program
in a few minutes.
The Division of Nuclear Materials Safety,
or DNMS, which is in the yellow, is composed of
inspectors and license reviewers who implement all
aspects of NRC's nuclear materials licensing and
inspection program within the region except for those
licensing and inspection activities that are
specifically delegated to the states that have
agreement state programs. Those agreement state
programs are overseen by two agreement state officers
that report to the director, Division of Nuclear
Material Safety.
DNMS licenses and inspects nuclear
medicine programs in hospitals, radiographers, nuclear
gate users, and well loggers. they also inspect
uranium mines and mills, a fuel cycle facility, and
power reactor independent spent fuel storage and
decommissioning activities within the region. The
materials inspectors in Region IV have a particularly
large challenge, since even though they're only on the
order of 625 materials licenses and 25 uranium
recovery facilities they're spread over large
distances, including the North Slope of Alaska and
Guam in the Western Pacific.
Finally, our Division of Resource
Management and Administration, or DRMA, which is shown
in the pink, is the administrative unit supporting our
technical safety mission. They handle such activities
as travel, budget, human resources, mail, information
technology support, and a host of other service
functions that keep the technical safety organizations
functioning smoothly, and we're proud of the high
level of service that our DRMA organization provides
to our inspection and licensing staff.
We have a very large region
geographically, as you will see on my next slide. Our
travel office issues more airline tickets than any
other NRC region and almost as many as our
headquarters offices. Kathleen Hamill, who's the
director of the Division of Resource Management
Administration, is here in the room with us today.
The next slide, which is my last slide,
depicts Region IV. It identifies the 21 states in the
region and the location of the 21 power reactors and
the 14 power reactor sites in Region IV. You'll
notice that two of our power reactor sites, the
Callaway Plant in Missouri and the Grand Gulf Plant in
Mississippi, are physically located in states where
the use of nuclear materials is regulated by a
different NRC region. This action was taken in 1994
as we consolidated NRC Regions IV and V to more evenly
distribute the power reactor inspection work load
across the regions and to place all the plants that
were then operated by Entergy Operations Incorporated
in a single NRC region.
If you look at the map that's in front of
you you'll see a purple triangle in Missouri. That's
Callaway. And a purple triangle in Mississippi, and
that's Grand Gulf. Grand Gulf is one of the four
Entergy plants that are located in NRC Region IV.
This slide also shows that 15 of the 21
states in the region are agreement states. The dark
purple and the middle purple shades are the agreement
states in Region IV. Notice that both Alaska and
Hawaii as well as the Pacific Trust territories are
included in the six states that are not agreement
states in Region IV, and those are the lightest shaded
states on the map.
What the map doesn't show clearly is the
important work we in Region IV are doing to bring a
higher level of radiation safety to work being
performed on offshore oil platforms and on pipeline
barges in federal waters in the Gulf of Mexico. It
also doesn't make clear that our regulatory arms reach
to Johnston Atoll and Guam located on either side of
the International Date Line. As a result of this
circumstance we were able to state on December 1, 1999
that Y2K both began and ended in Region IV.
With that, I'm prepared to answer any
questions that you have about the region overall
before we go to the next presentation.
(No response.)
MR. GWYNN: If there are no questions I'll
turn it over to Ken Brockman, the director, Division
Reactor Projects.
Ken.
MR. BROCKMAN: Thank you very much, Pat.
I have a strange feeling that I won't be quite as
lucky on the lack of questions in my presentation.
I'm passing around a set of slides I copied for
everyone.
Over the next 45 minutes or so I'm hoping
to have a very -- an opportunity for a good
interactive discussion as to the insights that we've
seen in Region IV with respect to the revised
oversight process and also the insights that we've
been able to gain from it. As Pat mentioned earlier,
we've been very active over the last 18 months in the
process. I've been a member of the pilot program
evaluation panel and the implementation evaluation
panel, which has given me an appreciation for FACA
rules that I did not previously have.
And Jeff has been involved with it since
the very beginning, as he has said. The presentation
that we're going to give you is basically going along
these lines where we're going to talk about the
process overview. We'll go with the time line as to
how it's proceeded, inspection assessment process, how
it's worked in the region, the insights we've got from
there, specifically the results that we've seen in
Region IV, and how we think that that has rolled into
our assessment of licensee performance.
Is the process working? Does it appear to
be getting us to the places? Does the gut match what
your head says with respect to this. Certainly
conclusions at the end. We've got questions and
answers listed at the end. I would encourage I think
however that at any time you've got something that you
want to interject to keep the presentation more free
flowing as opposed to in that manner. We have the
capability to fill up any block of time that we are
given with the presentation, and that may not get to
all your needs, so feel free to interrupt.
MEMBER POWERS: Ken, you're not going to
discuss the significance of the determination process?
MR. BROCKMAN: No. Per se, we would
discuss it only that we go through it. I think with
the SRAs and what have you we've got that -- a more
in-depth discussion on that later on. Some of the
successes of it, some of the challenges of it. We will
be sharing -- generally has it worked with an example
there, but not the details for this presentation.
Okay. We'll go with our next slide then,
and I'm probably going to start off with my old
teaching type of philosophy with the infamous
rhetorical question, do we need to go through a
discussion of the ROP process: performance
indicators, inspection findings, how they come
together. Would that be of benefit as a refresher to
everyone or is everyone here fairly familiar with
that?
MR. LARKINS: I think we can go fairly
expeditiously --
MR. BROCKMAN: Okay. Then we'll really
cover -- at the 30,000 foot level. New program,
performance indicators provided by the licensees in
several different areas, inspection still an essential
part of the program. We can't forget how that's come
together. We have baseline inspection similar to the
previous concept of a core inspection. Now there are
criteria by when you would either do supplemental
inspection based upon performance deficiencies. That
can escalate in its level, be a low performance issue,
be a higher -- be a very significant type of
supplemental inspection.
MEMBER POWERS: The first question that
comes up in this comparison between core and baseline
is that now the region's locked into a baseline
whereas in the past they could adjust for a round in
response to the needs of particular sites.
MR. BROCKMAN: We can flip back to our
member of ours -- and let me refer you to a chart
that's further within your packet.
MEMBER POWERS: If we're going to get to
it I can wait.
MR. BROCKMAN: I'll get there. Yes,
without a doubt the new program still allows us the
capability to respond to changes in performance. It's
just a criteria or a little more defined now, more
predictable than they used to be. That's one of the
insights that we have seen is anything that we have
felt we need to inspect we can get to.
MEMBER POWERS: Well, you know, when give
him a licensee, is this, what -- under the old
program, I was doing good and I had X number of
inspection hours, and I haven't really changed and now
I've got X plus delta inspection hours. I'm getting
more inspections under this, and my performance is
about the same.
MR. GWYNN: I'd like -- a few things on
this subject, because this was one of my concerns when
we first proposed having this new program, and it's an
interesting result. But under the core program we had
a minimum inspection program that we did at every
facility. That was the core. We had core
inspections, regional initiative inspections, and
reactive inspections, and we couldn't change the core,
so the baseline is like the core but the baseline
includes all of the inspection that we plan to do at
the facility, whereas the regional initiatives -- some
of that was planned. Some of it was added as a result
of performance insights that occurred during the
assessment period, and of course reactive inspection
only took place as a result of events.
And so for licensees that were high
performing licensees under the core inspection
program, that got very little regional initiative
inspection and essentially no reactive inspection
because there were no events at their plants, and as
a result they essentially got the core inspection
program.
Now we in Region IV had a relatively high
number of plants that were performing at a high level,
and as a result the majority of the plants in Region
IV were on core or reduced inspection programs, and so
when the baseline inspection program began its
implementation here they did experience an increase in
the total number of inspection hours. But as you can
see from Ken's chart, the increases weren't that
great.
MEMBER POWERS: The problem I see is that
when they put in this new reactor oversight they
didn't say, Tom, here's 16 more FTEs to help you carry
out this additional inspection. I'm very certain they
didn't do that. So it looks to me like you must have
the same problem that the licensee is facing in that
you did have a lot of high performing plants. Now
you're doing more inspections with the same number of
people. Something's got to give some place. What's
giving?
MR. BROCKMAN: It's a good insight, and we
might as well -- I'm going to stay free flowing in the
presentation, so you've got this chart in your package
about two-thirds of the way back.
What you can see off this chart right here
is a look at -- right here is the last year -- this
light colored bar -- it's the last year of the old
program. Now, that's not the year right before the
new one, because that was a transitional year. I've
gone back to '99 when the old program was solid in its
implementation and then compared that with the dark
line against the first year of the new program.
You're going to see some a little more, some a little
less.
Why is the variance in the different
plants? Remember, we've got some procedures -- big
team inspections that are done biannually. Some are
done triennially. So the first year you haven't
gotten all of the program done anywhere, and we
haven't tried to normalize the data here. So you're
getting the actual raw data that was conducted, and
you can see, some above, some below.
Now --
CHAIRMAN SIEBER: I think that question
then needs to be extended a little further because if
you increase the baseline inspection basically for all
plants then reactive investigatory inspections have to
decline because you have fixed manpower, and because
of that do you feel that you lose some versatility for
those plants that don't perform as well as the average
plant to gain appropriate insights into the failures
of that plant?
MR. BROCKMAN: What's happened because
of -- we have to visually try to capture this a little
bit. We had several plants before. We had everybody
who was all South Point, and they'd get a small
amount of inspection. Then we had those who may have
had three 1s and a 2, two 2s and two 1s. What we've
done now is about everything from three 2s and a 1 on
up have been all brought together with the new
criteria to where you're at. That's about the number
of plants we're talking about. Right now we've got
about 85 plants in America who are all in the all
green arena, the licensee response arena.
Therefore, the amount of inspection that
you need to have to maintain your comfort that that
performance level is now based on the lowest person of
that 85, not the highest person of that 85 -- my
gradations are different now. That's why plants that
were very good performers are now seeing more. My
inspection program was verified with comfort the lower
level of performance. That addresses I think the
utilities issue as to why they're seeing more
inspection.
What they're seeing less of is less
regional initiative. I've got an itch that needs to
get scratched. Everybody's getting that itch scratched
on a baseline now in that aspect of verifying, so have
I lost that flexibility? No. That flexibility is now
built into the baseline program.
Your reactive question is a superb
question. It was one of my big concerns going in
there is our capability to respond to events as they
arise. We're going to talk about a couple of those
and where they've gone. The criteria now are very
much more prescribed. Management directive 8.3
certainly gives us definitive criteria at which time
you start considering a special inspection, an AIT, an
augmented inspection team, an incident investigation
team. We use those criteria and they're based on
risk -- as an entry point into the decision-making
process.
We've got overlap where deterministic --
your gut comes into play on it -- yes, I could. No,
I couldn't -- so we've got some overlap. The way I
describe it is PRA number gets me to the ballpark and
then my gut tells me what position I'm going to play
out there, whether I go or not.
So we put that together and what we've
been able to find now is under the baseline program if
I have an event that occurs -- we're going to talk
about two events today. If we've got an activity that
goes on there is a baseline module called event
response that I go out there with, and the purpose of
that module is to identify what is the risk
significance of this occurrence? Get me to the
ballpark. Am I at the ballpark, am I not at the
ballpark?
And then I can use one of two options to
inspect -- or one of three options to inspect it. A,
I can pass. Risk number didn't get me to the
ballpark. It's not worth the investment of the
issues. I will follow up. I leave it in the
licensee's domain and I will follow up with problem
identification and resolution inspection later on to
see did -- verify that they addressed it properly.
That's one option.
The second option I have is the other end
of the spectrum. I'm there. It requires a special
type of inspection, so that's inspection AIT, IIT.
The instincts are there and we will, based upon the
risk insights, the deterministic insights, we will
launch a unique activity outside the baseline program
to do that.
The third option that you have then is I
am going to use this to define the samples that I want
to do under the baseline program. I have identified
a risk significant sample set. It's time -- I'm
supposed to evaluate emergent work activities. Well,
I have a potential transformer that has exploded that
doesn't have a risk number, but boy the licensee's
scrambling about. They're doing things that have
impact on the plant operations. How are they dealing
with it? It's a wonderfully appropriate sample to be
using right now, and the insight gets me there, and
the baseline program lets me inspect that in a real
time method.
As I said, we have not found a thing that
we want to inspect that one of these three legs of the
program will not let us get to. We've been able to go
out and inspect everything we want. One of the
insights we do have with respect to resources though
is they are very tight. We have our people scheduled
out to the week, and Art's impacted by this even more
than I -- 18 months in advance. We know when our
people's leaves are going to be taken.
MEMBER POWERS: I don't understand whether
that's an acceptable situation. That really does
impact your flexibility.
MR. BROCKMAN: One of the lessons I think
we learned nationally is in Region I with IP2. The
initial estimate for an activity -- if you get an
activity that turns up red and goes into our large
scale supplemental inspection, the 95003 inspection,
I think they would tell you the initial resource
estimates associated with that were not nearly what it
winds up becoming.
MEMBER POWERS: It expands like --
MR. BROCKMAN: We have been blessed in
that we haven't been challenged with one of those
activities. We would really have to do some
significant resource decisions with respect to what
we've got to do. We've been challenged with a couple
of things ANO this year. I had -- in one year I've
got the new program, steam generator replacements, and
license renewal. Steam generator replacements and
license renewal are not part of the baseline
inspection program.
Now, many of the activities that went on
as part of our inspection for those things were
appropriate risk informed samples to put into the
baseline program. They're doing plant modification --
major plant modification going on with steam generator
replacement. What better modification to look at
during this year's inspection than the replacement of
steam generators? I gain great insights there. I can
take credit for that under the baseline inspection
program while we're inspecting the steam generator
replacements. This makes sense.
Were we type at ANO? Yes. We're type.
One of the insights I've seen is here in the regional
office I have two project engineers which support each
one of my branches. Their inspection time is fully up
to in the neighborhood of 30 percent on the road
inspection time. Every region-based inspection -- we
don't call them a DRS inspection, a DRP inspection.
DRS and DRP share the inspection program. Some of the
modules are resident based. Some of them are region
based. The region based inspection -- many DRP people
support those.
We have a schedule worked out where I've
got a resident who is leased on one region-based
inspection a year. Every resident is. Every one of
my project engineers are. So you have these
scheduling dilemmas much more a part of the branch
chief's job, and they schedule those much further out
than they did in the past.
MR. GWYNN: I have a couple of comments
that I'd like to make.
One of the major thrusts of the new
inspection program was to provide consistency across
all licensees and across all regions, and I think that
goal has been advanced substantially by the new
baseline program. Ken used the term if we have an
itch that needs to be scratched. That's now the
agency's itch. When I was leading the Division of
Reactor Projects if we saw an area that we thought
needed to be looked at more closely across the entire
fleet of plants in our region we would go and do that.
But the agency wouldn't do that, and so three other
regions didn't receive that inspection.
Now those decisions are made nationally
and if in fact that itch needs to be scratched it's
scratched at every plant in the country, and I think
that's a significant improvement in the conduct of our
inspection program.
We had a different threshold for event
response. Now if the licensee has a good corrective
program and they're in the licensee response band we
typically don't respond to a low-level events that
occur at their plants. And so the things that we were
doing in the past we're not doing now that were unique
to this region, but we're applying additional
resources at plants in areas that have been deemed by
the agency to be of risk significance, and as a result
of that we've had some excellent findings that we
would not have achieved under the previous inspection
program, and that's focused attention for all of the
utilities in the countries in areas that it hasn't
been focused in before.
so I think that the new program has
brought a lot of value to the agency and has advanced
a number of goals, including the goal of consistency
across the regions.
CHAIRMAN SIEBER: I'm going to ask another
question which probably will take you beyond where you
are in your talk right now, and if that's the case
then just remember it and when you get there you can
address it. But we are about to introduce as an
agency the performance indicators, and it's purported
that these performance indicators will allow a
reduction in baseline inspections.
Do you feel that there is an equivalency
between performance indicators and reductions in
inspections such that the combination of the two will
result in an adequate regulatory program, or do you
have other views? And you can address this now or
later on.
MR. BROCKMAN: You've looked at my
presentation notes. Bear with me. That's a major
topic we're going to talk about in just a couple of
minutes.
CHAIRMAN SIEBER: All right.
MR. BROCKMAN: It's a great segue. Let's
move -- everybody understands how we're organized now
under cornerstones, that concept, cornerstones come
together under reactor safety, radiation safety, or a
safeguards application. Performance indicators feed
a cornerstone. Inspection findings feed a
cornerstone.
And, Jack, we will be getting to bring
those together.
Let's very quickly move to the time line
that we're talking about so everybody is together
there. The pilot program for the ROP started in June
of '99. There were feedback lessons learned
associated with that commission meeting on that. SECY
paper went up and what have you. We implemented the
initial year on April 2, 2000. That went on for a 12-
month period. We've changed our basic planning cycle
now to an annual planning cycle as opposed to the old
South methodology, which was 18 plus or minus your
comfort factor.
And that's -- another point Pat brought
up, to be consistent. We are now it looks like going
to transition and get that annual cycle on a calendar
year basis. That's one of the things you'll see
coming up -- a recommendation is to right now play the
next nine months as another transitional period and
get this on a calendar basis. That's an efficiency
issue with respect to the agency to be able to do
that. So there's the basic time frames we're talking
about.
If you'll look at the next slide we've got
here real quick you can see in the initial year the
pilot program -- there are the sites that were
involved in the pilot program. In Region IV that was
the Fort Calhoun Station and the Cooper Station, and
as we've mentioned Jeff was the senior resident
through all of that. He's been one of my key people
who's been involved as we have made that transition.
What we're going to do now is talk about
out of this -- and we're going to start moving, Jack,
right to where you want to go.
The next slide takes us to the end of the
first year. Where are we? What has this program told
us? This is off the web page. It's currently there
right now. The column on the left is the licensee
response column, and there is about 85 plants that are
in that column --
MR. CLARK: This chart would actually
continue down. This is just a representative --
MR. BROCKMAN: Yes. But even though a lot
of information that's been heard is the performance
indicators, the findings, we've only gotten 2 percent
of the performance indicators that are not green.
When they come together, when the synergism of the
process comes together if you look at the regulatory
response column --
MEMBER APOSTOLAKIS: These columns are
from the action matrix. Right?
MR. BROCKMAN: This is what comes out of
the action matrix. This is what differentiates the
performance that we've got now. This is equivalent to
the old south in the aspect of here's your ones with
a couple of twos. The next one -- here's the ones
that probably got a three or so in there, and there is
no correlation. I'm just trying to give you a visual
picture of where it goes. So even though the
individual data has 5 percent of the performance
indicators, 5 percent of the findings aren't white.
When you put them together you get a differentiation
of performance on plants.
And in fact it's greater than 5 percent.
We've got 15 plants out of 103 that are in the
regulatory response column, three in the degraded
cornerstone column, one in the multiple repetitive
degraded cornerstone column, each one of these being
a more significant level of performance deficiencies.
MEMBER POWERS: I guess I agree with you
that if you'd asked me before this matrix was done
about what the distribution would be this is about the
distribution we would have thought. Right?
MR. BROCKMAN: It's probably not far off.
MEMBER POWERS: Maybe one or two were up
in the multiple response region, but not many more in
the regulatory response.
MR. BROCKMAN: No. That's -- there may
even be a couple more here than we'd have gotten, but
as you're beginning to see a distribution of
performance come about.
One of the things with the new process is
it takes a little time. You've got to let this play
out. When you get into the risk consideration of
issues and you put all this together the processing of
the issue takes a little longer than the old process
did. Very deterministic in the past. Did you comply
or did you not comply with the regulation?
Significant non-compliance -- you could get to an
escalated enforcement decision fairly quickly.
It is a little longer process now to
really put a an appropriate risk perspective on the
issue, and Troy will be able to talk to that probably
in more detail later on when we get into talking about
the SDP and where that goes. Art's probably got some
insights that he'll be sharing too. But it gets you
there.
MEMBER LEITCH: A question about Calvert
Cliffs, for example, where you're dealing with two
almost identical units, one in -- Unit 2 is in column
one and Unit 1 is in column two. I suspect that
what's driven Unit 1 to column two might be the fact
that it had three SCRAMs in a fairly short period of
time, but one was as I recall was a lightning strike.
Another one was a failure in an electronic component,
which could have just as easily occurred on the other
unit. It doesn't represent a different program or
different level of management attention. It's the
same management team.
And I just wondered does this indicate
that your level of inspection would actually be
different on Unit 1 for example than Unit 2?
MR. BROCKMAN: What you would immediately
get out of this would be Unit 2 would get what we call
the 95001 inspection -- excuse me. Unit 1 would get
the first level of investigatory inspection. This is
approximately one inspector for a week, and that
inspector goes out there and says, Okay. What is
behind here? I have a performance indicator that
threshold's been crossed, or I have this type of
insight that is not very low significance, but it's
not big. Let's go out there -- and this inspection is
to put that in the context, and it may be just what
you say. I've had a piece of equipment that had a
random failure to it, could not have been predicted,
caused the threshold to be crossed. The licensee's
dealing with it aggressively. That's the extent of
additional inspection they received.
MEMBER LEITCH: But that additional
inspection in this case would actually focus on Unit
1 as compared to --
MR. BROCKMAN: Yes. It would focus on
Unit 1 to put that insight into context and then
identify what's the right response that there should
be. Maybe there is something that is broader and I
have an extent of condition of vulnerability in Unit
2 that is appropriate to follow up on when I do the
problem identification and resolution inspection.
Maybe it's not.
Maybe I have got a unit-specific --
something that's going on here. If I had looked at
ANO, which is our site where I've got two different
vendors and the organization is very common in some
areas. In some areas it's not quite so common. Maybe
I determine it is something unique or maybe it's more
cross-cutting on the different units.
MEMBER LEITCH: Okay.
MR. BROCKMAN: That's the beauty of this
program.
MR. GWYNN: I think that it's particularly
insightful that the plant that's at the top of the
degraded cornerstone column which we do know about --
in a way we're not very familiar with Calvert Cliffs,
but we do know about that plant, and the things that
contributed to that situation are I think important
outcomes of this new baseline program and its focus on
risk important activities at the plants. We'll be
talking about a couple of those as a part of the
agenda later today.
And that plant was a category one
performer under a reduced inspection program for a
very long period of time, both when it was part of the
Region III oversight and then as a part of Region IV's
oversight, so this new baseline program has made a
difference at that facility.
CHAIRMAN SIEBER: Let me ask the question,
let's assume for the minute that the new reactor
oversight program is effective in coming up with a
distribution of performance across the fleet of
plants. However, under the old process there was a
different kind of response from the NRC that has to do
with significance determination to a great extent
where civil penalties were enacted, pressure releases
occurred when you've got a level three finding,
sometimes a public meeting in a local community, and
as a senior -- former senior vice president and chief
nuclear officer I can tell you those are attention
getters for the licensee.
So my question is now that civil penalties
are down and you don't have a lot of this fanfare do
you feel that the licensee's attention is just as high
under the new process as it was under the old process?
MR. BROCKMAN: Let me address that. I
would challenge one premise --
CHAIRMAN SIEBER: Okay.
MR. BROCKMAN: -- that you're presenting.
The fanfare is not down. In fact, the fanfare is
more. The only thing that's different is right to
check. If you go to the action matrix, which we've
got a copy of in your handout here back -- action
matrix right here --
CHAIRMAN SIEBER: Right.
MR. BROCKMAN: -- when we have one of
these issues -- and now it's done real time in a
supplemental inspection -- you're going to get
regulatory conference, and depending upon it it will
be in the local area, and you're going to get the
press releases associated with the white issue.
One of the things we do in Region IV,
we've gone to quarterly integrated inspection reports.
By that I mean for a given facility on a quarterly
basis the resident report is combined with all of the
small level region-based activities, the one, the two-
person inspections. We would give an exit
presentation if it's a DRS an HP inspector. They
would give an exit when they left. But the written
part of their report would come in at the end of the
quarter.
What are the differences for those?
Exceptions would be major team inspections. I've got
an engineering team out there. That report doesn't
wait for a quarter. It's a big activity. We cull
that out. It gets a separate report. Problem
identification resolution, any major activity that
we've got going on gets a separate report. Any
inspection that looks like it's going to have a white
finding or above we don't wait until the quarter.
That is culled out right now. It gets its own unique
inspection report number and comes out.
So it's addressed very contemporaneously
and we go right into the process: public meetings,
that regulatory meeting, the press release that goes
along with it. All of the other as you described
fanfare that went on is still fully there under the
new process. The only thing that's not is the change
in the enforcement policy for writing the check.
CHAIRMAN SIEBER: Let me follow up just a
little bit. If you ask the average member of the
public in the old days they understood $50,000 or
$10,000 pretty easily because it related to things
that they do, and when you say they had a violation,
they paid this civil penalty, they admitted that they
did wrong, that was pretty clear as far as the public
was concerned as to what actually happened there. But
if you tell the public that you went from a green to
a white perhaps there's some head scratching.
And I know that the NRC has spent a lot of
time in public meetings trying to explain the process,
but I don't think the public has as clear a notion as
to what is going on now with the grade of performance
as it used to be when it was pretty clear. The fact
that there were violations found, penalties being
enacted, and so forth.
Do you have any insight to that as to how
the public perceives the new process?
MR. CLARK: Kenny, can I address that?
MR. BROCKMAN: Jeff can probably do it
very well because he's at a site that's had several of
these change issues.
CHAIRMAN SIEBER: Right.
MR. CLARK: To address it let me go back
and talk about going into the pilot process and going
into the revised reactor oversight process.
As the senior resident at Cooper, Cooper
had performance problems going into this process. I
dealt very closely with the senior resident at Fort
Calhoun, and we dialogued throughout this process and
we saw big differences throughout this. I also
dialogued with the public a lot. We had several
public meetings. I live in Southeast Nebraska.
Everybody knows what your neighbor does, so --
CHAIRMAN SIEBER: Well, there aren't too
many neighbors.
MR. CLARK: I can see one house from my
house, so --
VOICE: Is it occupied?
MR. CLARK: No. So you go to the grocery
store and you go to a church meeting and you will get
dialogue about what is happening at Cooper, and I saw
in the transition phase they were still asking about
are they going to get fined for this thing that just
happened last week? Are they going to get fined for
this? And it took some discussion up front, but we
said, No. The new process is doing this by channeling
through the action matrix what type of response we
take, and it's going to have indicators. We explained
the indicators to them. That was a little fuzzy, but
I think the public is, at least in the vicinity of the
plants, coming onboard with what these indicators
mean.
And I'm going to say that from the
standpoint of we just had a number of performance
issues in the emergency response arena in emergency
preparedness at Cooper, and I have the public asking
me, How many whites did it have to get? So now
they're on board. They know what the indicators are,
they know how we respond now, and I think they're
becoming more aware of what risk was.
If I could turn the tables a little bit as
a resident under the old inspection program it was
sometimes difficult for me to defend the agency's
position on why these particular actions resulted in
this type of penalty. When we were looking at it as
combined significance or not being risk informed it
was sometimes difficult to defend what those actions
were. Conglomerating actions, conglomerating some
inspection findings to get an escalated issue with the
licensee was sometimes harder to explain to the public
than it is to say that we're going to put these into
these arenas, into these cornerstones. As you see the
performance match out it's going to come out.
And as we've seen and we'll discuss later,
we're seeing over a period of time that we're getting
the distribution, we're getting those colors, and
we're getting the response from the plants that we
somewhat predicted.
MR. GWYNN: I'd like to add to what Jeff
just said, and my perspective is a little different
from his. I was in the position that he's in back
when we were first starting to implement the
systematic assessment of licensee performance.
Number one, we still issue significant
notices of violation and impose civil penalties on
licensees for significant violations of NRC
regulations. I think that Jeff just explained that we
have a better threshold for determining the
significance of those violations now than perhaps what
we did in the past so the public can better understand
why we consider the issues significant.
I can tell you that making a number of
public presentations of SOWP under the early stages of
the program the public didn't have a clue what we were
saying, and we did very little to educate them as to
what SOWP was and what it meant. For this new
baseline inspection program we've had significant
public outreach, lots and lots of communication as
Jeff just indicated with the local community to
educate them as to what the program is, how it works.
They're learning over time, and as we
continue to hold these public meetings, as we continue
to gain experience with the program I think that the
public will become much more educated and much better
able to understand the agency's decision-making
process.
Now, an interesting side light from this,
there were times in the past, for example, the
Waterford steam-electric station that you just
visited, where it was like somebody turned a switch.
They went from being all SOWP category I to having a
category III in engineering and almost being on NRC's
watch list essentially overnight.
How does that happen? Under the new
program it doesn't. We have our action matrix.
People watch over time. As our inspection findings
and as the performance indicators build leading to
increased agency attention and more significant agency
actions up to and including major inspections,
commission attention, and perhaps even a plant
shutdown. And so I think that our process under this
new baseline program, which was one of the major
desires at the outset, is much more scrutable by the
industry and by the public.
They can understand where we've been,
where we're going, and why we're doing what we're
doing much better under this program than what they
could under the previous program, and so even though
I was not a major proponent of the program at its
outset I've become a major believer in the program as
I've seen it work.
CHAIRMAN SIEBER: Maybe I can comment on
the answers so far. First of all, I would
congratulate the agency and the region for the
outreach that's occurred, and I think that's the prime
reason why you're getting some degree of public
acceptance and understanding of what's going on, and
had that been done in the old system to the same
extent you might have had a different result under the
old system. But the resident still says -- the first
question they ask me is will they get fined for this?
So that's the expectation of the public, just like
going 30 miles an hour in a 25 mile zone. In
Pennsylvania where I live that's $141. I understand
that.
On the other hand, that's what the public
expects, and so it takes some explanation to explain
what this new system is, and probably it's a better
system, and I'll leave it at that.
On the other hand, you did mention, Pat,
one aspect that intrigues me when you talked about
Waterford where you said they went from a SOWP I to a
SOWP III instantaneously, and that wouldn't have
happened under the new system which tells me then that
you believe that it's predictive to some extent, and
I would be interested in knowing whether it truly is
predictive or the same thing could happen under a
baseline --
MR. BROCKMAN: The same thing can happen.
CHAIRMAN SIEBER: Okay.
MR. BROCKMAN: You cannot -- it is not
going to be the rule. The premise is that you're
going to see gradual degradation that would occur, but
you can't -- for example, there's nothing I can go
against stupid, and that could happen somewhere that
you've got someone out there who intentionally does
something and puts it into a vulnerability. You get
a catastrophic piece of equipment failure that has
implications. We did not have -- IP2 did not have
some whites, going to yellows and then proceeded on
into red. They had the catastrophic failure and it
had the significance that it had.
The system is not a 100 percent that can't
happen. It can happen. But --
CHAIRMAN SIEBER: So it's a mixture?
MR. BROCKMAN: -- it will be an exception.
CHAIRMAN SIEBER: It will be a mixture,
much less likely --
MR. BROCKMAN: Much less likely. We are
seeing with plants that in our old system seemed to be
the ones that continually had performance problems,
and as the data is building up we are seeing the
things coming together in the performance issues and
in performance indicators not so much, but the
performance issues coming together along those
lines -- let me answer a different question you had
earlier now that I've touched on that.
The next couple of slides show you a
couple of printouts off the web page, which I know
everyone here is intimately familiar with, being able
to get all the data. You see performance indicators
and inspection findings. I have emphasized the fact
that the new program consists of performance
indicators and inspection findings. If you look at
that chart that we had up there with all the plants
you can pretty well -- I haven't looked at all the
region specific data, but I would guess I could pretty
well predict which one of these plants are in the
regulatory response based upon performance indicators
and which ones are on inspection findings, and all the
ones that are one site out of multiple unit sites my
first question would be I'm going to guess that's a
performance indicator problem that got them there.
Without a doubt all the ones where I've
got both Quad Cities 1 and 2 and what have you, most
likely those are coming out of inspection findings.
Our experience here in Region IV is the inspection
findings are without a doubt still the driving
component of this program. You cannot give away the
inspection findings. The performance indicators are
a good insight but the thresholds are such that
without the inspection findings that predictivity
you're talking about, Jack, in being there would not
be there nearly as comfortably as we want it to be.
MEMBER APOSTOLAKIS: What's wrong with
the -- can you elaborate on that?
MR. BROCKMAN: I'll give you an example.
We're recently seen the agency received a
communication from Mr. Lochbaum talking about the
threshold on reactor trips and how we don't gain
insights on crossing reactor trip threshold III or V
or whatever it is. The risk threshold for reactor
trips to go from green to white 19. We're not going
to set up 18 trips to have in a year is okay.
The absolute risk part of it doesn't
necessarily go in with your gut, and certainly from
what the history is and what the performance of the
industry is from where they're at doesn't go into the
match up what you should have as your deterministic,
and once again, the risk number gets me to the
ballpark. What position am I playing? My gut says
I'm behind the plate. Five trips is enough, thank you
very much. And you've got to bring that together. If
this thing becomes risk based then the difference in
the PRAs at the different plants -- you've got to then
bring all of the data into a perfectly common playing
field, and we've got to have total confidence in its
absolute accuracy.
The industry and PRA is not there yet.
That's why we need to maintain the deterministic part
of it.
MEMBER APOSTOLAKIS: So the green-white
threshold for initiators is the three. That's not
unreasonable, is it? I understand that the red is --
MR. BROCKMAN: But if I did it on nothing
but risk -- the initial number that came up on risk
when we were developing this would have been -- it was
a humongous number. I want to say 19 -- 25 I think
was -- it was a crazy number.
MEMBER APOSTOLAKIS: That has to do with
how these numbers are derived and stop already because
every such program --
MR. BROCKMAN: Yes, sir.
MEMBER APOSTOLAKIS: But I'm trying to
understand. Let's say we had the right numbers. Do
you think that the inspections give you insights that
the performance indicator will never give you?
MR. BROCKMAN: Absolutely. The
performance indicator gives me insights in one aspect.
The inspection gets to things we don't have
performance indicators for, and the overlap is my
verification. The inspection also does some
verification that the performance indicator is being
properly reported, appropriately focused, so that's my
overlap on my vin, but the inspection definitely looks
at parts that we don't have performance indicators
for. There's not a good way that we've been able to
identify yet to gain that indication off a
quantifiable, reportable data.
Problem identification resolution's a
great example. I don't have a number that gets
calculated to say how good a licensee's corrective
action program is, and we all know that's the basis
upon which this entire new program is premised. I
think one of the key things out of the IIEP report was
the executive summary. If you read anything on that
report read the executive summary, because it takes
the data and actually takes a step back and tries to
start drawing some conclusions about what it's telling
you: the difference between risk informed,
deterministic applications.
There is a difference. It's a
philosophical difference. It's changing the way in
which the public looks at things. It's very easy.
You're going to get a fine. I understand that.
$55,000. Wow. I look at my budget. That's a hell of
a fine. I look at the licensee's budget. No. That
press release caused much more concern than that
$55,000 check did in the overall scheme of things at
the level we're talking about for a licensee.
But that --
MEMBER APOSTOLAKIS: But it seems though
that we have again a conflict here, because it
appears -- I agree with you that an inspection gives
you a better picture of what's going on. At the same
time the agency wants to go the performance-based
route, so --
MR. BROCKMAN: I'll challenge that. Yes.
Performance based, risk informed. Yes, sir.
MEMBER APOSTOLAKIS: You're challenging
what, that the agency wants to go that way or that
it's a good idea to go that way?
MR. BROCKMAN: No, no. I misspoke. I've
had so many discussions with other people. The first
thing I hear is risk based and that's not what you
said. You said performance based. So, yes, I'm with
you. Performance based.
MEMBER APOSTOLAKIS: So it seems to me
that the performance indicators are consistent with
this philosophical approach, and you might say that
maybe we could have a first screening based on the
performance indicators, and then if you find that the
numbers are disturbing then you go and do a more
detailed inspection. Would that be a better --
MR. BROCKMAN: That's exactly what we do.
MR. CLARK: Let me address that.
MEMBER APOSTOLAKIS: Well, the baseline
inspection is independent of --
MR. CLARK: I see it from the other
perspective. As an inspector I see it as the
performance indicators are overall view of the
performance of the plant, and those are the roll-up
perspectives of the plant. The insights that you get
from the individual inspection items will be the
precursors to those initiating events or those things
that get you into the performance indicators.
So we're being somewhat predictive, but
also if you actually look in the details of what the
inspection attachments that we do are -- let me step
back and say when we initially went into this in the
pilot process -- I speak somewhat for many of the
inspectors throughout the region and throughout the
country -- we were skeptical, because we said we're
moving from a process where you follow your nose after
something you don't like to you fill the bins, going
out there and getting inspectable areas accomplished,
and we said we are not going to be able to follow what
we feel is risk significant.
Well, I can tell you -- I have some risk
background -- I misunderstood what risk significant
was. After going through the process for a period of
time, having findings, placing them through the
significance determination processes Troy and Kriss
will talk about a little bit later, we gained some
very valuable insights as to what the precursors to
these events are, what the precursors to performance
indicators are. We're seeing those come out,
particularly at my facility at Cooper. We're seeing
now connect the dots between some of these inspectable
areas then going into performance indicators.
Performance indicators haven't tripped
I'll say as yet, but you're actually seeing some
degradation in those areas, and I think with the
inspection findings we can go back and say this is
why, because they don't understand design basis. They
don't understand the performance of their operators.
MEMBER APOSTOLAKIS: I think that raises
another interest in philosophical question. This
business of leading indicators and trying to predict
what's going to happen. Again, you can say I have the
initiating events cornerstone and I would like to have
inspections before that to figure out when that
indicator of initiating events will go over the first
threshold.
Then you may stop and ask yourself why
would I want to do that? The initiating event
cornerstone is itself a leading indicator for core
melt, so there is no end to this. At some point you
have to draw the line and say enough is enough. I
don't really want to know that the plant is going this
way and eventually the initiating event cornerstone
will go over to white, because that by itself is
telling me something about the risk, and to say no, if
I do something else I will be able to tell in advance
when the initiating event cornerstone will go to
white, why would you want to do that? That was
against the performance based approach, was it not?
MR. BROCKMAN: Absolutely.
MEMBER APOSTOLAKIS: So where do you draw
the line? I understand the desire to know, but the
licensee on the other hand says, wait a minute. This
was supposed to be performance based.
MR. BROCKMAN: Let me put a different spin
on it, and I think you and I are very much cut from
the same cloth on this.
There's not a performance indicator,
there's not an inspection finding out there that's
predictive. Everything they've reported or we find
has already happened.
MEMBER APOSTOLAKIS: That's right.
MR. BROCKMAN: It's reactive.
MEMBER APOSTOLAKIS: Right.
MR. BROCKMAN: And we need to admit that
up front. It is reactive.
Now, the thresholds we set try to get us
to the point of saying it's becoming more than
coincidence. The licensee is not controlling their
destiny to the way they need to be. We need to get
interactive and provide assistance, provide more
oversight. That's the predictivity of it. It's not
that I'm going to predict when it happens. I'm not
going to do that. It's the level of interaction that
needs to be done to try to assuage a problem that's
moving from going further down the line. I think
that's very good for the individual items.
We've got the other thing that we
haven't -- the magic word we haven't talked about yet,
and I guess it's time we throw it on the table, cross-
cutting issues.
MEMBER POWERS: We're going to get to it.
MR. BROCKMAN: That might be the one that
has a bit of predictivity. And once again, as you've
told -- I talk with a little picture, and let me throw
my view of cross-cutting issues here. I have a house
sitting on stilts by the ocean. Each one of these
cornerstones is a stilt. When I have a degraded
cornerstone I've broken a stilt. My house tips a
little bit. If I break another cornerstone it tips
more. If I break enough and you get into degraded
multiple the house slips off and it falls down in the
ocean. We have a problem.
The cross-cutting issues -- I've got
somebody out there who's taking nibbles out of all of
my stilts. I get to the point finally where I have
not had a single stilt break, but the stilts as a
whole will not hold the weight of the house, and the
house catastrophically comes down, and I didn't have
the cornerstone fault beforehand. That's what cross-
cutting issues are trying to address, taking a bite
out of each stilt.
Typically in the licensee's corrective
action capabilities, human performance initiatives,
those are the areas that manifest themselves
throughout plant operations as we all know. That's
the concept of cross-cutting.
MEMBER POWERS: And your analogy is nice,
because we understand gravity. Now come to the real
situation. What's the phenomenalogical consideration
that leads me to believe that I can tell people who
are having the bites taken out of their human
performance activities and I can tell that because of
one of the performance indicators.
MR. BROCKMAN: I personally believe that
the cross-cutting issues we identify I'm finding more
out of the inspection findings. I've got to go into
the whys are these happening. I don't have a human
performance indicator --
MEMBER POWERS: It's really coming out of
your root cause analysis.
MR. BROCKMAN: You've got to -- and it
keeps on going back to their corrective action
program. Are they effectively managing -- have they
identified it? Are they dealing with it? Then I back
off.
MEMBER POWERS: But the trouble is are you
looking -- well, the question is are you looking at
the root cause analyses for all the non-cited, non-
written up kinds of inspection findings?
MR. BROCKMAN: We sample. There is a
sampling, and Art can probably speak very well. The
leadership for our corrective action inspection
problem identification resolutions under his domain --
you may want to share --
MR. HOWELL: Right. First of all, we do
try to identify those things that are potentially the
most significant to understand better the nature of
the extended condition and why they happen, and we use
not only the docket but we also use licensee records
to do that, and we get all that information.
So to answer your question directly, yes.
We look at issues that are not in the docket that we
have not necessarily already inspected and put into
our inspection reports. We try to assess trends and
patterns from our review of information and to make
some judgments about how effective a particular part
of the program is working.
The difficulty is what do you do with all
that? How significant is all those minor issues or
issues that don't trip an SDP threshold. So you have
a collection of insights that perhaps you can share
with a licensee but it's not at all clear what that's
telling you about performance given that we're only
sampling to a very small rate. A very small
percentage of issues ever get looked at in the form of
our reviews. We try to do the best we can.
MR. GWYNN: I have a question if you don't
mind. While you were at Waterford did the licensee
share with you its internal performance indicators --
MEMBER POWERS: Yes.
MR. GWYNN: -- the indicators they used to
manage their facility?
MEMBER POWERS: Well, they shared with us
some set of them and --
MR. GWYNN: Typically what I see is that
they have very different thresholds than what we use,
and it's appropriate. It's their -- they're in the
control bin. And I think significantly all of the
licensees that I'm aware of monitor human performance
and have human performance indicators that they rely
on to get them clues that things are not going in the
right direction at their plants.
That's perhaps the closest thing that I've
seen to a predictive indicator that licensees use, but
they're very -- there's a lot of variability. Every
organization has a different approach, and there's a
lot of unreliability in the data systems, and so we
wouldn't adopt those for the agency's use.
MEMBER POWERS: Yes. They can't.
Certainly Waterford -- they've identified human
performance as one of their concerns, whereas if it's
one of your concerns about Waterford it's not one of
your high level concerns, but it is for them, and
they've also looked at safety culture, which I don't
think you would ever try to look at. They probably
are looking at management philosophy, which I hope you
wouldn't look at.
Clearly they have a different set.
CHAIRMAN SIEBER: I think the tools that
they use are management tools and not regulatory
tools, and you can't use one for the other, and
actually the Waterford system is pretty common. I can
name you a dozen other plants that use basically the
same system. Wherever that steward went that system
went with him. Look at Palo Verde and --
MEMBER APOSTOLAKIS: We will discuss the
cross-cutting issues later.
CHAIRMAN SIEBER: Yes. One of the things
I would point --
MR. BROCKMAN: If it's a topic and you're
not tied to the agenda this would be the time to talk
about it.
CHAIRMAN SIEBER: Okay. One thing I would
point out -- and I think this has been a great
conversation because we're finding out the things that
we needed to learn to do our jobs from you, and that's
a great benefit for us. On the other hand, I keep
looking at the schedule and my airplane ticket, and I
would like to move on.
MEMBER APOSTOLAKIS: The cross-cutting
issues though -- if there is a place to discuss them
then we should. Otherwise we do it now.
CHAIRMAN SIEBER: Yes. It's important.
MR. BROCKMAN: This would be where we
would do it. Now, also if it's an individual thing
we've got the entire noon hour if you would like to
talk about that. I'm not trying to suggest -- however
you all want to do it we're here to support you.
MEMBER APOSTOLAKIS: The thing about the
indicators that we saw at Waterford yesterday when it
comes to human performance I don't know how much
they're telling you, because there is an implicit
assumption there that -- when they plot the human
error rates these are during normal conditions.
Right? In fact, they told us that every morning they
have a senior management meeting where they evaluate
what happened and they declare something as being a
human error. I think that's a reasonable thing to do
because it's obvious what is a human error.
But these human errors are found to occur
during normal operations, and there is an assumption
there that if you're doing well in that respect then
if you actually have an initiating event you will also
do well. And it's so clear to me that that's the
case, that if you're doing well with respect to
routine maintenance then if there is a need to decide
to go to bleed and feed it will do equally well. I
don't see that --
MR. BROCKMAN: In fact, you can build the
argument it could take you in either direction. The
higher sensitivity and the urgency makes people more
focused, they'll do better, and the other side is is
the infrequently performed activity and the stress
will come up as they perform less efficiently.
MEMBER APOSTOLAKIS: That's right.
Exactly. So again, I'm not arguing that you shouldn't
be doing well because you don't know. I'm not saying
that. But I think to feel comfortable that one was
switched to this -- when the initiating event occurs
you have a very different culture perhaps, so if that
doesn't help me that the human error rate goes down
what does? It seems to me that I have to do
inspections and evaluate what is happening and maybe
also use questionnaires because now the issue of
safety culture in my mind becomes much more important.
Now, at the same time I know that the
commission has cooled to the idea of the agency
looking into safety culture issues, so they're clear
it's a problem, because if they say don't do it you
don't do it. But we have this problem it seems to
me -- and maybe -- first of all, I would like to know
what your reaction is to these thoughts and second,
perhaps we should try to sensitize the commission to
these issues.
But I just don't see how normal indicators
help me understand what the operators are going to do
under extreme time pressure in a critical situation.
MR. BROCKMAN: Let me give you my
thoughts, and I want to ask Troy to inject a point too
here based upon your November finding over at River
Bend where you made the cross-cutting issue finding.
MR. PRUETT: Okay.
MR. BROCKMAN: One thing that I would say
with respect to human performance if they can't do it
well under normal conditions I have no faith they'll
do it right under stressful ones.
MEMBER APOSTOLAKIS: And I think that's a
very good point.
MR. BROCKMAN: It establishes that's why
we're looking at it from the normal. At least it
says -- I have not lost confidence. I can't say I've
got it, but if they don't do it right under normal
then I have lost my confidence they'll be able to do
it under more exigent conditions. So I think that's
the value that brings. It answers that question. Not
the other side of the coin.
Now, Troy was my senior resident out of
River Bend, just recently has come into the site. He
mentioned that to you. One of the things that he has
done -- the new program allows us as part of the
normal inspection program to try to identify cross-
cutting issues in this area, and he's one of the few
who's been able to put together logic and have a
respected inspection finding in this area nationally,
and I'd like him to be able to share what his logic
was on going about that last fall.
MR. PRUETT: Essentially we've developed
a human performance cross-cutting issue in the
operations area which involved questioning attitude
and operator awareness of plant conditions, and
initially that started with -- we looked at
performance indicators associated with the risk
significant systems of the plant. None of those
performance indicators had crossed a threshold over
into the white band, but we were seeing an increase in
hours in plant unavailability on selected systems,
mainly service, water, and some diesel generator
systems.
With that we decided to take a multiprong
approach and look at -- implement the baseline
inspection program by -- we used a maintenance rule
procedure to look at those systems to see if they were
accounting those unavailability hours correctly, if
they classified the deficiencies properly and
implemented the appropriate corrective actions.
We also went after post-maintenance
testing in those areas as well as surveillance in
those areas, and our op evals inspection focused on
those same systems, and what we were able to come up
with was a number of deficiencies involving each of
those inspection modules on those systems, and as it
turned out there were inappropriate engineering
evaluations with inappropriate operator reviews
associated with those that involved a lack of
understanding of the system or a lack of awareness of
plant indications associated with that issue, or
inappropriate post-maintenance test methodology which
was due to a lack of operator or engineering or
maintenance craft understanding.
And eventually we developed a trend of
approximately 20 to 30 findings associated with some
type of poor or inadequate human performance aspect
with each of those inspection modules, and we rolled
those up together and termed it a cross-cutting issue.
And it gets to what Ken was pointing out
earlier. There's a lot of stilts out there, and what
we were seeing was bites being taken out of a half a
dozen or ten different areas.
MR. BROCKMAN: The key thing is what do
you do with that? We brought it forward as a finding.
The licensee in fact embraced the finding. They
didn't necessarily like it being documented. That's
a different issue. But they had no disagreement at
all with the insight, with the assessment, with the
finding being brought forward. And they have
initiated corrective actions to be dealing with that
within the licensee response arena, and that's what we
did. We brought it forward and then we sat back and
watched the licensee deal with it.
You would notice from our annual
assessment letter that came at the end we see they are
making progress. They are doing what you would expect
a licensee to do in the licensee response man, and
that was not a conceptual problem with respect to our
annual assessment. We didn't carry it on as an annual
level concern because they were dealing with it in a
manner that was responsive to try to improve and make
that problem go away.
CHAIRMAN SIEBER: The big question here
though is -- obviously, Troy, you've done a really
good job. The question is do the other 12 resident
offices in your region -- can they do the same kind of
job and can they do it nationwide to gather together
these insights to make it work?
MR. PRUETT: There's only one of me. We
don't have --
MR. BROCKMAN: There is no pride in Troy's
family. He has garnered it all in his --
MEMBER APOSTOLAKIS: But that was my next
question is very much related to what Jack said.
Let's say the commission said go ahead and do
something about safety cultures and work environment.
Do you --
VOICE: And they will say that eventually.
MEMBER APOSTOLAKIS: But do you think that
it is possible to identify a number of indicators that
will tell me something about the safety culture,
because this is the argument right now. In fact,
Commissioner Diaz came to me and we asked why do you
feel that we shouldn't be looking into this? He says,
You can't measure it so leave it alone. Essentially
that's what he said.
So is it -- measuring it probably is a
very ambitious thing to do, but at least can we
identify if your indicators say if I look at A, B, C,
D then I can tell something. Now, my colleagues with
the utility experience sometimes tell me that the
moment you walk into a plant within a minute you know
whether the culture is good. Right? And if they talk
about Coca-Cola cans being left --
VOICE: In the ventilator ducts.
MEMBER APOSTOLAKIS: Yes.
MR. PRUETT: I think you can take some of
the performance indicators we have right now, the
SCRAMs or the safety systems or BSF actuation type
indicators and look at those and provided there's not
a single issue with -- where you take fault exposure
hours that put you into that threshold, but if you
have multiple instances of where you're increasing
your unavailability numbers and you actually look at
the data, that's an insight I believe into human
performance.
MEMBER APOSTOLAKIS: So it's the
repetitiveness --
MR. PRUETT: I think so.
MEMBER APOSTOLAKIS: -- because it points
towards an underlying cause.
MR. PRUETT: That's right. And you have
to use the inspection program to go find out what that
underlying cause is.
CHAIRMAN SIEBER: It's not performance
indicators that's doing this though. It's analysis.
MR. PRUETT: Right.
MR. BROCKMAN: Absolutely. And the
challenge is going to be how thin do you want to slice
this? How good do you want it to be? We're going to
talk later on today about some things we're doing with
California plants. PG&E right now has declared
protection under Chapter 11. We know that. I have
specific things that the residents are following up on
on basically a daily basis as part of plant status
reporting that gives us indications that the safety
culture that I'm talking now at 30,000 feet is being
properly focused, that we're not losing it.
Yes. I can come up with something at that
level pretty good. Now, if you want to know do I have
the ultimate confidence that everybody's going to
record every single issue no matter what and bring it
in, that's a much thinner slice and becomes much more
difficult to do. So the answer is where we want to
set that threshold to be able to do that.
MEMBER APOSTOLAKIS: So to close this
subject so Mr. Sieber will not have a heart attack or
high blood pressure --
CHAIRMAN SIEBER: No. I already have
that.
MEMBER APOSTOLAKIS: -- you would not
discourage the ACRS from pursuing this issue and
coming back -- going back to the commission and saying
this is something we have to look into? Look into it
doesn't mean establishing a regulation tomorrow,
because that's a common misunderstanding sometimes
among the licensees, but understand it a little
better. What do we mean by safety culture, and maybe
are there any insights one can draw by looking at
certain things and saying something about it? Would
you discourage us from doing that?
MR. GWYNN: I think this is a very
difficult subject. When you're talking about true
safety culture you're talking about are the operators
sleeping in the control room? Are the operators and
the maintainers performing their duties by the book so
that you have confidence that the surveillance tests
have really been performed, that they've really met
their acceptance criteria, that the logs in the
control room haven't been tampered with, that the
strip charts from the control room recorders haven't
been flushed down the toilet. That's very difficult
to get at from the outside. I think that it's almost
impossible to get at from the outside.
And so I don't know and I don't have a
clue as to what this agency might be able to do to get
at that type of safety culture issues that are I think
at the root of what the industry and the public ought
to be concerned about. I know from inside the
organization you can get at those problems.
VOICE: Yes, you can.
MR. GWYNN: But from our position it would
be extremely difficult if not impossible in my view to
be able to deal with and identify safety culture
problems. That's just a personal opinion.
MR. BROCKMAN: -- morally I can't argue
with that. Your premise has the moral high ground
totally captured. The difficulties of implementing an
inspection program in this area though are
significant, especially with no rules or regulations
to fall back on. You have to -- and this program does
more to get there than anything else because it's
performance based.
We make findings now -- we've made
findings in the first year that under the old program
would have not even been documented that have been in
observation, and we've got white findings out there
now. It's a performance finding. It was not a
violation. You did not violate the rules, but your
performance is of such significance that it's white.
We've got other ones on the other arena.
I think those issues go very much toward the aspect of
the safety culture there.
MR. GWYNN: I think that we -- if the
agency did put together an inspection program to deal
with safety culture we could do it, but I think that
we would be fooling ourselves that it had any
meaningful results in terms of evaluating the true
safety culture at the facility.
MEMBER APOSTOLAKIS: But there is a later
question. Maybe I agree with you that this would be
very difficult for us to do, but there is also another
side, that what we do intentionally or unintentionally
does affect the safety culture of the plant, does it
not? Should we try to understand then our impact on
the safety culture of the plant? Would that be easier
to do in terms of the inspections we do, in terms of
other things we do?
There was this report in England where
they had as an example of an overly prescriptive
system that had a negative impact on the safety
culture of the licensees, the American system. Now,
should that tell us something that we should be doing
something about it, or no, they don't know what
they're talking about, because that's something we are
doing now. It's not that we're trying to evaluate
what the licensees' processes are. We are doing that
to them. Do we understand enough to do that or is
that a hopeless thing or maybe shouldn't be very high
on the priority list?
MR. BROCKMAN: Our processes -- put
yourself in the laboratory with yourself being the
professor. I now have a process going on that has
10,000 input variables to it, and I want to identify
what's the impact of this one, and it has both
positive and negative impacts and I want to determine
are the negatives greater than the positives. It's
easy to do as long as I can separate out the other
9,999, and that's what I don't know how to do.
MEMBER APOSTOLAKIS: Okay. I think I've
got basically -- you will be out there fighting with
us.
MR. BROCKMAN: The other thing that would
cause me a concern is the further we get down this
path the greater the expectation by external
stakeholders that we could be totally predictive on a
step change would never occur. You won't -- if you
can do this you'll never go from green to yellow.
That can still happen no matter how much of a handle
we've got on their safety culture --
MEMBER APOSTOLAKIS: All right.
MR. BROCKMAN: -- and I would be concerned
about that.
MEMBER POWERS: It seems to me the insight
that Ken -- that I need to spend more time thinking
about with respect to safety culture is the
examination of the corrective action program and the
root cause analysis. I think if what I have is a
great deal of confidence that there are a number of
licensees that know exactly what they mean by safety
culture. I see documentation that they have
identified deficient safety culture, they've sat about
correcting it.
Those corrections that they have
documented, written down in magazines say we address
these things are to my mind safety culture issues, and
they seem to have gotten better performance by their
metrics.
Their metrics are a little more sensitive.
They're a little more comprehensive than yours, but
they're their metrics and they did well.
It seems to me Ken's offered us an insight
here that we can get an appreciation appropriate for
the regulatory program by looking at how they handle
the root cause analyses in their corrective action
programs, and that might be a better way to pursue it
than looking for performance indicators and things
like that.
MEMBER APOSTOLAKIS: And again, by safety
culture -- maybe we should have said that much
earlier -- I don't just mean the attitudes of people.
It's the totality of how they do business which
includes the organizational issues, how certain
analysis are done, and these are more tangible in my
view. I agree with Dana that it would be easier to
see what would you do -- how would you do the root
cause analysis here rather than trying to figure out
what the attitudes of people are, which is really a
hopeless task?
So I think I got your input --
CHAIRMAN SIEBER: Enough to write your
report?
MEMBER APOSTOLAKIS: Well --
CHAIRMAN SIEBER: Why don't we move on?
MEMBER LEITCH: Another question about the
reactor oversight process. There seems to be some
confusion regarding the difference in the meaning of
the green color between performance indicators and
inspection findings. Does that difference cause any
confusion in the agency? It causes us a little bit of
confusion. We see green meaning one thing in
performance indicators and green meaning something
different in the inspection finding areas.
MR. BROCKMAN: Green means the same thing
in both. Green as -- but let me -- as has been
defined, green means the issue of significance such
that it is in the licensee's control bin. That's what
green means.
However, the American public does not see
green that way, and we as engineers can define it all
we want to and they don't accept that definition, and
that's Dr. Lippoti's argument is you call it green,
you've told me what it is. That's very nice but I'm
sorry. I forget about that ten seconds after you tell
me and green is good, and in performance indicators
green is good, and all my residents have a sign out
there at their resident's office, green is not equal
to good when it comes to inspection findings. It's
still an issue.
MEMBER APOSTOLAKIS: So it doesn't mean
the same thing.
MR. BROCKMAN: And that's the dilemma you
get to is we as engineers can define it all we want,
which we've done in this program, and it is a
continual challenge to put that in perspective. More
and more that it's out there the more people are
understanding what we're saying.
There was a point that Jeff brought up
earlier where he -- everybody is understanding what's
going on at Cooper, in the neighborhood of Cooper. I
can promise you at Fort Calhoun the public does not
have an understanding of white issues and how they're
dealing to the degree they do at Cooper. Why? They
haven't had any. And until you get this being played
out in the local arenas and they see one and have to
deal with it there's going to be confusion out there.
Art, your thoughts?
MR. HOWELL: No. They clearly are
different. Licensees strive to maintain themselves in
the green band for PIs and they strive very hard not
to have any green inspection findings or any other
inspection findings for that matter.
MEMBER LEITCH: I have another question
about the reactor oversight program. It seems to me
that there are apparently different weights
unconsciously applied to the different cornerstones.
For example, there was one plant in Region IV that we
read about in our briefing material -- I think it was
Callaway -- that had three radiation protection
issues, and so they had three white findings in
radiation protection. There was another plant, San
Onofre, that had a major operational event, switch
gear fire, wound up melting the turbine bearings down
and grinding to a stop, and that got a non-sited green
violation. At least that's the way I read it.
VOICE: You're accurate.
MEMBER LEITCH: I think -- and it seems to
me that those are just disproportionate. I'm not
questioning the significant determination process if
the blanket was properly followed and correctly led
you to those conclusions, but do you find in your mind
that there's something disproportionate about those
two findings?
MR. HOWELL: Really, one of the challenges
that we have is how to deal with issues that don't
lend themselves to PRA analysis, and that's really
what we're talking about. And we've made an effort to
define deterministically what's important and what
isn't in this first year, and as we've gone along
we've found as Pat indicated that issues heretofore
that perhaps we wouldn't have considered to be
particularly important or spend a whole lot of time
looking at have been elevated in importance vis a vis
the new process, and certainly that's also true in the
other direction.
And the question is are we in the right
place yet, and I think there's still a number of
questions out there and a number of these
deterministic SDPs where the results are getting us to
the right place. Are we truly treating -- is it truly
appropriate for example to have ALARA findings cross
a green-white threshold or a white-yellow threshold
for that matter when on the other hand you can have a
fire at a plant melt your turbine, challenge the
operators, put them under stress, et cetera, and so
it's very difficult to make comparisons in terms of
significance.
MR. GWYNN: I'd like to just make a
comment at this point that I think helped me to put
the ALARA findings at Callaway into good perspective
from a safety standpoint. I was visiting the Palo
Verde plant with Commissioner Merrifield not too long
ago and as we were being briefed they raised the issue
of the Callaway white findings in ALARA, but right
behind the head of the vice president at the plant
were their ALARA statistics, and for three very large
power reactor units their total dose to their
operating staff was less than the dose to the
operating staff at Callaway for one smaller unit.
And how can you say that we're not putting
our attention in the right place at Callaway by
focusing on ALARA when in fact they have those types
of results at their facility? On the other hand at
San Onofre there were no safety systems that were
challenged as a result of the fire and explosion that
occurred. And so I think from a risk standpoint the
program is taking us in the right direction at both of
these facilities. It's just -- I may be wrong, but
that's my belief.
MEMBER LEITCH: I don't mean to down play
in any sense the Callaway incident. In fact radiation
safety is a critical part of our business. That's not
where I'm going. What I'm trying to say is did the
process -- and I believe the process was properly
applied as per the process, but my question really is
did the process lead us to reasonable conclusions?
MR. BROCKMAN: We asked the same question
when we were processing the Callaway aspect. There
was a lot of debate going back here -- three whites as
to where this is going. It was a great deal of
exactly what you're saying. Is this taking us to the
right point?
One of the things we used to reach our
decision was we're going to follow the process in the
first year and then we're going to identify that as
part of the feedback process, this needs to be looked
at. We're not going to set off down the path and in
the first year, which is the initial implementation
year, say first time we come across a bump in the road
we throw away the process. What credibility do we
have with our stakeholders if the first time we hit a
bump in the road we abandon the process? We chose not
to.
If that in fact had not been given as one
of the issues to be looked at at the end of the year
of lessons learned -- and it was if you remember, and
the internal working groups and the external working
groups, the SDP for ALARA was one of the issues that
needed to be looked at to see is it coming up in the
right spot and if in fact it's being looked at and
there are revisions coming out.
So I would say your concern is one a lot
of people had and there are certainly some marginal
adjustments that are being made to it that may
preclude such an imbalance in the future. I'm not
sure exactly where it's at at the moment, but I know
it's something that's definitely being looked at
because it just didn't pass the initial wow test.
CHAIRMAN SIEBER: When I looked at that I
didn't come to the same conclusion because in my
opinion the regulator's job and the licensees' job are
the same, which is protection of the public health and
safety, protection of the health and safety of their
workers, which is Part 20 and the protection of the
reactor and cone system pressure boundary and your
mitigating systems and so forth, but if you melt a
turbine bearing that's dollars and outage time, not
safety related, so that tells me the whole
significance determination process one way or another
worked in this case to distinguish between what is
important from a regulatory standpoint from those
things even though they may be costly are not safety
significant, and so that's what I got out of that.
That's the way I would have looked at it.
MEMBER LEITCH: But it wasn't just the
main unit though. There were other aspects of fire --
failure to identify precursors that could have led
them to the --
MR. BROCKMAN: Yes. And there's a lot
there, and I can go into that, but very much all of
that was in the power generation side of the house.
And what it really becomes is appropriately
communicating that to all the concerned stakeholders,
because that's what we're talking about. Three whites
versus one white. Will that define the action that we
took? And we were questioning not whether it was a
white issue. It was how many.
The other part of it very much though is
to us doing our job in communicating that, generating
confidence in our external stakeholders that we're
appropriately regulating the industry, making sure the
industry is appropriately focused on the corrective
actions in addressing embracing issues, addressing
them, correcting them. Those are where you get out on
some of the other parts of it. And it's an
interesting dilemma at the moment when everything is
not perfectly risk informed.
CHAIRMAN SIEBER: But that's what safety
culture is, is being able to make these decisions
between what is significant from the standpoint of
human beings and the safety of the plant versus what
is significant as far as being commercially viable is
concerned, and that is something that has to be taught
by the agency.
MR. GWYNN: We have both of these issues
on the agenda for today, and --
CHAIRMAN SIEBER: We may have covered
them.
MEMBER POWERS: I think there's a lot more
that we want to go into in a couple of those issues,
but they follow this track.
MR. GWYNN: Yes, and I would like to note
that Gail Good, who's the branch chief for our
emergency preparedness health physics and safeguards
inspections here in Region IV has joined us in the
room, and she will be presenting the Callaway ALARA
experience a little bit later this morning. And we
have the SONGS electrical fire on the agenda for this
afternoon.
VOICE: So what's next?
CHAIRMAN SIEBER: Let me suggest at this
time since we are a few minutes behind, if you are
finished, which it appears that we are, maybe we can
take a 15 minute break at this point.
(Whereupon, a short recess was taken.)
CHAIRMAN SIEBER: The next presentation
we're going to listen to is the significance
determination process as it's implemented here in
Region IV, and I think after that we'll break for
lunch because lunch is a hot lunch, and if we don't
break then it will not be a hot lunch. And so let's
move briskly through the SDP.
MR. GWYNN: Our two senior reactor
analysts, Kriss Kennedy and Troy Pruett, will be
making this presentation. I've asked Kriss, the
primary presenter, to try to skip through some of the
information and maximize the time focus on areas that
might be of interest to the committee.
Kriss?
MR. KENNEDY: Good morning. My name's
Kriss Kennedy. I was selected as SRA, started the job
in November of 2000, started the training in December,
and I'm still in the qualification process as is Troy,
who you met earlier. My background is I started out
in the agency as an operator licensee examiner. I've
been the resident inspector at Comanche Peak and the
senior resident inspector at Arkansas Nuclear 1.
The senior reactor analysts in Region IV
are assigned to Division of Reactor Safety. Art
Howell is our boss and we are the focal point for risk
informed activities in the region. In addition to
Troy and myself we have a branch chief in the Division
of Reactor Projects that was previously qualified as
an SRA, and we also have three staff members that are
going through the advanced risk training that some of
the regions are sending their people through. In
fact, they're in their second week of training this
week, so those are the resources we have available in
Region IV.
We're going to go ahead and skip the next
couple of slides where I was prepared the discuss the
SRA functions in Region IV, the various tasks that we
perform, and we'll go directly to the slide entitled
status of risk tools. I think that may get us more
into some of the discussion areas that you are
interested in.
CHAIRMAN SIEBER: One quick question which
would prompt a yes or no answer --
MR. KENNEDY: Okay.
CHAIRMAN SIEBER: -- you said that these
are the resources available to Region IV to conduct
these functions. Are those resources in your opinion
adequate, two people? Yes or no?
MR. KENNEDY: Yes or no.
CHAIRMAN SIEBER: Everyone is ready to
take notes.
VOICE: You will be quoted.
MR. KENNEDY: Yes. I think right now they
are. If the process goes where the program office
wants it to go it will be enough also. There -- I
guess I'm not going to give you a yes or no answer.
CHAIRMAN SIEBER: I accept that.
MR. KENNEDY: During the first year of --
CHAIRMAN SIEBER: You've already said
enough.
MR. KENNEDY: During the first year of
implementation and during even into the second year of
implementation there's a lot of startup costs with
using the new process. The phase two worksheets which
we'll talk about more are just coming out, inspectors
are learning how to use them -- actually using them
and so we're pretty busy.
CHAIRMAN SIEBER: I imagine.
MR. GWYNN: I'd like to just make a
parenthetical note here that Region IV management made
a decision early on in the process that we were going
on select the very best people that we could to be
senior reactor analysts in the region because they
were such critical positions, and as a result those
people are also very promotable. We had two of the
very most talented senior reactor analysts that were
available to the agency. Both of them were promoted
to branch chief positions and that's why both of our
SRAs at this point in time are in training.
But we have two highly talented SRAs in
training. Their work load will go down as soon as
they complete their training, and I think that we'll
be back in a more normal mode of operations and then
Kriss might have been able to answer yes to your
question emphatically.
CHAIRMAN SIEBER: Thank you.
MR. KENNEDY: And Troy didn't get an input
either, so Troy may have --
MEMBER POWERS: I guess the question goes
on. It will probably get into it as you go through
your presentation, but I note one of the slides that
you skipped over is the development of comprehensive
risk informed resources, and I'm going to be anxious
to know what kind of risk resources that you have in
the area of fire risk, shutdown risk, and seismic
risk.
MR. KENNEDY: You haven't looked at the
last slide. Those are actually listed as challenges
that we'll get into.
MEMBER POWERS: If the resources are
adequate then why is what we have adequate?
MR. KENNEDY: If we could go on to a
couple of slides I'll hold that as a question and
we'll go on to that.
This portion I wanted to discuss the
status of the risk tools that we have available to us,
and primarily these risk tools come out of manual
chapter 609, significance determination process for
the first part. The risk informed inspection
notebooks also known as the SDP phase two
worksheets -- in Region IV NRR has issued eleven of
the 15 worksheets for Region IV plants. We're at 73
percent there. NRR has also has a processing program
to go out and benchmark those phase two worksheets,
make a site visit, sit down with the licensees, PRA
folks, and go through system by system, compare the
results that the licensees get with their models,
compare the results that we get with the worksheets,
and identify any changes or errors that we need to
correct on the worksheets.
MEMBER POWERS: I take it this has not
been done with Waterford?
MR. KENNEDY: It has not been done with
Waterford. No.
MEMBER POWERS: Because they were wincing.
I mean, they feel left out. They feel hurt and
unloved and unwanted.
MR. KENNEDY: Well, they shouldn't.
There's only been four benchmarking trips to date.
Three of them have been in Region IV, so it's a
process that's ongoing and will continue at least
through -- to completion, which may be the end of next
fiscal year, so some plants will wait -- will have to
wait.
The other risk tool -- one of the other
risk tools that we use is the standardized plant
analysis risk models, the SPAR models. Those were
developed by INEL. They've come out with revision
three for some plants. In Region IV we have eight of
15 revision three models out, and of those eight none
have been QA. None have gone through a site QA
process.
MEMBER POWERS: What is the meaning of QA?
They've presumably complied with the NRC's mandates on
software QA.
MR. KENNEDY: By QA I really mean similar
to a benchmark trip where they go out to the site with
the model, compare the results of the SPAR model to
the results of the licensee's model and identify where
the differences are.
MEMBER POWERS: So it's really a
verification then?
MR. KENNEDY: Yes. The term QA comes from
the revision two models where they issued a
revision -- what they called 2I and then after the QA
process they would call it revision 2QA, so we're at
revision 3I for these plants and once they're QA'd
they'll be a rev3QA.
CHAIRMAN SIEBER: Quick question. When
you make a benchmark trip to a licensee's facility
you're comparing the results of the SPAR model against
a licensee's PRA. What criteria if any do you use to
judge the quality of the licensee's PRA?
MR. KENNEDY: We're not really there to
review the quality of licensees' PRAs. That's the
first part. But what we do is when we identify
significant differences in the results of the
worksheets and the results of the licensee's model
then we start asking questions, figure out what they
have in their model, why they're getting different
results, and if we're looking specifically at that
area and there's a specific problem with the
licensee's model in that area -- although that's not
the norm. It's typically a problem with the
worksheet -- then we'll point that out.
And we had one example of that at South
Texas I believe where they -- we identified an error
in their model. It was a minor error with the steam
generator PRBs, and --
MR. PRUETT: The PRBs. They assumed they
only needed one PRB for an accident. In reality, we
challenged that, and I believe they needed to have a
minimum of four.
MEMBER POWERS: This is not a trivial
mistake.
MR. KENNEDY: Well, in the overall impact
on the PRA it was not a large significant error.
CHAIRMAN SIEBER: Now, if you're using the
SDP process for enforcement for example or to evaluate
a licensee application to NRR even though NRR will
probably do that examination, or ask CENED-ED-EH to do
it, as they have in the past, would you do some
different kind of evaluation of the licensee's PRA?
MR. KENNEDY: The SDP is designed to
evaluate inspection findings, performance issues that
are identified at the plant. So for in the case of
amendment requests where a risk analysis is done that
is done using standard risk analysis techniques and is
done by headquarters or other contractors.
CHAIRMAN SIEBER: Okay.
MR. GWYNN: When we get into the
enforcement arena and we're talking about the risk
significance of an issue, then typically that is
extensively discussed at the enforcement conference
with the licensee and differences between our results
and their results are determined as a part of that
pre-decisional enforcement conference.
MR. BROCKMAN: But if it's a regular
conference which is what the new process has, as
opposed to the old pre-decisional enforcement
conference, those same rules apply. Significant
discussion on the risk insights that they gain. In
fact, we've recently had one with Cooper and there was
a lot of subsequent submission of material back and
forth because of inadequacies we found in their
presentation on their risk assessment.
MEMBER APOSTOLAKIS: A related question --
I noticed in the -- in attachment two of our notebook
here, which is the attachment to the letter you
transmitted to Mr. Ray of Southern California Edison.
It says somewhere here that the team concluded that
the risk assessment was conservative. Using the
current leading probablistic risk assessment model in
the San Onofre office safety monitor in Unit 3
condition of core damage probability for the event was
calculated as 1.4 x to the minus four, and the team
noted that the assessment did not take that into
account.
Now, the thing is it seems that you are
using additional risk tools in addition to SPAR and
the SDP --
MR. KENNEDY: Right.
MEMBER APOSTOLAKIS: -- worksheets, and in
this case it was a safety monitor signing off. Now,
has anyone from the agency reviewed this safety
monitor to know what's in it and that it does a good
job calculating core damage probabilities?
MR. KENNEDY: I don't know that there's
been any formal review of that particular tool at San
Onofre, although just to note -- and we'll get into
this -- we also used the safety monitor when we did
the benchmarking trip at San Onofre and compared those
results too. But as far as a formal review of their
safety monitor, I don't believe that's been done.
MEMBER APOSTOLAKIS: But the South Texas
Project PRA has an excellent reputation in the
community, and we were just told --
MEMBER POWERS: They couldn't even get
their success criteria right.
MEMBER APOSTOLAKIS: So, I mean, just
because they have television screens in every room at
San Onofre that doesn't mean that their underlying
models are meaningful.
MR. KENNEDY: And we agree 100 percent
with you, and that's why we don't rely solely on the
licensee's models and tools and information to come up
with a risk assessment. We --
MEMBER APOSTOLAKIS: So in this case you
also did your own calculations, because it says the
core damage probability was calculated at San Onofre?
MR. KENNEDY: Yes.
MR. BROCKMAN: We did. In fact, we used
the -- actually I was only here for the very beginning
of this event and then I was in training the next
week, but we did run this on this SPAR model.
MEMBER APOSTOLAKIS: You did?
MR. BROCKMAN: Yes. In fact, if my memory
serves me correctly, Jack Shackelford had that -- ran
that particular -- was our SRA who did that. Our
process would be -- is any time on a daily basis that
we identify an issue -- an operational issue we get
the SRAs involved with it very early, and for
something like this, a regulatory conference, we would
have our SRAs running their independent analysis. We
would have that being confirmed with insight from
headquarters, research, IIPB, the NRR risk insights so
that we would have a relatively consistent position as
an agency.
This statement here then would be made
because there was a reasonable agreement between the
two numbers.
MEMBER POWERS: I guess I'm curious what
you mean by you ran it on the SPAR model. A SPAR
model's not a fire model. It doesn't have a fire
growth model in it. It doesn't have a smoke model in
it. So what does it mean that you ran this problem?
MR. KENNEDY: Essentially we input the
transient into the SPAR model.
MEMBER POWERS: Yes. But that doesn't --
MR. KENNEDY: The transient that was
caused by the fire.
MEMBER POWERS: That doesn't explore what
the fire could do. That wasn't even questioned.
MR. KENNEDY: It did not explore what the
fire could have done. We evaluated what actually
happened. The transient that resulted from the fire
is what was evaluated.
MR. GWYNN: And that's our typical
approach, including the typical approach of involving
both NRR PRA experts and research PRA experts in
validating our results for those significant events
that they were contemplating to respond to as a result
of our risk assessments.
MR. BROCKMAN: And this is an essential
difference. An event under the new program is
evaluated for what happened, whereas an identified
condition is identified for what could happen.
MEMBER POWERS: We'll come back to that I
suspect. For instance, in one of your findings was
that there were unqualified fire barrier penetration
seals --
MR. KENNEDY: Right.
MEMBER POWERS: -- and a conclusion was
reached that that was not risk significant based on
ignition frequency. I don't really understand
ignition frequencies myself, but when I say I look at
risk significance on a penetration barrier I really
should be looking at the ignition frequencies on two
sides of the barrier, and I should be looking at the
probability if the barrier fails, none of which show
up in most fire protection models and certainly don't
show up in a SPAR model.
MR. KENNEDY: That's correct. A SPAR
model does not model fires, external events, and most
of the fire studies done at the plant are really
screening type studies and not risk studies.
MEMBER POWERS: And most of them assume
100 percent liability of fire bearing penetration
seals.
MR. KENNEDY: Right. That's true.
MEMBER POWERS: And so when you're looking
at the risk significance of a penetration seal it's
going to come up zip.
MR. KENNEDY: It depends on the issue. In
the event where the inspector has identified that a
fire wrap around a cable in a room is degraded or is
not in accordance with the tested configuration --
MEMBER POWERS: I can do that one by hand.
But a penetration -- that's a real risk item. I'm
sure I can do that one by hand.
CHAIRMAN SIEBER: Well, that tells us as
we said in our research report we need to do more work
as an agency on fire, because there's a lot of stuff
that isn't --
MEMBER APOSTOLAKIS: It's not just fire.
It's also a bigger issue here. We've got to move into
risk information inspection processes of the
regulations in general. It seems to me that we are
not spending or paying enough attention to the tools
that we will be using --
CHAIRMAN SIEBER: That's right.
MEMBER APOSTOLAKIS: -- to make these
assessments, and even the SPAR models there is an
underlying computer problem which has never really
undergone any kind of review.
Now of course the situation is not very
bad because you have independent assessments. You use
SPAR. They use -- the licensee uses his own model and
so on, but here is a safety monitor -- people have
been talking about the San Onofre safety monitor for
a long time now, and pretty soon it will be accepted
because we've been talking about it. It's like a
celebrity. You're well known for being well known.
MEMBER POWERS: The other problem --
inconsistency that I see is we plow down through these
thermohydraulic codes worrying about every twitch in
the computer language, and make arguments for
compensating errors and things like that to the third
decimal point --
MEMBER APOSTOLAKIS: That's right.
MEMBER POWERS: -- and then in the risk
assessment tools we say, Well, we use SPAR for a fire
problem.
MEMBER APOSTOLAKIS: There is a reason for
that, because the risk guys are better than the
thermohydraulic system.
MEMBER POWERS: Granted.
CHAIRMAN SIEBER: Let us move on.
MR. GWYNN: I'd like to just mention that
this is a risk informed program. We have very smart
people. We pay them a lot of money to be smart.
MEMBER APOSTOLAKIS: Do they agree?
MR. GWYNN: If in fact there was a
significant potential associated with a fire
protection feature at a plant that could have and
would have significantly adversely contributed to an
event had some circumstance not occurred, some
unplanned and undesigned circumstance not occurred
then we would pay close attention to that, and we can
make regulatory decisions even though the risk numbers
don't quite get us there.
MR. BROCKMAN: That's a good point. All
I want the risk number to do is get me to the
ballpark, and I want it to bring me to the ballpark on
several nights when the game's going to be rained out
too.
MEMBER POWERS: But I think -- I'll accept
that argument. I even like that argument, but here
I'm wondering if it gets you to the entirety of a
ballpark or are you only looking at first base, and
when you've got a tool that you're jerry-rigging to
work on one kind of a problem because you don't have
a real suitable tool for that -- it's not your fault.
You only have the tools that people are willing to
produce for you, but it seems to me that you've got to
squat.
It's the squeaky wheel that gets the
grease in a time of limited resources, which is the
problem the agency has. They've only got so many
guys to generate models that here's an area that what
your challenges -- it's really important. This
affects the way you do your job. This is a front line
problem the agency -- there's nothing the agency
shouldn't be pulling out to address for the guys that
are out on the line doing things. If this is what
they see as a challenge address it. Don't put it off
and say we don't need to do this. If you guys need
these tools you need these tools.
MR. KENNEDY: Let me comment on something
you said earlier. I agree with I think everything you
said. We rely on licensee IPEs that have been
reviewed but not QA'd. We don't get -- necessarily
licensees don't submit updates to their IPEs to us,
and our tools don't -- are not very good, and we'll
get into this more on considering external events. I
think Troy and I agree with you 100 percent.
MR. HOWELL: But I would add that the
exercising of the tools we do have has put the
spotlight on some of these questions.
MEMBER POWERS: Don't get me wrong. My
that goes off to you guys. I think you do a fantastic
job with the tools you have. I just think that
getting you better tools needs to have a higher
priority in the agency and plowing down through
thermohydraulic codes to the fifth decimal point --
it's a useful exercise. Don't get me wrong. And it
may be important, but right now you've got a problem
now, today. Future licensing actions that had to do
with realistic assessments of thermohydraulics are
things that can be put off.
MR. KENNEDY: This slide --
MEMBER POWERS: Not to mention the risk
analysts are better than the thermohydraulics --
MR. KENNEDY: This slide is a summary of
the results of our first three benchmarking trips in
Region IV, and as it turns out the first three in the
country. The only one that has a final report out is
the Diablo Canyon one, but at SONGS -- let me go
through what these mean.
Rev zero indicates the worksheets that we
had issued when we arrived onsite, and we did a
comparison between those rev zero worksheets and the
licensee's model, and by non-conservative I mean that
the SDP came out with a lower color than what the
licensee's model would have indicated, and so 13
percent were a lower color than they should have been.
Twenty-two percent were a higher color than they
should have been, and 65 percent were the same
results. We identified some corrections to be made to
the worksheets, and you can see the final numbers
there, 4 percent non-conservative, 9 percent
conservative, and 87 percent same results.
Keep in mind that the process when we --
if we get a white or greater color we're going to do
a phase three evaluation, so this tool tells us when
we need to go on and do a more detailed evaluation.
The SPAR model --
CHAIRMAN SIEBER: Looks like that is the
worst of the bunch --
MEMBER APOSTOLAKIS: It's very bad.
MR. KENNEDY: Not plant specific.
MEMBER APOSTOLAKIS: Not plant specific --
MR. KENNEDY: It's supposed to be -- they
take aspects of the plant model or the plant
configuration and they put it into the SPAR model, so
it's supposed to be a --
MEMBER APOSTOLAKIS: Well, they have done
30 plant specific -- they developed 30 plant specified
models. Is San Onofre one of them?
MR. KENNEDY: Yes, sir. That's a Rev 3I
no QA done on that model yet.
MEMBER APOSTOLAKIS: Sixty-four percent?
MR. KENNEDY: Yes.
CHAIRMAN SIEBER: Non-conservative.
MEMBER APOSTOLAKIS: That means it may not
be accurately non-conservative. Just disagrees with
the licensee's assessment?
MR. KENNEDY: Yes.
MEMBER APOSTOLAKIS: And it's not that
much better for Diablo.
MR. KENNEDY: Well, it actually is
significantly better.
MEMBER APOSTOLAKIS: Twenty-nine percent
non-conservative. My goodness.
MR. PRUETT: That's non-conservative to
the licensee's model or to the notebook?
MR. KENNEDY: Non-conservative to the
licensee's model.
MR. PRUETT: Okay.
MR. GWYNN: Before you go on to Diablo
Canyon I think it would be of interest to hear whether
this site visit identified any anomalies with the
licensee's model as the South Texas facility.
MR. KENNEDY: None jump out. I don't
remember that there were any. Of course, they use the
PLG model, so it's very difficult to find problems
with those large event models, so --
CHAIRMAN SIEBER: Right. They've got a
lot of chains.
MR. KENNEDY: But in SONGS' case I don't
think we identified anything where the licensee said,
Oh, yes, this is an error in our model that we need to
do something about.
In the Diablo Canyon case you can see the
numbers there. The SPAR results were a little better.
The -- and the final results with the fixes were very
similar.
CHAIRMAN SIEBER: Who's their PRA vendor?
MR. KENNEDY: PLG also.
CHAIRMAN SIEBER: PLG?
MR. KENNEDY: Yes. The first three were
all -- San Onofre is not. Right. So Diablo and South
Texas were PLG.
CHAIRMAN SIEBER: Who was San Onofre, do
you know?
MR. KENNEDY: They used -- I don't know
who their vendor was, but they used the typical small
event tree, large -- see the numbers for Diablo
Canyon? The other thing we looked at that was
beneficial was San Onofre, Diablo, and South Texas --
their models all purport to include some aspect of
external events. And at Diablo Canyon we found that
the affects of fire, flood, and seismic initiators in
some cases increased the results by one order of
magnitude, so for some scenarios, not all, the SDP
would give results that were one order of magnitude
lower than the licensee's model when you considered
external events.
MEMBER APOSTOLAKIS: So Diablo doesn't
have external events?
MR. KENNEDY: Diablo does.
MEMBER APOSTOLAKIS: Does?
MR. KENNEDY: Yes. It does have, and
that's --
MEMBER APOSTOLAKIS: So the 29 percent
refers to -- the licensee did it with external events?
MR. KENNEDY: Yes. No. I'm sorry. Let
me go back. The numbers that you see are internal
events only.
MEMBER APOSTOLAKIS: For Diablo?
MR. KENNEDY: For Diablo.
MEMBER APOSTOLAKIS: And the South Texas?
MR. KENNEDY: And -- well, South Texas is
two numbers, but at Diablo the external results are
not listed but the words there indicate that it's kind
of a summary that -- for those -- we found up to an
order of magnitude difference when you considered
external events.
MEMBER APOSTOLAKIS: I was always under
the impression that by using the worksheets you would
be getting very crude results and that you should be
using PRA models, but this SPAR thing now --
CHAIRMAN SIEBER: It's the other way.
MEMBER APOSTOLAKIS: It's the other way.
CHAIRMAN SIEBER: That's the way it looks.
MEMBER APOSTOLAKIS: And both for Diablo
and San Onofre I would rather go with the sheets.
MR. KENNEDY: Yes. A couple of things
about the SPAR model though. They -- we don't rely on
them too much right now for this reason, because we
don't really trust the numbers that we're getting, and
so --
MEMBER APOSTOLAKIS: But the worksheets
are also based on SPAR, aren't they?
MR. KENNEDY: No. The worksheets are
based on the licensees' IPEs.
MR. BROCKMAN: One thing to look at
here -- let's look at the worksheets revenues with the
fixes. At SONGS we would basically be saying that 91
percent of the time -- that's the 87 plus the 4, the
regulatory posture -- 87 percent of the time the
regulatory posture that we would propose off the
worksheets would be what we would anticipate would be
the licensee agreeing to for the reg conference.
The key thing -- look at Diablo. SDP is
conservative. Thirty-six percent of the time the
results of our regulatory conference would be to
decrease the significance of the issue. Now, that's
great from the aspect that we're looking at
everything. It certainly can result in a public
relations challenge.
MR. HOWELL: Which it's why it's important
to do more than just exercise the worksheets before
you ever get to that point.
MR. KENNEDY: What we typically do is when
we -- and typically we haven't done a lot of these,
but if we come out with some results greater than
green on the worksheets the first place I don't go to
is -- I don't go to SPAR the first thing. I go to the
licensee's IPE and make sure I have enough data at IPE
and I'm looking at the systems they have and what
their risk achievements are for those systems and --
MEMBER APOSTOLAKIS: But why when the
office of research comes to us and they advertise SPAR
as a major achievement they never tell us this?
MR. KENNEDY: I think they use SPAR -- I
don't want to be put in the position to defend
research, but I'll provide some defense.
When they use these SPAR models they use
them for accent sequence precursor evaluations, and
they are much more skilled in going into the model and
making changes to the model than most SRAs are, so
they actually get into the model and do a lot more
manipulation, do a lot of research to determine the
proper way to model whatever they're trying to model
and use it for that.
MEMBER POWERS: I come back to my
thermohydraulics. We don't let people do that in the
thermohydraulics code. That code -- you can't change
anything once it's been approved, and it doesn't do
you -- it doesn't help you to get a model that has to
be tweaked to get the right answer.
MR. KENNEDY: We would agree.
MR. PRUETT: We agree. Kriss can speak
for himself, but from my perspective I'd like to see
more time spent on developing the SPAR models,
improving the end-user interface so that I don't have
to make significant manipulations to the model. I can
point and click on certain basic events and initiating
event categories and get a reliable answer. Right now
I can't do that.
MEMBER POWERS: You've got a full-time
just interpreting the results.
MR. PRUETT: That's right.
MEMBER APOSTOLAKIS: Now, why shouldn't
the agency demand that every licensee do a complete
level to PRA? How much is it? Is it the million
dollars? Big deal. Look at the --
VOICE: Level two?
VOICE: Big deal to you.
MEMBER APOSTOLAKIS: Well, look at all the
uses. We have to fight and try SPAR, and there is
nothing and do this and do that. If we're going to
have risk informed regulations we should have good
risk assessment tools.
CHAIRMAN SIEBER: The risk informed
regulations is optional for the licensee.
MEMBER APOSTOLAKIS: Right.
CHAIRMAN SIEBER: And so you can't make
him do something that's optional.
MEMBER APOSTOLAKIS: Speaking of optional,
can they tell you do not use the revised oversight
process when you inspect us, oversee us? Can they tell
you that? So it's not optional.
MR. BROCKMAN: Yes, they can.
MEMBER APOSTOLAKIS: They can?
MR. BROCKMAN: They could do that.
MEMBER APOSTOLAKIS: But has anyone done
it? No.
MR. BROCKMAN: The only thing that was
done Cook as they were coming up said we're not quite
ready yet. We don't have the data. They were
captured in O-3 process, that we need to get our
baseline going and they wanted about a six-month delay
in getting into it because of the lack of
historical --
MEMBER APOSTOLAKIS: First of all, it's
not a million dollars because they've already done the
IB. We're talking about documenting the IB, having a
serious review of it, and then all these issues are --
MEMBER POWERS: If you're talking about a
level two.
MEMBER APOSTOLAKIS: That's what we're
using.
MEMBER POWERS: I don't think you can get
a level two done for a million dollars, and you
certainly can't get one that anybody would agree with.
MEMBER APOSTOLAKIS: You can get a full
level three for a million and a half, so --
MEMBER POWERS: You can't get one that
anybody will agree with.
MEMBER APOSTOLAKIS: What, because of the
nature of the severe accident -- those are you guys.
MEMBER POWERS: But --
MR. GWYNN: The South Texas Project folks
tell me that they spend about a quarter of a million
dollars a year just maintaining their PRA, and so the
initial cost is not the entire picture. But whether
or not the licensees are required to have level two
PRAs is a matter of policy that we don't have -- it's
not our decision, and so --
MEMBER APOSTOLAKIS: I understand that.
Sometimes these simple questions come to you and you
say, Gee, why didn't I think of that? Here we're risk
informing a lot of things, and yet we are willing to
leave with models that have not been reviewed, that
are incomplete, and everybody knows that, and the
question is why? I can see a reporter asking that
question if there is a nuclear incident some place.
You're doing all this and you don't have the
underlying tools.
CHAIRMAN SIEBER: Well, this is why it's
risk informed instead of risk determined.
MEMBER APOSTOLAKIS: It seems to me if
it's risk informed you should be able to assess the
risk to the best of your ability.
MR. GWYNN: If you look at the nuclear
power industry historically when we first started down
this road we would never have built the first power
reactor if we took the approach that it's got to be
perfect before you build the first one, and so these
tools are being improved over time. The question is
whether or not they're adequate for the thing that
we're using them for today. And I think that
they've -- based on the results that we've achieved
over what we had before and what we have now I think
that we've seen an improvement as a result of
implementing this tool --
MEMBER APOSTOLAKIS: There's no question
that there's an improvement. It's just it's kind of
odd we don't have the right tools.
CHAIRMAN SIEBER: Well, we know that, and
we have determined that we don't know how much they
cost.
MEMBER APOSTOLAKIS: No, no. We know very
well.
MR. KENNEDY: Not to add fuel to the fire,
if you look at South Texas, when we -- this was the
third visit made in the country. We showed up in
South Texas with the rev zero worksheets and found
that there was a fatal flaw in the worksheets. They
considered -- the worksheets contained a mitigation
strategy for high pressure recirculation that South
Texas doesn't do, so we couldn't run through the
samples using the worksheets as --
MEMBER APOSTOLAKIS: Wait a minute. The
worksheets we were told come from the IP.
MR. KENNEDY: Yes.
MEMBER APOSTOLAKIS: And the IP for South
Texas is really a PRA, so how come there -- the PRA
itself had this flaw?
MR. KENNEDY: No.
MEMBER APOSTOLAKIS: It was in the
translation?
MR. KENNEDY: It was in the translation.
Yes. So we did run a revision zero, but that was a
fairly easy fix. We did it onsite and corrected the
worksheet and ran the examples through. The number in
parentheses compared the results considering external
events to the worksheets, and that's what those
numbers are.
CHAIRMAN SIEBER: Well, I guess I have a
question then. It would appear that we got better
results for South Texas than other places. It also --
MR. KENNEDY: Well, in what area?
CHAIRMAN SIEBER: Well, in comparison
between worksheets and their PRA.
MR. KENNEDY: Okay. But keep in mind the
South Texas -- the only numbers we have for South
Texas are the final numbers. Those are after the
changes were made onsite.
MR. PRUETT: Yes. The high pressure re-
cert was not the only change made.
MR. KENNEDY: Right.
CHAIRMAN SIEBER: Okay.
MR. PRUETT: There were several that we
made as we made a high pressure re-cert change.
MR. KENNEDY: Right. And so what we're
missing is the rev zero which would have been just
terrible.
MEMBER APOSTOLAKIS: Diablo looks very
good. Read the fixes.
MR. KENNEDY: Yes. Diablo looks good, and
SONGS doesn't look too bad.
MEMBER APOSTOLAKIS: Tom told us earlier
that SDP conservative means that you go into
conference with the licensee and you find that 36
percent of the time for Diablo for example you back
off. You were conservative.
MR. KENNEDY: Well --
VOICE: Maybe.
MEMBER APOSTOLAKIS: So 15 percent of the
time then the licensee tells you, No, Mr. Regulator,
you are not conservative enough so you have to give us
a white instead of a green?
MR. KENNEDY: No.
MEMBER APOSTOLAKIS: Is that what it
means?
MR. BROCKMAN: No. In fact that's really
the type error that we need. Our goal has to be to
get that to zero, because --
MEMBER APOSTOLAKIS: No. But what does it
mean?
MR. BROCKMAN: -- the potential exists
there that I am not going to pursue a white issue
because I come up with a green determination. My goal
on that has to be to get that number to zero, and
that's the challenge. I never want to have an issue
that I don't pursue because I have underclassified it.
I need to get that to zero but on the
contrary my public relations dilemma is the other side
of the coin. I don't want to have too many times
where it looks like all I do is back off, and I get
the reputation of not being an effective regulator.
I cut deals in dark, smoke-filled rooms. And there
are certain people out there right now who make those
accusations.
MEMBER POWERS: Then they've got type one
and type two errors.
MR. BROCKMAN: That's it. Type one-type
two errors traditional.
MEMBER APOSTOLAKIS: But you actually find
out if the licensee's assessment was worse -- the
result was worse than yours?
MR. KENNEDY: No. Let's step back a
minute. The only thing we're really concerned about
is do we come up with a green on the worksheet that is
really white?
MEMBER APOSTOLAKIS: What do you mean,
really white? There isn't such a thing as really.
MR. KENNEDY: Well --
MEMBER APOSTOLAKIS: Somebody else's
assessment is white?
MR. KENNEDY: Yes.
MEMBER APOSTOLAKIS: Okay.
MR. KENNEDY: The worksheets are
underestimated the risk, the actual risk --
MEMBER APOSTOLAKIS: Right.
MR. KENNEDY: -- and so the results of the
worksheets are a green, and in our process we don't do
anything. We do some other things, but we don't go to
a reg conference. We don't engage on further risk
analysis.
But right now if we do come up with
something greater than green, a white, yellow, or red,
we don't go straight to the reg conference based on
the results of the worksheet. We engage their risk
analysts onsite and do a phase three type analysis to
determine what the risk really is. So we would avoid
this 36 percent downgrade in the color even before we
went to the reg conference because we're doing that
phase three analysis.
MR. PRUETT: Right now I'd say about half
of that 36 percent that Kriss is talking about is due
to the way we implement the county rule in the
significance determination process, so if we have
three greens adjacent to a white block we're going to
call that white finding. In reality it may really be
a green finding, but for the purposes of the phase two
analysis we're going to call that white.
MEMBER APOSTOLAKIS: So you're referring
to the action matrix?
MR. PRUETT: That's correct.
MR. KENNEDY: No --
MR. PRUETT: Not the action matrix.
VOICE: The SDP --
MEMBER APOSTOLAKIS: That takes you to the
headings of the action matrix. Isn't that the same
thing?
MR. PRUETT: Well, no. You've got the
greens next to whites. You're right. The output from
that would take you as to where you start going in
the --
MEMBER APOSTOLAKIS: Are you happy with
the headings? I think they're very arbitrary, but two
whites or three greens or -- do these make sense? And
then all of a sudden the last one -- this is changing
the subject a little bit, but I don't think we
discussed it at all.
MR. BROCKMAN: Well, there was --
MEMBER APOSTOLAKIS: What's the basis?
MR. BROCKMAN: The one thing with three
greens next to a white was to try to prevent the error
of missing one. It's too close and we know there's
uncertainty in our tool, and if we come up with three
greens next to a white we say we're going to pursue
further. It's like a performance indicator. I don't
know there's a problem but I need to look further
because I'm in my uncertainty band, and that's where
we're trying to -- should it be three next to a white?
Should it be two next to white? We started with
three.
MEMBER APOSTOLAKIS: All right.
MR. KENNEDY: If you go to the next slide,
Troy, I think we've discussed almost all the
challenges that I have listed here. By challenges I
think these are challenges that Troy and I faced that
regional management faces and the inspectors face out
in the field, and that is the accuracy of the SDP
phase two worksheets.
We have to sit down -- the inspectors
implement the phase two worksheets. They fill they
out, and they have to sit across the table from the
licensee, and if there's errors in those worksheets
that the licensees are pointing out to them that's not
desirable. And the second one, availability and
accuracy of the SPAR models, we've discussed that.
And to get on the question that you asked
earlier, Dr. Powers, the tools that we have for fire
protection shutdown operations and containment
integrity, in the case of the last two those are
really under construction. There's procedures out
there, but what you -- they're really screening
procedures that you end up going back to NRR whenever
you have some issue, and the fire protection SDP is
probably harder than it needs to be.
MEMBER POWERS: I don't even understand
it. You come in here and you say, Okay. Is the
manual fire question capability degraded a little bit,
half way, a bunch. I have no idea, but having made
that determination then I start -- I get an exact
number.
MR. KENNEDY: Right.
MEMBER POWERS: That turns out to be an
exponential. Now, there's a numerical error in it,
but that's okay. We get these numbers out. I have no
idea how to do that.
MR. KENNEDY: We share the same
frustration.
MEMBER POWERS: I don't even know where
the exponential numbers are. I know exactly where
they come from. They come from five, but that doesn't
help me. Where did five get them?
MR. KENNEDY: And the numbers that you get
from five are screening values and they don't really --
MEMBER POWERS: And they did things that
I think are obnoxious in fire protection modeling.
MEMBER APOSTOLAKIS: That's another
mystery to me, again, and it has to do with these
simple questions I mentioned earlier. Why did most of
the licensees choose to do a screening analysis for
fires when we have all this risk informed regulatory
system facing us? Very useless. You just screen
things out and say they're not important. How does
that help me implement a significance determination --
I don't understand these things.
MEMBER POWERS: Whenever they have an
inspection finding you tell them it's green because it
got screened.
MEMBER APOSTOLAKIS: It got screened out.
MEMBER POWERS: It doesn't matter if the
fire protection seals all fail and it's going to be a
roaring inferno in there in the event of a fire, but
that's -- it's screened.
MEMBER APOSTOLAKIS: Okay.
MEMBER POWERS: The fire's smart. It
knows. It goes around those --
MR. KENNEDY: But in all these -- in these
three areas in particular NRR does have some projects
going on to further develop the shutdown SDP, the
containment integrity SDP --
MEMBER POWERS: Right.
MR. KENNEDY: -- and I'll be honest with
you. Their efforts on the appendix F improvements --
I'm not sure they're headed in the right direction,
but they are trying to do something with it. From
what I've heard it doesn't simplify the process
though. I think it goes from 60 pages to 100 pages,
but --
MEMBER POWERS: -- as long as I'm just
rolling dice and guessing at a number to begin with.
CHAIRMAN SIEBER: It seems to me these are
areas where we have to pay a little closer attention.
MEMBER POWERS: There's no question about
it. We're getting the same story from both sides of
this coin, and -- all apologies, Kriss. You're not
the first to tell us this.
MR. KENNEDY: I'm glad. I didn't think I
was.
MEMBER POWERS: And so when we prepare our
September report to the commission -- they've got to
understand what's going on, and I like this. It's
challenges to the one guy -- one set of people that I
really don't want to throw any more challenges to, and
that's the guys that are out in the front line dealing
with the plants, and then they should go in with a
measure of confidence that what they're doing has a
good technical, sound foundation, that the
uncertainties in it have been examined fairly closely.
I don't think it's a fatal flaw, but I
think it's an issue of priorities.
CHAIRMAN SIEBER: Do any other members
have questions?
(No response.)
CHAIRMAN SIEBER: Well, thank you, Kriss,
for your discussion and I would point out that even
though this has been more dialogue than presentation
so far, this method is important to us to get a really
good insight in a short period of time as to what your
problems are and how do you perceive the operation of
the agency.
What I'd like to do is we are on schedule
if we ignore the fact that we have not covered topic
five. What I'd like to do is perhaps go until 12:15
rather than 12:30 for lunch. We can gain at least 15
minutes in the process and so I would suggest we break
for lunch right now.
MR. GWYNN: If I could I'd like to ask the
Region IV staff to allow our guests to go first for
lunch, and the lunch is in our executive conference
room just around the corner here. We'll go in, pick
up our lunch, then come back and eat it here if that's
all right.
CHAIRMAN SIEBER: Fine.
(Whereupon, a short recess was taken.)
A F T E R N O O N S E S S I O N
(12:20 p.m.)
CHAIRMAN SIEBER: I think in the plant
operations area I think a number of us have
questions about the general topic of Callaway grid
experience and how that impacts other plants. We're
aware the information notice that was published in
the incident in 1999, but you may want to give us
some insights as to what your expectations are for
the future under the burn energy situation and what
it is Region IV is doing about it.
And so with that I will turn it back to
regional management for their next presentation.
MR. BROCKMAN: Thank you, sir.
We're really in what I'll call our segue
transitional part here of moving along and focusing
on the electrical part and then we'll be moving into
the fire protection part. The first thing we want
to do is share with you a little on the SCRAM
trends. This will be very quickly. This is a
transitional issue.
As we've looked over the last couple of
years as to what have been the trends that we have
seen in our SCRAM data and what have you and the
insights we're getting and how that's trying to
focus us in different areas, and you're going to see
it's going to lead us right into this afternoon's
topic.
So with that, Bill Johnson, who is my
chief of the Branch B in reactor projects which just
happens to be where Callaway resides --
MR. JOHNSON: This is some data that was
put together by regional personnel on total SCRAMs
across the nation for years 1998, '99, and 2000. I
don't see any distinct trends from this presentation
of the short-term SCRAM data. I did notice one
interesting point that the number of manual SCRAMs
in year 2000, 33, was the same as the number of
manual SCRAMs in here 1999, also 33, which indicates
that the new performance indicator which counts both
manual and automatic SCRAMs might not have had much
of an effect on the number of manual SCRAMs. It's a
good sign.
Since we noted that a number of the
SCRAMs in Region IV were caused by electrical
systems a further review was performed, and later on
the agenda Mr. Pruett will summarize the results of
that review.
CHAIRMAN SIEBER: Just a quick question.
Licensees complain that including manual SCRAMs
prevents or induces an operator to try to wait it
out as opposed to taking a safety protective action
before an automatic action occurs, which potentially
might not occur as we would like it. In view of
that is there any consideration or any thoughts that
you would have about counting manual SCRAMs and the
total number of scams as an unintended consequence
or an unintended driver to rely more on the
automatic action rather than the operator's
intuition?
MR. BROCKMAN: In fact, I think an
accurate characterization is is there were two or
three individuals placed in the industry who
expressed a personal concern that this could be an
unintended consequence. Across the board in all of
the trips that I think we have taken out to our
licensees they have unequivocally stated, No. This
performance indicator would have absolutely no
impact on the intent of their operators and the
actions of their operators.
It was a couple of people who said this.
MR. GWYNN: Every licensed operator that
I've spoken with in a control room and asked that
question of has said, I'm going to follow my license
requirements and my boss is going to be very upset
with me if this thing goes out automatically when I
should have punched it out manually, and it has --
the performance indicator had no bearing on their
thinking in that arena, and the data that Bill just
put up I think supports, at least during the first
year of initial implementation that there hasn't
been an impact.
MR. BROCKMAN: But with that said, NRR
is revising the performance indicators to preclude
that. There's activities going on to revise it and
get it into an arena where that potential supposedly
could not even exist.
CHAIRMAN SIEBER: Another quick
question. Are there any other performance
indicators that come to your mind like the counting
of outage hours and certain risk conditions that
might have an unintended consequence?
MR. BROCKMAN: Yes. Probably the one
that comes to my mind most easily is unplanned power
reductions.
CHAIRMAN SIEBER: Okay.
MR. BROCKMAN: Currently there is -- it
was the old AEOD performance indicator that had
absolutely no risk association to it but was without
a doubt the highest correlation factor toward those
plants that degraded in the NRC's overall
assessment.
For plants that had unintended power
changes, unplanned power changes, the more they
occurred it wound up being that those were the
plants of concern. Not anything to do with risk.
This was brought forward in the new program.
Without a doubt you have the what is an unplanned
power change? Are you talking about an automatic
run back? Are you talking about a condition evolves
and I've got to take action within the next six to
eight hours to reduce the power to make that happen?
In the old AEOD performance indicator
that would have been an unintended power change,
doing it within that time, but currently the way the
performance indicator is done is any power change
done within 72 hours is an unplanned power change.
Give you adequate time to get all your things
together, plan the activity, prep your people, and
embedded into more of your normal processes. If
you're a utility and you've got the choice of doing
this at hour 68 or at hour 73 it's a no-brainer.
You're going to do it at hour 73.
CHAIRMAN SIEBER: If I have a --
MR. BROCKMAN: We have seen indications
where decisions are being made -- now, they're being
risk considered into it, but if risk is not an issue
and they have a choice of doing it in less than 72
hours or quicker or after 72 hours, they're doing it
in longer than 72 hours so they don't take the PI
hit.
CHAIRMAN SIEBER: So if I have a small,
below tech specs reactor cooling system leak in a
joint, which is allowable, I should allow it to leak
for 72 hours before I go in and do something about
it?
MR. BROCKMAN: I'm not sure that they
would take it at that particular point, but we've
had -- and your memory is always better on these
things than mine where once again, if risk isn't an
issue, if the tech specs aren't an issue, and if
I've got reactor cooling system leakage I'm going to
be in a short action statement there, but if it's a
valve packing leakage, which we know is right there,
and I've got a choice of reducing the plant down
tomorrow night or waiting until Saturday night to do
it, they'll probably figure two things with respect
to that, and that's going to be with the load, the
system load is requesting on -- they'll factor that
in there, and then they'll look at that outage time
too on the hit for the PI.
CHAIRMAN SIEBER: Yes. Well --
MR. BROCKMAN: And if they don't think
it changes their risk profile they'll wait.
CHAIRMAN SIEBER: The reason why you do
it is for ALARA, and the reason why you don't want
the leak to stay there for 72 hours is because leaks
never get better. They always get worse.
MS. WESTON: Are there any plans to
change that possible consequence?
MR. BROCKMAN: They're looking at that
one, but I don't know what --
VOICE: That's one that's being
reviewed. The power reduction is being reviewed.
I'm not sure whether there's a work force on it.
I'm not sure exactly --
VOICE: That one could be manipulated
two ways. One is a 72 hour and the other is whether
or not you go to 81 percent or 79 percent, because
the cutoff is 80.
MR. BROCKMAN: And that becomes an ALARA
consideration too, and that's one thing they used to
take it down to 75 and say, If I've got no
additional ALARA --
CHAIRMAN SIEBER: Okay. Thank you very
much. You may go on.
MR. JOHNSON: I pulled a couple of
trends graphs out of SECY-01.0111 just because I
thought they were interesting and probably worth a
quick demonstration. And overall there aren't any
industry trends that seem to be heading in the wrong
direction.
For ASP program results there were no
significant precursors in fiscal year 2000, and it
looks like an overall downward trend in the overall
number of the precursors.
Looking quickly at some of the ex-AEOD
indicators the one for automatic SCRAMs overall
trend of course is still down. We've noted on this
one as well as on the first slide in 1999 there was
an increase. I don't know exactly what that means,
but it still fits within the expected boundaries.
Safety system actuations also down.
Looking at a couple of the raw
performance indicators I wanted to look at unplanned
SCRAMs per 7,000 annual critical hours. Don't see
much of a trend on that, but this is short-term data
and you couldn't draw a very firm conclusion from
it. Scrams with loss of normal heat removal -- I
still don't see a trend there either, but it will be
interesting to see this data accumulate for a few
years and see if it tells us anything.
And the other one I wanted to look at is
safety system failures. I do think I see a trend
there, even though it's short term. That's for
PWRs. And the similar curve for boiling water
reactors -- there's a similar possible trend that a
statistician could figure out.
And that's the ones that caught my
interest. We're open to questions if you have any,
sir.
(No response.)
MR. JOHNSON: Okay. Thank you very
much.
MEMBER POWERS: It seems to me that the
question that arises, especially when we look at
what the risk significant thresholds for PIs are
that we've really chosen PIs that are too limited.
It's really combinations of things together that are
really the PIs that we want. Unplanned SCRAMs --
that frequency combined with frequency of something
else is really the indicator that we want to have.
Do you have any thoughts on that?
MR. JOHNSON: I'm not well versed on
that, but I do know that the unplanned SCRAMs in
itself does not have a lot of risk significance, but
the unplanned SCRAMs with loss of heat removal might
well have serious significance, and that might be
one to watch more closely.
MEMBER POWERS: I'm wondering about more
complicated combinations. When you go through and
you come out and you find out I've got to have 19 or
something like that unplanned SCRAMs to get to a red
level, you know that's never going to happen. It's
just looking at the wrong thing, because that
particular measure is just in itself not risk
significant, but it's some unplanned scams -- a
couple is something else -- where having one might
get you certainly to a white.
Is there --
MR. HOWELL: That's why we look at every
one to see --
MEMBER APOSTOLAKIS: If we had a good
safety monitor and calculated the core damage
frequency every time we have something happening
then that would be a good indicator, would it not,
because then you could set it at levels of CDF, and
you don't care how you got there. It could be a
combination of ten things.
MR. HOWELL: And that's why --
MR. BROCKMAN: True. That's why we look
at it on the front end.
MR. HOWELL: Yes. Our inspection
threshold looks at the CDP that comes up there that
instant. Basically, that instantaneous
probability --
MEMBER APOSTOLAKIS: No, because when
you do the SDP and performance indicators really the
thresholds are such that the change in that
indicator would cause a level CDF greater than some
threshold. Not a combination.
MR. HOWELL: Correct, but we do look at
that on the front end for events, and even
conditions too. So Kriss and Troy, they do that,
using the tools that we have we talk to the
licensees and we'll ask San Onofre, What does your
monitor indicate, and if it trips the threshold
the --
MEMBER POWERS: Then you didn't believe
him.
MR. HOWELL: You have to get the
information the best you have.
MEMBER POWERS: Well, they came back
with 1.4 times ten to the minus four, and you said,
We don't believe that. That's way too conservative.
MR. HOWELL: But we still did a special
inspection though. We sure did.
MR. BROCKMAN: You've got two different
things. What you bring up here is very interesting
to the performance indicator, but as I tried to say
earlier, the inspection is without a doubt still a
critical component, and we'll look at exactly that
for an event or condition that occurs.
And this weekend you saw the 5072s where
the potential transformer at San Onofre that
disassociated itself all over the Pacific Coast
Highway, and we also had one at Cooper.
So we took -- the risk guys looked at
that right away. Where are we at on that thing --
the startup transformers lining out out at Cooper.
Well, it becomes a risk interesting issue if that
startup transformer is out five days. They're at
about two and a half. Are we monitoring that as
we're correcting?
We're inspecting right now on it, and if
they get up to five days with the other issues that
identify themselves in some other areas there we'll
definitely be looking at changing that inspection
threshold, which then gives us an additional vehicle
to identify the issues that we've been talking about
corrective actions and things so we can get those
insights.
MEMBER POWERS: I know what I want to do
for sport on the 4th of July. I want to get an
inspector proponent like Ken, lock him in a room
with a risk guy like George, and see who comes out
alive. I've had numerous discussions with some of
the staff risk guys.
MEMBER APOSTOLAKIS: If the safety
monitor could be trusted that would be the best
method, really, to core damage treatment, the
condition of core damage probability, but
unfortunately, we can't trust it.
MR. KENNEDY: But there's also a
deterministic aspect to the threshold that's been
picked for SCRAMs, and that is it's a pretty good
indicator irrespective of risk that if you have too
many there's a problem at that site, and --
MEMBER APOSTOLAKIS: So what do I care
if it's an element of risk? Ultimately it has to be
connected to risk. Right, because we are
regulating -- protecting public health and safety.
If they want to lose money, that's their business.
MR. KENNEDY: There's a lot of
deterministic SDPs out there though, and several of
the SDPs are deterministic.
MEMBER APOSTOLAKIS: Well, there
wouldn't be if you had a very good reliable safety
monitor.
MEMBER POWERS: Well, don't get over
enamored with this risk analysis. There are other
issues.
MEMBER APOSTOLAKIS: Like?
MEMBER POWERS: Like sabotage, site
security that you can invest in that, and there are
elements not only of the regulations but of the
oversight program that address those things. And as
I often say to you when we discuss defense in depth
even if the probability of event is low if it occurs
I'd really like something between me and the bad
stuff.
MR. BROCKMAN: My residents will all
echo that.
I think next up is Ms. Good, who is our
plant support branch chief, to talk about the
Callaway ALARA issue which we agreed to wait until
now to discuss.
MS. GOOD: Thank you.
Good afternoon. My name is Gail Good.
I'm the chief of the plant support branch here in
Region IV. I am responsible for reactor inspections
in the area of security, emergency preparedness, and
radiation protection, and my presentation this
afternoon will focus on the radiation protection
area and specifically on some problems that were
identified at the Callaway Plant in Fulton, Missouri
that involved their ability to implement their ALARA
program. And ALARA stands for as low as reasonably
achievable.
My presentation will cover the findings
that were identified during the initial inspection,
the specific performance problems that were
associated with the findings, the NRC's assessment
of the findings, and that would be the significance
using occupational radiation safety significance
determination process and any enforcement issues.
It will cover the licensee's response to the
decisions that we made and then the NRC's actions to
address the licensees' appeals, and then finally
I'll discuss the special follow-up with the
supplemental inspection that we conducted.
In August of 2000 Region IV conducted a
baseline routine inspection of the licensee's ALARA
program. That inspection focused on a review of
jobs that were completed during refueling outage ten
that was in 1999. Specifically we reviewed those
jobs where the actual job doses exceeded the
projected job dose by greater than 50 percent and
accrued more than five person rem, and based on that
review we identified six jobs that exceeded that
criteria.
CHAIRMAN SIEBER: Just a real quick
question.
MS. GOOD: Yes.
CHAIRMAN SIEBER: If I were the RCM at a
plant and I knew you were going to operate this way
why would I not fudge the estimates so that I
couldn't miss? Do you have a way of looking at
absolute values?
MS. GOOD: We have a way of looking at
their justifications for the projected doses that
they're assigning, and if we see a significant
increase from doing a similar job in a previous
outage we might question why they were saying there
would be an increase in the projected dose for this
particular job. So we would be reviewing their
justifications.
CHAIRMAN SIEBER: But you would be on a
different kind of philosophical framework that way,
saying, I don't really have great confidence in the
way you're doing your estimates, as opposed to the
numerical issue of you're double what you said you
were going to be.
MS. GOOD: It's a concern that we have.
CHAIRMAN SIEBER: Thanks.
MS. GOOD: And so with respect to the
six jobs, the six jobs included all of the
scaffolding work that was done in the reactor
building. That was all considered to be one job,
and the actual dose for that job was 46 person rem.
The second job was the removal and installation of
the steam generator manway covers and inserts, and
the actual dose for that job was 8.5 person rem.
MEMBER LEITCH: My question here is are
we talking about bad estimates or bad performance?
MS. GOOD: Bad performance.
MR. GWYNN: As a matter of fact, there's
a screening criterion that says that if these
conditions exist but the overall ALARA results for
the facility are good then we don't pursue them.
Correct?
MS. GOOD: We would expect that there
would be a performance problem. Our initial look at
it is for those jobs that are greater than five rem
and where they exceeded the projected dose by
greater than 50 percent, and we're using that
greater than 50 percent as a filter to say we need
to go out and take a look at these jobs to determine
if there is a performance problem associated with
it.
MEMBER LEITCH: So it's just not that
the job proceeded along an unexpected course but
there were some performance deficiencies --
MS. GOOD: Yes. There were performance
deficiencies.
CHAIRMAN SIEBER: It also would seem to
me though in the process of estimating -- and I'm
thinking like a licensee now -- if I would project,
for example, scaffolding erection to be 20 man rem I
would automatically have at least six jobs called
scaffolding erection. Okay. And --
MS. GOOD: They actually had -- I think
it was about 160 individual scaffolding tasks.
CHAIRMAN SIEBER: At one job.
MS. GOOD: But they considered it to be
one job and the ALARA planning and controls were
done at that higher level, and that was one argument
that the licensee tried to make when we had the
regulatory conference was that we really should have
been looking at the individual scaffolding work
tasks.
CHAIRMAN SIEBER: The licensee should
have been planning at the lower level.
MS. GOOD: And that was the argument we
made, that there weren't sufficient ALARA planning
and controls established at the level they wanted us
to look at.
CHAIRMAN SIEBER: Right. Thank you.
MEMBER POWERS: Will you give me a
feeling for the context? This is all part of one
refueling outage?
MS. GOOD: Yes, it was.
MEMBER POWERS: And what was the
duration of that refueling outage?
MS. GOOD: I don't know.
CHAIRMAN SIEBER: Roughly?
VOICE: About 35, 40 days.
MS. GOOD: About --
VOICE: It was a little bit longer,
right, because of the -- went over -- 40, 50 days.
MEMBER POWERS: We see a lot of this I'm
going to set the record for outage for this kind of
plant, or I'm going to break my current record,
things like that. We've got a whole dose of it at
Waterford. This is -- I'm happy for them to have
good planning and do their outages quickly, but this
setting record business is going to lead to this
kind of problem.
CHAIRMAN SIEBER: But generally when the
outages get shorter the man rem expenditures get
lesser.
MR. HOWELL: Yes. But that didn't
happen in this case.
MS. GOOD: In some cases.
MR. HOWELL: But that was one of the
arguments that they said. We took into account as
part of our planning. We want to have a shorter
outage. We'll do the hotter work early in the
outage and then we'll get done quicker and the
overall cumulative dose will be less, but that's not
what happened.
CHAIRMAN SIEBER: This is one of the
snupps plants?
MR. HOWELL: Yes.
CHAIRMAN SIEBER: Did they use the hot
boron injection to try and get the source turned
down?
MS. GOOD: I'm not sure they did.
MR. HOWELL: I think so, but they
were -- I don't know, but they were doing work
before they cleaned up the RCS. They were erecting
scaffolding before they cleaned up the RCS. They
were --
CHAIRMAN SIEBER: That sort of explains
it.
MR. HOWELL: Right.
VOICE: And their source terms was
complicated by the anomaly that they had --
MR. HOWELL: And they were trying out
electrosleeving of the steam generator tubes for the
first time, new technology here in the states, and
it had complications which contributed to some of
this.
CHAIRMAN SIEBER: But none of those were
scaffolding, and scaffolding was 40 something man
rem?
MR. HOWELL: Yes.
CHAIRMAN SIEBER: Okay.
MR. HOWELL: A lot.
CHAIRMAN SIEBER: That's a lot. That's
two outages.
MR. HOWELL: Steal some of Gail's
thunder -- to cut to the chase, they went from 305
man rem in refuel ten to 100 in refuel eleven as a
result of corrective actions --
MS. GOOD: So they can do it. It can be
done.
CHAIRMAN SIEBER: I apologize for
interrupting.
MS. GOOD: All right. Moving along with
the jobs, the third job that I have listed here is
the eddy current testing, the robotic plugging, the
stabilizing, the electrosleeving, and that job
actually was the highest, and it accrued a 58 person
rem.
MEMBER UHRIG: How much of that was
electrosleeving were normal procedures?
MS. GOOD: I don't have that figure off
the top of my head because they lumped all of that
together under one job, under one RWP, and I can
attempt to get that but I don't have that answer for
you right now.
The fourth job was the health physics
support for the primary and secondary steam
generator activities, and the actual dose for that
job was 5.6 person rem. Fifth job was the foreign
object search and retrieval, and the actual dose for
that job was 6.4 person rem --
CHAIRMAN SIEBER: That was one steam
generator?
MR. HOWELL: I think it may have been a
couple of objects that they dropped in --
CHAIRMAN SIEBER: But they went in
through the -- where the flow blocking device is?
Most of that was probably extremity. Right?
MS. GOOD: I don't --
MR. HOWELL: We'll have to get the
report. It may have actually --
MS. GOOD: I think we had that --
MR. HOWELL: -- been during refueling.
I don't know if it was necessarily the steam
generator. It may have been the --
CHAIRMAN SIEBER: It must have been
extremity dose?
MR. HOWELL: I can get you the report.
MS. GOOD: I'll move along then.
As I mentioned, the sixth job was the
reactor coolant pump seal removal and replacement,
and the actual job dose for that was 13 person rem.
And again, I'd like to point out that all six of
these jobs exceeded that filter that we use for
focusing our inspection activities, that they were
all over five person rem and they all exceeded the
dose projection by greater than 50 percent.
CHAIRMAN SIEBER: Industry experience is
mockups for coolant pump seal replacement are
invaluable. Did they use mockups in their -- did
they have a mockup seal?
MS. GOOD: Some but not enough. That
was one of the areas that was a performance issue
was the lack of the use of mockups.
MEMBER UHRIG: On an object search and
retrieval is not a normal part. That's sort of an
accident? Did somebody drop something?
MR. HOWELL: Yes. Right.
MEMBER UHRIG: So this is just simply
the fact that it went over five rem, because
normally that would be zero.
MS. GOOD: Well, they planned to do this
job and they said, We think it's going to take this
much dose to do this work --
MEMBER UHRIG: Right.
MS. GOOD: -- and they went over that by
greater than 50 percent, so it was work that they
planned to do.
Getting into the performance problems,
the licensee conducted post job reviews and had
prepared an outage report, and the licensee actually
identified five performance problems that caused the
higher than predicted doses. And those problems
were the maintenance activities were conducted in
the vicinity of the reactor coolant system during a
time soon after shutdown when area dose rates were
temporarily elevated by a chemical cleaning process
and without taking any additional protective
measures for personnel.
The second performance problem --
maintenance activities were conducted in the
vicinity of the steam generators before the steam
generator bowl drains were flushed resulting in
higher than normal dose rates, and again, without
taking any additional protective measures for
personnel. Third, the maintenance activities were
conducted on the reactor coolant pumps and the steam
generators without the secondary sides filled with
water resulting in higher than normal dose rates,
again, without taking additional protective
measures.
The fourth performance problem was that
maintenance activities were conducted without
sufficient practice training to familiarize
worker -- contract workers with plant equipment, the
use of tools, and techniques to effectively reduce
the dose that they would receive. And then the last
performance problem, maintenance activities were
performed with ineffective communications between
radiation protection personnel and the primary
contractor, which resulted in additional worker
exposure due to ineffective planning and the
sequencing of work activities.
Now, in addition to these performance
problems the NRC was aware that high collective dose
was a problem at the plant. The collective doses
had increased between 1997 and 1999 and exceeded the
135 person rem which is the industry median for
pressurized water reactors. They were -- at the
time we did this they were at about 178 person rem,
and there was only one other PWR that had a greater
person rem, and that was Indian Point 2.
MEMBER LEITCH: Were there any concerns
with individual exposures?
MS. GOOD: No. There were no
overexposures.
MEMBER LEITCH: Do you know if any of
the licensee's administrative limits were violated
for individual exposures?
MS. GOOD: I don't believe they were.
MEMBER LEITCH: Okay. Thanks.
MEMBER APOSTOLAKIS: What exactly is
ineffective communication? What does that mean?
MS. GOOD: They didn't -- some
individuals, some groups didn't know when other
groups were planning to do work. They didn't have
good briefing so they weren't able to plan things
out so it could be done in the most efficient way to
reduce the doses. So it just -- not confusion, but
it took more time for them to figure out what was
going to happen next.
CHAIRMAN SIEBER: This plant's been
running since the 1980s?
VOICE: Yes.
CHAIRMAN SIEBER: So it's not lack of
experience.
MR. HOWELL: She's going to touch on
that. They did a root cause analysis.
CHAIRMAN SIEBER: All right.
MS. GOOD: After conducting a regulatory
conference with the licensee in November of 2000,
reviewing the supplemental information that the
licensee provided and conducting a series of
significance and enforcement review panels -- and
those included regional personnel, NRR, Office of
Enforcement, the Office of General Counsel, and the
inspection program branch the region then issued its
final significance determination and violation, and
we issued that in January of 2001.
Now, our letter indicated that we had
identified three white findings, and in the reactor
oversight process those are findings with low to
moderate safety significance. Now, the two jobs
that accrued greater than 25 percent rem were
determined to be individual white findings using the
occupational radiation safety significance
determination process. And again, those were the
scaffolding jobs and the eddy current and
electrosleeving that I discussed earlier.
Now, the other jobs -- and I won't go
over that list again -- were all grouped together to
make the third white finding, and the significance
determination process assigns a white significance
if there are greater than two jobs that exceed the
five person rem and the greater than 50 percent dose
projection. You get over 25 person rem it's a stand
alone finding.
MEMBER UHRIG: Had they not mis-
estimated the exposure here, just the fact that it
was greater than 25 person rem would have been
sufficient to get the white rating?
MS. GOOD: No.
MEMBER UHRIG: It would not?
MS. GOOD: They would have had to have
exceeded the projection --
MEMBER UHRIG: By 50 percent?
MS. GOOD: By 50 percent. That's right.
MEMBER UHRIG: Okay.
MS. GOOD: And then lastly we issued a
violation for failure to use to the extent practical
procedures and engineering controls based on sound
radiation protection principles to achieve
occupational doses and doses to members of the
public that are ALARA, and I've got the citation
list in there.
MR. GWYNN: This was a precedent-setting
notice of violation for a power reactor. There had
only been, to my knowledge, one other before that,
and it was a 4 that was not reviewed by the program
office before it was issued in Region II, so this
was a precedent-setting notice of violation, and I
believe it was at the right plant at the right time.
CHAIRMAN SIEBER: That's a lot.
MS. GOOD: Yes, it is.
CHAIRMAN SIEBER: -- for a plant of that
size and age.
MR. GWYNN: Right.
MS. GOOD: So in response to our January
2001 letter, the licensee submitted two separate
appeals that covered four areas. I have the first
one here. First they asserted that the NRC had
imposed a regulatory staff position that is new or
different from a previously applicable staff
position; in other words, a backfit. Second, they
denied the violation; third, they asserted that our
significance determination process creates a new
regulatory burden and that it's fatally flawed and
should be suspended; and finally, they appealed the
staff's determination of the three white findings.
But other than that they were really happy with the
letter.
MEMBER POWERS: This is the classic my
dog didn't bite you, my dog doesn't bite, I don't
even own a dog approach.
MS. GOOD: So after a great deal of
careful review by a significance determination
appeal panel, a backfit panel, and evaluation of
each of the licensee's arguments -- and this again
was a small army of people that again included the
region, somebody from another region as an
independent evaluator on the appeal panel, members
from NRR, OGC, OE, Inspection Program Branch, the
NRC issued a response to the licensee's appeals in
May of 2001.
CHAIRMAN SIEBER: This seems to be a
licensee's response and your response seemed to
involve legal issues. I presume they had their
attorney and you had yours?
MS. GOOD: Yes.
MR. GWYNN: Like I said, it was a
precedent-setting enforcement action.
CHAIRMAN SIEBER: Did anybody
participate besides NRC and the licensee, like NEI?
MR. GWYNN: No, but there was --
MR. BROCKMAN: On the stage, no. Behind
the scenes, yes.
MR. GWYNN: And there were interested
members of the public from the State of Missouri
who --
CHAIRMAN SIEBER: Very interesting.
MS. GOOD: So we issued our response to
their appeals and our response said we determined
that there was no backfit, and that applied to both
the significance determination and the violation.
We determined that the violation occurred as
described in our notice of violation, and that the
occupational radiation safety significance
determination process is fundamentally sound even
though there are some areas that could be enhanced,
and currently the NRC is working with NEI to work
through those specific issues.
And then lastly the significance
determination process appeal panel concluded that
there were no significant discrepancies in how the
staff had applied the significance determination
process, so in accordance with the reactor oversight
program -- and the region conducted a supplemental
inspection, and we did that to provide assurance
that the root causes and the contributing causes are
understood for the performance issues to
independently assess whether the root causes for the
performance issues affected other plant processes or
human performance, otherwise known as extent of
condition; and three, to provide assurance that the
corrective actions for the performance issues are
sufficient to address the root and contributing
causes and to prevent recurrence of the performance
issues.
CHAIRMAN SIEBER: The licensee did not
go to or consider the appeal board?
MS. GOOD: We understood that formally
the only appeal that existed at that point was to
appeal the backfit, and we've not heard whether they
intend to do that or not, and certainly there are
some informal processes that they could use. And
we've not gotten an indication at this point that
they plan to appeal anything. We've gotten a sense
that they may just let the NEI and the agency work
through the issues with an occupational rep safety
SDP.
CHAIRMAN SIEBER: Thanks.
MEMBER UHRIG: What if the next time
they came in and estimated these at a hundred person
rem and you said that's unreasonable, what's going
to happen then?
MS. GOOD: I don't know. We've not gone
down that path. Obviously we're going to be looking
carefully, and we will in fact be looking because as
Art mentioned this most recent outage that they had
their total dose for the outage was 100 person rem.
Well, they had estimated -- if you added up the sum
of all their radiation work permits it came out to
160 person rem. We haven't done an inspection yet
to really discover why there is this big difference,
so we would just have to look and see if they have
good reason. If they don't have a good reason we're
going to have to pursue it and see where we end up.
MEMBER UHRIG: Are you going to adjust
their estimates?
MS. GOOD: Are we going to adjust their
estimates? No. We would just ask them why they
adjusted their own estimates and what was their
justification for doing it. So we would -- then
it's going to be our opinion against theirs if we
don't agree with their justification.
MR. HOWELL: We've seen one or two
examples of inflated dose estimates we believe.
It's -- they're more modest in nature. They're not
100 rem. They're --
MEMBER UHRIG: I think the one you
alluded to was at Turkey Point in the steam
generator change out. I remember it involved that
one.
MS. GOOD: We've seen a couple of other
instances at plants in Region IV since this action
occurred where we at least had some questions, but
at this point we felt that everybody has had a good
answer when we've asked those types of questions, so
far, the plants in our region.
MR. LARKINS: Let me ask you a quick
question. You said that NEI and NRR are working
together to work out some of the nuances in the
significance determination process for this area.
Did the region find any -- take any issues with the
SDP process as currently constituted for handling
this type of problem?
MS. GOOD: I think initially we felt it
ought to be just one white finding because these
were based on activities that occurred in one
outage, and we questioned whether they were really
the same problem rather than multiple different
problems. But I think what we arrived at was --
what we have here is really a programmatic breakdown
in the area of ALARA. Everything in the ALARA
program was broken, and so from that standpoint I
think we did feel that three whites and the actions
we would take based on three whites was really the
appropriate thing for the region to do.
MR. BROCKMAN: That's the key issue when
you go to the action matrix was the actions that
were responsive to this particular problem with
ALARA that we would send out a couple of person team
inspection to follow up on this with their root
cause analysis, and I think we thought that was
right on where we should be. If it's less I've got
one person out there for two or three days or Art's
got one person out there for two or three days which
wasn't the right type of response to be able to
address the issue.
So you've really got to look at the
action matrix and where it puts you.
MS. GOOD: I'll go on then and go into
the root causes. They identified several root
causes. First they identified that it was
management's failure to establish expectations for
keeping doses ALARA, management's failure to
communicate a priority for keeping doses ALARA, a
culture that did not support an ALARA concept, and
then finally administrative controls that didn't
assure that documented ALARA concerns would receive
proper priority, appropriate consideration, and
comprehensive resolution.
MEMBER APOSTOLAKIS: How did they decide
that the culture did not support the ALARA concept?
I thought we can't say anything about culture.
MR. HOWELL: Those are their words.
VOICE: That's their finding.
MEMBER APOSTOLAKIS: Their as the
licensee?
MS. GOOD: Yes.
VOICE: That's not ours.
MS. GOOD: This was what came out as
their root cause analysis.
MEMBER APOSTOLAKIS: So our guys doing a
supplemental inspection --
MS. GOOD: Yes. We looked at their root
causes and their extent of condition to determine
whether we agreed with them and whether they took
appropriate corrective actions to address those root
causes.
MR. BROCKMAN: We can't say it but we
can endorse them saying it.
MS. GOOD: So after conducting our
inspection, looking at what they provided to us on
their root causes and their corrective action, we
concluded that the licensee had conducted a thorough
evaluation of the causes and had correctly
identified the extent of condition and had
implemented appropriate corrective actions.
We found that some corrective actions
were not completed before they started the most
recent outage. They were actually in an outage when
we did our supplemental inspection, so that was good
timing for us, and that some corrective actions had
not been institutionalized, and by that I mean that
they hadn't been incorporated into procedures and
processes to ensure that the lessons learned would
be lasting.
So that ends my presentation on ALARA.
I don't know if there are any further questions for
me.
VOICE: I think you've heard from the
presentation that we learned something with respect
to the initial implementation of the oversight
process through this, that it can become very
burdensome in terms of staff hours to address these
controversial issues where the licensee took issue
with virtually everything that we found but
subsequently agreed they had a major problem that
needed to be fixed.
CHAIRMAN SIEBER: I would think -- and
I'm not speaking on behalf of the agency but more on
my experience as a licensee that if you had dose
rates like that to your people you would be
concerned right off the bat. You would be concerned
before you --
MR. HOWELL: That was our sense too,
that clearly some of these things that they did not
do during that outage were lessons they had already
learned because of those dose rates and they chose
for various reasons not to implement --
CHAIRMAN SIEBER: Well, the industry has
moved way beyond this point. This is 15-20 years
ago behaviors.
MR. HOWELL: Yes.
MR. GWYNN: And I think that's why it
was easy for them to come to the conclusion about
the safety culture because there were such glaring
examples where the culture should not have allowed
the activities to progress to where they did.
CHAIRMAN SIEBER: Do you believe that
cultural issues as would reflect itself in one
technical area spread to other areas in the plant?
MS. GOOD: That's part of what we looked
at, what they had to do when they looked at the
extent of condition, and the only area where they
felt that there was some involvement in other areas
was administrative controls. The issue having to do
with the administrative controls did apply to other
areas and not just the ALARA area.
CHAIRMAN SIEBER: And did you all agree
with that conclusion of the licensee?
MS. GOOD: Yes, we did.
MS. SCHOENFELD: Did a contractor do
their assessment or did they do it?
MS. GOOD: Do you mean provide their
response to us?
MS. SCHOENFELD: No. Do their root
cause --
MS. GOOD: I don't know the answer to
that.
MS. SCHOENFELD: -- to identify these
root cause findings.
MS. GOOD: I don't know if they used a
contractor to do that. I can get that answer for
you but I don't believe that they did, but I'd like
to check on it.
MEMBER POWERS: The thing that interests
me is the decision to make it three findings instead
of one in order to get it into what you felt was the
appropriate place in the action matrix.
MR. HOWELL: We applied the SDP as --
literally as it was developed, and that's the
outcomes, three white findings. It's clear. To go
to anything else would have been a manipulation of
the SDP. Now, you can argue about whether that's
right and certainly they did, but we implemented it
and that's what you get. You get separate findings
for each of those categories.
CHAIRMAN SIEBER: Any further questions?
MEMBER APOSTOLAKIS: Yes. Why are you
the only one using Power Point?
MS. GOOD: I think there's going to be
someone else this afternoon.
CHAIRMAN SIEBER: Well, thank you very
much. That was a very good presentation.
MS. GOOD: Thank you.
And with that I'd like to introduce Troy
Pruett. Troy is going to cover the Callaway grid
experience.
MR. PRUETT: Once again, my name's Troy
Pruett. I'm a senior reactor analyst in Region IV.
Today I plan to discuss an overview of the Callaway
plant trip that occurred in August of 1999, and at
the tail end of that I'll go through a review we did
of electrical related SCRAMs and ESF actuations
occurring in Region IV since 1995.
Kriss didn't get a chance to mention
some of the functions that the SRAs performed, but
one of the things we do is an independent review of
operational events as they occur and then again when
the LERs make it into the region. And during one of
these independent reviews a senior reactor analyst
had identified a potential concern involving
switchyard voltages being below the tech spec
requirements following a reactor trip at the
Callaway plant.
Based on that concern the NRC initiated
an inspection activity which involved the senior
reactor analyst that initially identified the issue
as well as a resident inspector from the Diablo
Canyon plant.
There were three general issues of
concern that the inspectors took with them. One was
a plant trip that results in a loss of offsite power
condition. A second concern was a plant trip which
results in a potential for double sequencing of
safety-related equipment, and then the third concern
would be a plant trip that would result in a partial
actuation of safety-related equipment.
And then we're also going to talk about
specific areas of concern that were identified as a
result of the inspection that involved operator and
dispatch center awareness of the degraded voltage
condition, and I'll get into the specific inspection
issues.
The first one --
CHAIRMAN SIEBER: Let me ask a couple of
general questions first.
MR. PRUETT: Okay.
CHAIRMAN SIEBER: The inspection report
talked about high inner system loads. Was that
reactive or real power delivery? You can get a lot
of current going and no power going.
MR. BROCKMAN: You were running in a
large demand in the Chicago area --
MR. PRUETT: No. There was a load
demand in the north because of cold weather, very
high demand in the south, and very high demand in
the grid area that the power is being wheeled
through.
CHAIRMAN SIEBER: Okay.
MR. PRUETT: Callaway's function during
this time frame was to provide grid support in the
form of reactive loading. They had boosted the VAR
output of the generator.
CHAIRMAN SIEBER: So they were pumping
VARs as opposed to delivering energy.
MR. BROCKMAN: But the grid itself --
and it was a freight train with power going through
it.
CHAIRMAN SIEBER: Now, did they have
automatic tap changers or manual tap changers?
MR. PRUETT: At the time they just had
the standard transformers. After the event --
CHAIRMAN SIEBER: No tap changers?
MR. PRUETT: No. After the event they
installed automatic load tap change transformers.
CHAIRMAN SIEBER: And they are the kind
that will change taps under load?
MR. PRUETT: That's correct.
CHAIRMAN SIEBER: Because there's two
different kinds, one of which does you no good.
Thank you. That helps me to understand
a little better.
MR. PRUETT: I was going to explain the
phenomena with that slide right there. I don't have
to address that now other than to say that as a
result of the event and the inspection findings the
plant revised procedures to limit the amount of VAR
output that they could put out through their grid.
The next slide -- the licensee's
procedures for verifying offsite power did not
account for post trip voltages or instrument
uncertainties. In this case the dispatch center
uses a post-contingency computer model to determine
what the grid condition would be for several
hypothesized transmission failures. In the event
that the computer model detects a potential low
voltage condition it's supposed to activate an
alarm. They in turn were supposed to contact the
Callaway plant and inform them of that alarm
condition.
CHAIRMAN SIEBER: And this occurs before
any actuations of protected devices occur on the
system?
MR. PRUETT: This is all hypothetical.
CHAIRMAN SIEBER: Right. This is in
advance.
MR. PRUETT: In advance.
CHAIRMAN SIEBER: So this is a real load
flow calculation?
MR. PRUETT: Right. That's correct.
Some of the deficiencies involved in
that communication process and some of the computer
alarm setpoints -- the inspectors identified that
the computer point alarm setpoints were non-
conservative, and that was both at the plant end and
at the dispatch center end. On the plant end the
maintenance personnel incorrectly set the alarm
setpoint on the plant computer associated with grid
voltage.
Even had they set it correctly the
setpoint was non-conservative in that it did not
account for instrument uncertainties associated with
monitoring switchyard voltage.
On the dispatch center end they didn't
have an appreciation for what the tech spec allowed
value was for voltage, and consequently their
predictor model alarm setpoint was set too high, so
even though an actual low voltage condition existed
their predictor model did not detect it, provide the
appropriate alarm, and consequently the plant wasn't
notified.
CHAIRMAN SIEBER: Now, this issue went
back to the early 1980s industry wide?
MR. PRUETT: As far as --
CHAIRMAN SIEBER: Low voltage --
MR. PRUETT: Low voltage condition?
CHAIRMAN SIEBER: Low voltage and low
flows. It goes back a long ways.
MR. PRUETT: Long ways.
MR. BROCKMAN: But it's really raised
its head back up now when you're looking at all the
implications with the wheeling that's being done and
the artificial support.
CHAIRMAN SIEBER: That's my question.
There is an information that was published about
this one and four or five incidents like this
over -- from '97 to '99 I think it was. What is
going on now in Region IV since this condition is
getting worse day by day as the energy situation
does not improve would be a good way to say it that
would make other plants vulnerable to the same kinds
of things? Has somebody gone in and said, Do you
have tap changers? Have you had a load flow -- a
recent load flow that tells you what these settings
are? What happens if -- what happens otherwise?
MR. PRUETT: You mentioned one
information notice that came out directly after this
event. There is also a regulatory information
summary that came out --
CHAIRMAN SIEBER: Right.
MR. PRUETT: -- and in that summary,
essentially it acknowledges that NEI committed to
communicate these grid reliability concerns to the
industry. Out of that NPO is also conducting a
review of grid reliability concerns that's supposed
to be completed in 2002, and my recent discussions
with the NRR folks indicates that the NRC may
initiate a review of those implementations to
resolve grid reliability concerns following the NPO
review depending on the findings that come out of
that.
CHAIRMAN SIEBER: So the answer is no?
MR. PRUETT: Well, that's long term
plans. In the near term several utilities reacted
in response to those information notices and
improved their communications with their dispatch
centers. Most of the utilities in Region IV have
agreements with the dispatch centers and those
dispatch centers use a post-contingency type of
model to predict grid voltage conditions for those
plants.
There's only -- I think there's only one
that I came across, Cooper, that does not have a
post-contingency model for a predictor. All of the
other sites that I've talked to did have such an
agreement and model in place. On top of that,
specifically for the Entergy plants, since I'm most
familiar with that, they will have no touch days. I
know the west coast plants have no touch days based
on loading on the grid, and at that point that's
communicated through the plant status aspects of the
inspections that the residents do, and the residents
follow up on the onsite contingency plans associated
with those.
MR. HOWELL: And as you indicated, this
is not a new issue, and we've had -- dealt with
similar problems at some of our specific facilities
in the past, Arkansas.
MR. BROCKMAN: But to bring it up, we've
got several less formal channels that have been
used. We have many regional utility group --
engineering, licensing managers, plant managers and
what have you that typically all the executive
management whenever they have a meeting participates
in. This has been a topic of continual drum beating
by us, and these forms bring it to their attention
and to drive on there. It's become an area of focus
for my resident inspectors out there especially with
respect to VAR loading, whereas they're doing plant
status -- we really pay a new type of attention to
that now as opposed to in the past.
We don't just look at the spider graph.
If you get 200 or above we start inquiring as to
what's going on, because it's getting in the realm
of a concern, could start coming up with artificial
holding out.
CHAIRMAN SIEBER: Let me just ask one
more simple question, and hopefully not engender a
complex answer. But I would be concerned about the
west coast area network and the rolling blackouts
and whether or not there have been load flows
performed for calculations prior to deciding what
they're going to black out and when, because that
really changes the flow in the grid, changes the
amount of VARs that get pumped around, changes the
voltages at the substations of all these stations,
and you can figure this out in advance. Has anybody
done that?
MR. BROCKMAN: There's been a lot done
on that area. We've got one or two down where we're
going to talk about California and I think we can
get into that quite a bit.
CHAIRMAN SIEBER: Thank you.
MR. PRUETT: I just wanted to touch on
some of the corrective actions that were taken.
MR. GWYNN: Just as a matter of going
back to our focus on the initial implementation of
the reactor oversight program, in the past with this
type of a learning experience in Region IV we very
well may have initiated a regional initiative
inspection where we would go to all 14 sites in
Region IV and look at this, but we have not done
that. The agency is determining what the agency is
going to do across the entire industry, and so
that's why we don't have substantive inspection
activities that we can say we've gone out and looked
at this at every plant in the region. We are
waiting for program office to make decisions about
those types of inspections.
CHAIRMAN SIEBER: As an administrative
process do you consider that to be timely and
responsive to an evolving situation, or would you
feel more comfortable just going and doing it
yourself if you had the resources to do it?
MEMBER POWERS: The other thing I worry
about is if it's a western problem and headquarters
weights it with --
CHAIRMAN SIEBER: Eastern problems.
MEMBER POWERS: -- eastern problems
maybe it doesn't come out with the weighting that it
deserves in this region.
MR. BROCKMAN: A little bit of a dilemma
that you get into here is the licensees have
certainly been put on notice that they have to have
the appropriate management controls and technical
controls in place to ensure they're in compliance
with their license, and that's what basically --
what they've got to do is have a reliable grid to
operate under. We feel very comfortable we've
communicated that to them.
Now, with the new program I have no
reason at the moment to follow up in that area when
they have all assured to me that is going to happen.
We are monitoring some of the indicators. We think
that if they start violating for example VAR loading
and what have you that we would follow up on that.
We'll be able to share with you in
California's case we're doing a little more. We've
taken some additional steps on looking and
challenging and staying interactive on there because
of its exceptional vulnerability and the high public
interest. We're really back into the discussion we
had before, and one of the points that got brought
up in the IIEP corrective action is the
differentiation now between having a responsive
inspection program versus a predictive inspection
program, and what you would be suggesting here would
certainly be predictive type of inspection.
CHAIRMAN SIEBER: Anticipatory.
MR. BROCKMAN: Anticipatory, yes. A
better word than predictive, but I think we're
sharing the same vision, and the new program doesn't
put us into that arena.
CHAIRMAN SIEBER: I guess I look at some
of these things a little differently too. If you
say here's your tech specs and here's all the
setpoints and here's all your procedures and so
forth, and you, Mr. Licensee, are responsible for
maintaining this plan inside that envelope that's
one thing. On the other hand all these other things
are happening from the outside in, and the licensee
may not have control over it. System operator now
is running stuff as opposed to individual
dispatchers, and --
MR. HOWELL: And they were in full
compliance with their tech specs.
CHAIRMAN SIEBER: Absolutely.
MR. HOWELL: They passed all the
surveillances for offsite power availability.
CHAIRMAN SIEBER: Yes, sir.
Well, those are my concerns.
MR. HOWELL: We understand.
MR. PRUETT: I'm going to move on to the
fourth item, which was the plant operators did not
detect the low voltage condition following the SCRAM
and additionally, the plant operators were not aware
of the operability requirements associated with
offsite power.
Now, once they identified the moisture
intrusion issue at the power supply for the alarm
set point they dispatched maintenance personnel to
correct that. The next day they again had low
voltage conditions in the switchyard, picked up the
alarm set point on the plant computer, but the plant
operators didn't recognize that that alarm had
activated and consequently didn't take any actions.
Secondly, when interviewed the plant
operators indicated that even if the dispatch center
called and said the predictor model showed voltages
would be below their minimum requirements following
a plant trip they would not consider the offsite
power source inoperable, and licensee management
revised procedures and instituted some guidance to
have the operators consider offsite power
inoperable. The predictor model showed voltages
would be insufficient.
The next slide the inspectors identified
if there was no agreement related to switchyard
voltage between the Callaway plant and the energy
supply operations personnel. Following the
inspection the licensee implemented an agreement
between themselves and the transmission provider,
and procedures were revised on both ends to notify
Callaway of changes in grid system characteristics
and to notify the plant 15 minutes before an
anticipated out of range condition.
MEMBER POWERS: When you look at this
the immediate question is is this the only area
where they needed to have an agreement between
themselves and their electrical supply center. Is
it the only topic where they didn't have an
agreement they needed one, or are there other areas?
MR. PRUETT: Between them and the
dispatch center?
MEMBER POWERS: Right.
MR. PRUETT: I don't know the full
details of what that agreement involved.
MEMBER POWERS: It may be the only one.
MR. BROCKMAN: I'm hard pressed to think
of another area.
MEMBER POWERS: Nothing came to mind.
MR. BROCKMAN: -- grids going to be
jeopardized we've got an agreement. I'm trying to
figure out a different area that the load dispatch
center and the plant would be involved with. I know
we had a major thunderstorm three or four months
later after this and was on a Sunday afternoon, and
various parts of the grid came crashing down over
there, and this was weather induced, and we didn't
have the problem at that stage of the game.
Some of the interim corrective actions
they had implemented seemed to work on that weekend.
MR. HOWELL: I know that the NRR does
have grid reliability coordinators, and they have
been making visits to these operators and
understanding the agreements and interfaces, and
that's the only thing that's come out of it so far.
MEMBER POWERS: When you read the
writeup on this -- that's the first question that
emerges in this discussion. Is this the beginning
and the end of it or is there something else, and we
just have to wait for another incident to come along
to discover that something else.
MR. BROCKMAN: We were left with no
incident --
MEMBER POWERS: Nothing comes to my mind
either.
MR. PRUETT: Kriss has the next slide
up. The load flow analysis underestimated the
system loading conditions. Specifically the load
flow analysis was modeled on peak winter loading
with an additional 5 percent conservatism. In
actuality the peak load conditions of the Callaway
plant occurred in the summer of '99 and 2000.
The corrective actions that came out of
that were to update the load flow analysis
following -- prior to each peak season, and also to
include the sensitivity due to system transfers
through their grid system.
CHAIRMAN SIEBER: There are some systems
that have on time, real time line loss and load flow
programs to manage the system, and I don't know if
you have any of those in your region but that
capability is there to some extent, and that really
helps.
MR. PRUETT: Something else Callaway
did, we mentioned that they installed the automatic
load tap changing transformers following this event.
They also installed capacitor banks to support a
block start if needed.
The next side there -- the information
notices weren't dispositioned in accordance with
licensee procedures. Specifically there was a
IN9807 offsite power reliability challenges from
industry deregulation was reviewed by the facility
and closed with no further action required. That
prompted them to review all the information notices
issued since 1996, and there were additional
corrective actions that came out of INs that weren't
appropriately dispositioned.
CHAIRMAN SIEBER: Going back to the
capacitor banks, these are switchable banks? Switch
them in, switch them out.
MR. PRUETT: I don't have the full
knowledge on that.
CHAIRMAN SIEBER: The other question is
are they onsite or are they someplace else?
MR. PRUETT: No. They're onsite.
MEMBER UHRIG: They're probably
switchable but they're onsite.
CHAIRMAN SIEBER: I imagine because it's
either that or change to the field --
MEMBER UHRIG: Yes.
CHAIRMAN SIEBER: -- and you can't do
that without getting into instability sometimes.
MEMBER LEITCH: Concerning other areas
of potential interface that may be required with the
dispatcher as was Dr. Powers question, I have run
into some situations where the dispatcher has
certain understandings as far as off normal
frequency operations, that is when you trip the unit
and so forth in 61 cycles or 59 cycles, and some of
those are not necessarily consistent with the best
practices. Some manufacturers of large turbines
recommend against operating at power other than 60
cycles right as is normally done because those --
particularly those large last-stage buckets are so
carefully tuned that the operation at other than 60
cycles at full power may cause the blades to fail
and there could be nuclear safety implications
associated with turbine missiles and that type of
thing.
So I guess that's a little bit of a
stretch, but it might just be an interesting area to
consider; that is, what is the relationship between
frequency -- that is are these large nuclear units
allowed by the practices with the dispatch office,
are they allowed to operate for extended period of
times at other than 60 cycles.
MR. PRUETT: Okay.
MS. WESTON: I have a question. That
information notice issued in March of 2000 indicates
that there was a similar problem in '89 and '01, in
'91 Millstone, in '93 Palo Verde, in '95 Diablo
Canyon in '95. What was being done in the interim
to deal with this problem since obviously there were
a number of related issues? Was there anything done
prior to now with regard to this issue?
MR. PRUETT: That predates my tenure in
the --
VOICE: It's a binary answer.
VOICE: Done from what perspective? The
licensee's perspective, or the NRC's?
MS. WESTON: The NRC.
VOICE: You know, I know at ANO, ANO
addressed their specific issues, but --
MS. WESTON: NRC.
VOICE: Right. In the early '90s. Yes.
There's no direct inspection of this area. There
was previous information notices as the issues were
emerging.
In the case of Callaway the previous
occurrence was not known until the investigation was
conducted for the '99 events, so in every case it
wasn't known necessarily that those occurrences
occurred at that time.
CHAIRMAN SIEBER: My memory isn't too
great, but back in the 1980s I seem to recall a
round of questions coming out on this subject which
we did line losses and load flows and ended up
putting in tap changers and changing bus
configurations and especially if you change out a --
I mean, a transformer and the impedance of the new
one is a little different than the old one you end
up with a whole host of different problem, because
you may end up with surges too big that your circuit
breakers will hold together when they trip, you
know, and you could end up blowing out the breaker
here and there.
So there's a lot goes into these
calculations, and we modified the plant in the 1980s
for that issue.
MR. BROCKMAN: But the process would
have been with the TI or incorporating something
into the old core inspection program --
CHAIRMAN SIEBER: Right.
MR. BROCKMAN: -- during that time
frame, and it wasn't either.
CHAIRMAN SIEBER: It was NRC initiated
that.
MR. HOWELL: But there were -- as you
may recall the electrical distribution system
functional inspections made.
CHAIRMAN SIEBER: Right.
MR. PRUETT: Lastly, there were some
generic communications that were issued following
the Callaway event, and we touched on those already
as well as NEI's involvement with NPO and NRR.
That's all I was going to talk about as
far as the Callaway event is concerned, unless
there's other questions.
CHAIRMAN SIEBER: I think if would be
good if we could move on to California grid.
MR. GWYNN: We're far enough behind
schedule that I'd like ask if you'd be agreeable to
our just eliminating the discussion of the
electrical design operations issues at Cooper from
the planned agenda.
CHAIRMAN SIEBER: I think we could.
MR. BROCKMAN: Or let's just put it at
the very end. If we want to -- recovering all the
time at the end of the day we'll come back to it,
which I'm doubtful of, but --
CHAIRMAN SIEBER: Another area that we
may be able to cut back on to some extent is the San
Onofre electrical fire because we have a lot of
materials and pictures of that, along with --
MEMBER LEITCH: We have a very short
presentation if you have a very few questions. One
thing I'd like to hear if someone's up to date on it
is apparently I believe there was a fire at Cooper
earlier this week, and I'd like to be briefed very
quickly on the events there. Evidently they're a
single loop operation. I'm not sure if that was
related to the fire.
MR. GWYNN: Both San Onofre and Cooper
on the same day experienced potential transformer
explosions that affected the plants. The Cooper
effect was much greater than what it was at San
Onofre, but essentially the same event.
MR. BROCKMAN: And David is acting as
the branch chief for the branch that owns both of
those plants, and he was also the regional duty
officer for those two nights.
VOICE: Why don't we go with California
and then come back and touch on that briefly?
MR. LOVELESS: As Ken told you, I'm
David Loveless, currently acting as the branch chief
with responsibility for Cooper, Fort Calhoun, and
San Onofre, and I'm here to talk about some of the
things that San Onofre and Diablo Canyon were seeing
in California. I guess if you've read any
newspapers or watched the news you know that the
people in California aren't really happy right now
and that they have a lot of problems going on with
their electric deregulation.
While we empathize with those
individuals we basically are concerned that the
plants continue to operate in a safe manner and that
the financial conditions of the utilities aren't
affecting that safe operations, and that's what I
wanted to talk about today. A couple of things that
I will cover here is a little bit of history of the
electric grid in California and how it got to where
they are, what the current situation is under
deregulation, what our response is to our concerns
at the plant based on that condition, and then I'll
provide a brief summary of where I think we are.
So California has for some time been an
importer of power. The last numbers that I heard
ranged on the 20 percent range on the average, so
they import a lot of power. They don't have the
resources to produce all the power they're using.
Another of the problems that set them up
for this was the BANANA principle, which is build
absolutely nothing anywhere near anybody, and that's
been in existence for about ten years at least in
California. They have legal restrictions to
building their plants. They don't want them in
their back yard. They don't want any that burn coal
or gas because they have emissions. They don't want
nuclear because they're scared of it, and so the
bottom line is they just haven't been building new
plants.
But at the same time they've had
significant electric power growth. No one even
began to understand how much power the internet and
Silicon Valley was going to take, but it's using a
lot.
The current situation -- this number I
got from an Enron report that they've been
collecting an average -- and this is averaged over
night when it's real cheap and the peaks when it's
real high -- $138 a megawatt wholesale power rates.
Well, our two major utilities, Pacific Gas and
Electric and Southern California Edison, have a
regulatory cap retail of $60 a megawatt hour, so
where does that lead? Well, Pacific Gas and
Electric is in Chapter 11 bankruptcy, and Southern
California Edison is working on a memorandum of
understanding that they have developed with the
State of California that will make substantial
changes to their business, but they are hoping will
bail them out and keep them from going bankrupt.
Now, Pacific Gas and Electric is the
owner-operator of Diablo Canyon, and Southern
California Edison is the primary owner and operator
of San Onofre, so both of these companies are having
significant financial difficulties right now.
What are we doing about it? We
developed a list of what we call financial impact
observables. I'll talk about those in just a
minute, but we have basically a punch list that our
resident inspectors go out on a weekly basis and
keep in the back of their mind as they're doing
their routine job, and they look at these
performance indicators with a goal of determining
early on that the financial situation of the
utilities is affecting safe operations.
We have senior management visit both the
sites once a month. We've been doing that since
January, and we talk with senior people out there.
We have retained shorter inspection report periods
and we're including more details in the scope of
those inspections in our inspection reports to
assist the public and other interested members in
understanding what we're doing out there to ensure
that plants remain safe throughout this evolution.
We've had additional public meetings
specifically tied to bankruptcy and the financial
conditions, plus we've taken every opportunity we
could to have meetings where the public was
available and have press conferences associated with
some of those meetings. Senior management in the
region has meetings weekly. We have a call with
both of the senior resident inspectors and discuss
these observables. I'm going to talk about other
impacts: morale, what the public's doing, all kinds
of things.
We also have biweekly meetings with
utility managers. They've supported those to the
point that we get senior managers in the corporate
office and senior managers at the plant on the phone
that specifically discuss their financial condition,
things that they wouldn't be willing to discuss
under any other circumstances, or not publicly, so
we're getting as much information as we can to help
us understand what they're doing.
The items that we've asked the resident
staff to take a look at -- one of them is staffing.
We are looking at just their basic level of staffing
at the plant with the assumption that one of two
things might happen that could affect safety. One
is the better employees start to realize that the
company is going down and look for other jobs and
leave, and they start losing people that way.
Another would be -- excuse me.
MEMBER APOSTOLAKIS: I think it's
happening at Southern California.
MR. LOVELESS: They haven't seen any
increased attrition right now. They are watching
that closely as you might guess, but their staffing
levels are staying fairly steady. They're dropping
a little bit but they're dropping along the line of
a gap review that they did a couple of years ago and
have had in place for quite some time. And so
currently we're not seeing anything there but we are
definitely looking on a routine basis.
We're looking at plant maintenance. We
look at the backlogs in corrective maintenance. Are
they creeping up? If they do go up we look for why.
Is it because suppliers won't provide the parts
because they're afraid of not getting paid? Is it
because they don't have the money to buy things,
that sort of thing. That's the type of thought.
So far, again, we've seen nothing.
We're seeing the normal ups and downs -- the
corrective action process in that maintenance area.
We also looked at the preventive
maintenance to make sure they're continually in full
force, and so far they have been. We're looking at
outage in plant modifications. Are they changing
the scope of any outages? Have they canceled or
postponed any risk significant modifications?
Again, so far we've seen nothing here. In fact, the
unit 3 outage at SONGS we took some additional time
in that outage to do some things that would make
them more reliable in order to stay online
throughout the summer, which is where the demand's
going to be.
We've looked at a number of things in
emergency preparedness. We looked at training, make
sure they're continuing to provide training to
people, that they aren't trying to cut back in that
area. We look at the facilities, make sure they're
ready. We've looked at the emergency sirens, make
sure that as they go into blackouts that they're not
blacking out the emergency sirens so they wouldn't
be available, and we actually found a couple of
sirens that they had in blackout zones, and they've
gone back and blocked those out so that they won't
be -- so they won't lose power during an emergency
if they're needed.
Also as Troy was talking, we're looking
at grid stability. The primary indicator of that is
the VARs plan we've been looking, and they're pretty
much staying at their historical levels. We also --
we looked at the ISO's responsibilities with respect
to their emergencies and the licensees have asked
the ISO to go back and look at grid stability, and
they've actually recently changed their threshold
for entering a stage three and going into blackouts,
because they decided under certain conditions they
wouldn't be as stable as they'd like to be at that
point, so they're doing it earlier than they were
back in January.
MEMBER UHRIG: What authority if any
does the utilities have over the ISOs?
MR. LOVELESS: The utility has no
specific authority. They have agreements
contractually for certain powers to --
MEMBER UHRIG: When they gave up the
grid they had lost all control of it.
MR. LOVELESS: Pretty much, except that
they are part owners in it, but, yes. They don't
operate the grid.
MEMBER UHRIG: But the grid can impact
the plant.
MR. LOVELESS: That's true.
MR. BROCKMAN: And so they have
agreements associated with them for the operability.
Yes.
MR. LOVELESS: Also, the way the ISO
works in California they don't black out
transmission trunks. They tell the utilities what
their share of the blackout is and it's the
utility's responsibility to select the blocks that
they're going to black out and how. So they're
blacked out at a distribution level so that the
transmission and the stability of that grid
throughout remains there.
MR. BROCKMAN: And all of the thresholds
that you're hearing there are premised for the
blackouts and everything to maintain the stability
of the grid, not in response to instabilities. All
of those blackout activities of when you get to a
certain level is to make sure you maintain an
adequate margin, so that's a key philosophical
application to understand.
MR. LOVELESS: And it's also --
actually, you could look at it as a benefit, because
the ISO's responsibility and primary concern is the
maintenance of that grid. That's how they make
their money, the transmission network, staying up
and stay -- where the utilities want to sell power,
and so the ISO being independent can direct
blackouts in times that the utilities might have
tried to push it. I'm not saying they would. I'm
just telling you that having that independence has
some benefits too.
MEMBER UHRIG: You say they changed
their threshold recently? I assume you meant the
trigger a blackout sooner?
MR. LOVELESS: Yes. That's correct.
They changed -- they were at 3 percent. What's the
new -- 5 percent. Five percent's being reserved.
MR. MARSCHALL: Five percent total
reserves.
MR. BROCKMAN: The speaker is Charles
Marschall.
CHAIRMAN SIEBER: Actually, the hardware
and the procedures for doing this came out of the
failure at Big Alice in New York many years ago, and
I think across the nation those procedures and
equipment are in place to block shed distribution
centers as opposed to transmission lines, and so
that's the way you respond to an undergeneration
issue.
MR. BROCKMAN: It also lets them
localize it very much, let's them do this several
small areas for a period of time and be able to
rotate that around and not get a large metropolitan
area covered and what have you.
The interesting part is the need to
coordinate this with the local law enforcement.
Everybody says, Well, why don't you just put out --
if we go to blackouts today here's the areas that
are going to get a blackout, and everybody knows
between 4:00 and 5:00 I don't want to be on an
elevator. And the working agreements with the local
law enforcement authorities are we can't do that.
Every crook in California will be in that
distribution area between 4:00 and 5:00.
And so you get very much a security
aspect that goes along with this where you have to
have those types of considerations that you wouldn't
necessarily think of right off the bat when you're
looking at the philosophy.
MEMBER LEITCH: Does your observation of
staffing levels include not only the utility
personnel but the contract?
MR. LOVELESS: Total station numbers is
what we've been looking at.
MEMBER UHRIG: Does that include ISOs?
MR. MARSCHALL: Charles Marschall again.
But the financial situation doesn't really affect
the ISO. It affects the utilities could have a
shortfall because of the fact that they can't
collect their costs, and so the ISOs are affected
and the chances are that staffing really isn't a
concern for the ISO.
MEMBER UHRIG: Most of the staff came
from the utilities anyhow that had the transmission
lines before. When they created the ISOs they
didn't start out restaffing them from scratch. They
took the people who were there and --
MR. BROCKMAN: But it came from the
large wire part of the utility, not the plant
operation staff.
MEMBER UHRIG: Yes. You're correct on
that.
MR. LOVELESS: So where does that leave
us? The current safety impacts that we've seen at
the plants are none. Both plants report having
large enough cash supplies to continue to operate
the plants in a safe manner, and the fact that long-
term success of these companies depends on those
plants continuing to run safely.
The utilities are working with the
bankruptcy judge and with the state on a memorandum
of understanding for SCE, and we are keeping close
eye on that to make sure that none of the decisions
made at those levels will impact the safe operations
of these two plants.
MR. BROCKMAN: The Department of Justice
in fact has the responsibility to represent us as an
interested party in this and is actively pursuing
that responsibility in all of the proceedings.
MR. LOVELESS: And so we as a region
realize that we need to continue our vigilance, not
relax in California because we have to ensure that
they maintain safety at those plants, and we also
have a very real role in public confidence through
this crisis if you will, so --
MEMBER POWERS: You raised this issue of
the Department of Justice. Without impugning my
legal friends too much their skills in the area of
reactor safety sometimes are less than optimal.
They have an adequate understanding of the financial
requirements to maintain safety to represent
adequately?
MR. BROCKMAN: Yes. I feel pretty good
on this. Larry Chandler is our representative in
OJC interacting with them, and he and I have talked
on numerous occasions, so while DOJ has that
representational authority the communications
channels are very good and all of the right people
even down to us to provide that information back
on -- allow the good information flow. If there's
any question that would come up and we would need to
actively participate in it they would be calling on
the right technical people to go up with them.
We're not caught with the hoitiness
here. This is our job. Stay away.
MEMBER UHRIG: I assume the solicitor
general is the government's lawyer so to speak, and
they're the ones that have to act on this, and I
assume that they're coordinated in that Department
of Justice very carefully?
MR. BROCKMAN: I'm not sure on that. I
think this is all being done in state -- it's all
right there.
Now, one final thing with respect to
David's public confidence issue. It is amazing how
quickly we forget. I was just out there last week
at Diablo Canyon making -- we had a wonderful public
meeting afterwards with placards and chants and all
manner of people there, and there haven't been any
of the blackout applications in several weeks in
California, and everyone was willing again --
everyone was out there, Yes. We'll hang out laundry
out on the lines. Close down Diablo Canyon. And
the intervener organizations were just a couple of
weeks ago quoted in the press as saying this is a
very important part of our energy mix. We're glad
they're here and we need them.
So the public -- a lot of what we're
doing there is really trying to make sure the public
understands that we are being attentive to
monitoring the activities and that right now while
we don't see an actual consequence we're paying
attention to this and if it would start going down a
path we will be very active in it, and that's the
level of confidence we're trying to provide them, is
just it is being monitored. It's being looked at.
CHAIRMAN SIEBER: I would prefer to see
the agency acting in that role as opposed to
responding to events, and so I hope you keep up the
good work.
MR. BROCKMAN: I have the schedule
through December for all the monthly visits out
there if anyone's looking for a trip.
CHAIRMAN SIEBER: Are there any more
questions on this topic?
(No response.)
CHAIRMAN SIEBER: I think since we're
going to try to roll on as far as we can it might be
a good idea to take a ten minute break and come back
at two o'clock.
(Whereupon, a short recess was taken.)
CHAIRMAN SIEBER: At this time we'll
resume the meeting.
MR. GWYNN: I believe that we had a
couple of questions that we needed to respond to.
David, there was a question about the transformer
explosion that occurred at Cooper Nuclear Station.
MR. LOVELESS: Sure.
MR. GWYNN: Did we answer that question?
MR. LOVELESS: I answered it during the
break.
MR. GWYNN: Okay. In that case I'd like
to turn the meeting over to Art Howell, who will
present the fire protection experience in Region IV.
Art.
MR. HOWELL: Good afternoon. Once again
I am Art Howell, the director of the Division of
Reactor Safety. What I'd like to do is share with
you the results and experiences that we've had with
implementing the new fire protection inspection
program, and the last slide in your package is a bar
chart, but before I get there I thought a little
background would be appropriate.
The new baseline inspection program --
I'm on page 2 of the slides -- resulted in a
significant increase in the inspection level of
effort compared to the old program. It's
approximately a tenfold increase. If you look at
the old program it was roughly 25 to 30 hours every
other SOWP cycle, so every three years, and the new
program is one team inspection performed every three
years at about 200 hours of effort plus another 33
hours spread out over four quarters by the resident
inspector, so when you annualize that it comes out
to about a hundred hours, so it's about a tenfold
increase.
MEMBER POWERS: It's a big increase
relative to what was in the past. The question is
is that big enough?
MR. HOWELL: Our experience has
indicated that based on what we know, yes. And so
far as the number of risk significant fire areas
it's fairly limited, and so if your premise is that
you can get to all those in a reasonable period of
time on a sampling basis is the level of effort
enough. And Rebecca, who is one of our team
leaders, Rebecca Neece, has been heavily involved in
this, and I'd say overall from an overall
perspective it is, but we have been challenged
during individual inspections to get everything done
in the inspection procedure.
And in particular what we have found is
that it takes quite a bit of effort to exercise the
fire protection and significance determination
process, and oftentimes that has to be done after we
get back from the site and so I don't think that was
envisioned in the process, but it's recognized, but
we are getting the inspections done. So from that
standpoint, yes, do we need to work on streamlining
for the inspectors the use of the SDP? That is true
too.
CHAIRMAN SIEBER: I guess the question
is is the size and scope of the inspection geared to
the FTEs available to perform, or is it geared to
the actual risk in the plant?
MR. HOWELL: It is risk informed in the
sense that the goal of the inspection is to focus on
the most risk significant fires. During a pilot the
level of effort for this inspection was half of what
it is right now --
CHAIRMAN SIEBER: Right.
MR. HOWELL: -- and it was recognized
that that clearly was not enough to accomplish the
individual inspection objectives associated with the
inspection.
Now, at that time there was two elements
of the inspection during the pilot that we're not
implementing now, and yet we've doubled the level of
effort. And so whereas it's a challenge to perhaps
get everything done and look at the extreme limit of
the sample, which is five fire areas -- it's three
to five fire areas -- we are getting that done. The
impact in on the tail end, not on the front end.
MEMBER POWERS: Well, you're getting it
done, but quite frankly, you're essentially giving
everybody a bye on the associated circuits analysis.
MR. HOWELL: That's correct.
MEMBER POWERS: And that is a non-
trivial inspection.
MR. HOWELL: That's correct. That's
true.
MEMBER POWERS: That would be a big
effort.
MR. HOWELL: Right. And in fact, when
we get to the results our first pilot was Fort
Calhoun, and at that time we were still doing
associated circuits, and Rebecca Neece was the team
leader. She had to go back out to the site and it
took several weeks of SRA involvement to disposition
the inspection findings.
And it was partly because of that
experience that the program office increased the
level of effort when we went into the initial year
of implementation, and quite frankly, the other
regions were experiencing similar outcomes and so
that's why the level of effort was doubled, but it's
still challenging. It is challenging.
CHAIRMAN SIEBER: I would imagine though
that the moratorium on associated circuits
inspections is going to end some time.
MEMBER POWERS: It depends on how long
NEI can string it out.
MR. HOWELL: I know the testing -- that
some testing has been completed and the results are
being reviewed by the expert panels, and we have a
number of open issues in that area that we can't
disposition that we're waiting for guidance, but
you're right.
Also one of the things that we were
supposed to be looking at that we really couldn't do
was reactor coolant pump lube oil collection
systems, and you really can't look at those at
power, and we don't do team inspections during
outages, at least routine team inspections, and
we've had issues in that area in the past that we've
identified. In the case of one plant they actually
had a fire from leaking lube oil that wasn't
collected that soaked some lagging, and because the
wicking had started a fire in the containment, and I
believe Kriss Kennedy responded to that event, and
we've had others.
So we're not looking at that. We're not
looking at associated circuits.
CHAIRMAN SIEBER: The point I was trying
to make was if you look at the overall risk of fire
based on IAPEEEs or level threes it's about equal to
the risk of operating the plant.
MEMBER POWERS: I can find plants,
especially among the population of boiling water
reactors, where fire outstrips the normal operating
events. If you do that split we'll all be fire
protection engineers.
CHAIRMAN SIEBER: I guess the point is
on a real risk basis if you were scheduling based on
risk there would be more effort put in.
MR. HOWELL: Right. But the point I
was -- and I understand that. The point I was
trying to make is that there are only a limited
number of risk significant fire areas, and how often
do you have to look at them before you gain some
confidence in how they're being maintained, how the
engineering features are being controlled, et
cetera.
MEMBER POWERS: It really boils down to
the transient combustible issues as far as frequency
it seems to me.
MR. HOWELL: We've had issues with
transient combustibles, and I believe every time
that we've looked at them using the tools that we
have that they haven't had a significant impact on
the fire loading in the particular fire areas.
We've had a number of them. We've had some under
the old program, and in fact, the old inspection
program was primarily focused in looking at the day
to day operating-maintenance testing transient
combustible implementation of the program, so we had
those issues.
Slide three -- the inspection is broken
down into two areas of responsibility. One is
performed by the resident inspectors on a quarterly
basis, as I indicated. They are looking at the same
types of things that we principally looked at under
the old core program which was performed by the
region based inspectors, so that's really the only
significant difference. And the level of effort is
higher. It's 33 hours a year instead of 25 hours
every three years, and then once a year they observe
a fire drill.
Slide four -- the region based
inspection is more focused on achievement and
maintenance of safe shutdown and everything that
goes with it. I touched on the areas that aren't
inspected, associated circuits being a major
omission until that's straightened out. I already
talked about the comparison to the old program on
slide five, so we skip over that.
Going on to slide seven, results of the
team inspections, we are finding instances of
failure to meet separation requirements,
inadequacies with passive barriers, inadequate
emergency lighting, problems with suppression and
detection not meeting code commitments that they're
committed to, you name it. Everything except -- we
haven't really had many if any findings associated
with manual actions of operators to achieve either
safe shutdown or ultimate shutdown, which is
somewhat surprising given that a number of our
licensees do rely on manual actions, and many of
them are time critical. I would have thought just
as an inspector that that's an area that might be
potentially weak.
MEMBER POWERS: How many plants in your
reviewing have self-induced station blackouts?
MR. HOWELL: I'll have to get back to
you on that.
MR. SINGH: What was your question?
MEMBER POWERS: How many plants in this
region use self-induced station blackout?
MR. HOWELL: She just mentioned Arkansas
Nuclear 1.
MR. SINGH: There's only one that I know
of. Even they abandoned that if I remember.
MS. NEECE: I'm Rebecca Neece. I was
the team leader for the recent Arkansas inspection.
We just got off the site from performing the
Arkansas fire protection inspection, and one of the
areas we looked at had a number of manual actions
they had to take credit for because they decided not
to wrap or protect one train of redundant safe
shutdown equipment.
And in listing the number of items that
could happen all these things that could happen, we
ran across one where they assume a loss of ISET
power but they could also lose DC power which means
that they could also lose service water to the
diesels, which mean they would have to -- the
actions are to trip the diesel.
At the same time if they didn't have DC
power they would be in a station blackout for a
certain amount of time before they could get the
diesels back up. And it's not exactly a self-
imposed station blackout but it is in response to
some spurious actuation that could happen in an
area. It was a short period of time. I think 7-1/2
minutes.
MR. HOWELL: One of the things that we
did note during the inspection was that just prior
to us coming out there they had spent a lot of time
on operator training and making sure that they could
meet the time lines that they had established, and
so it's not at all clear that until they did that
that they would have achieved those time lines.
We have completed -- on slide eight
you'll notice the number of triennial fire
protection team inspections we've completed. We've
completed eight. Rebecca mentioned ANO. That one's
not listed because it's not completed yet.
On the next slide, which is also
reflected by the chart up there is a breakdown of
the findings by type for the eight baseline
inspections, the team inspections, as well as the
findings from the resident portion of the inspection
procedure. With respect to the team inspections we
found findings at six of eight sites, the two
exceptions, Palo Verde and River Bend.
I think it's interesting to note River
Bend is a plant that has had chronic fire protection
issues throughout the '80s and '90s. Jake himself
has been responsible for finding some significant
issues in the early '90s that resulted in escalated
enforcement. This is the plant that started the
thermo-lag issue. They also received a fire
protection functional inspection in 1997, a number
of issues there with associated circuits.
And so we went out there just recently
last month I believe and we had no significant
findings, and my read on that it's a testament that
after all this time they've finally implemented some
corrective actions to address issues in the fire
protection area.
As you can see, we have issues in
separation, which also includes passive barriers,
and that's really the only major trend if you will
or pattern. A few issues in detection and
suppression, emergency lighting, transient
combustibles, and fire watch training. I talked
about some of the conspicuous absence of findings,
lube oil collection system findings because we don't
inspect those any more under this procedure, and
associated circuits. We have about a half a dozen
unresolved items, apparent violations on associated
circuits at both BWRs and PWRs that we are waiting
to disposition.
Just looking at the groupings, in the
separation area this represents a gamut of unlatched
on inoperable fire doors, degraded fire wrap, holes
in ceilings that separate fire rooms or fire areas,
degraded seals in one case, intervening
combustibles, and lack of cable separation either
not meeting the 20 feet in 3G2 or not meeting what
they said in their exemption requests.
CHAIRMAN SIEBER: Fire dampers -- are
they continuing to be a problem or don't you know?
MR. HOWELL: I believe we may have,
what, one issue involving unqualified fire dampers
in these 19 findings.
MS. NEECE: There might have been one.
It was a resident --
MR. HOWELL: Right. In detection and
suppression -- this is primarily involving not
placing detectors per the NEPA code, or in one case
there were sprinklers that they changed the diameter
of the sprinkler head holes without evaluation.
Emergency lighting, inadequate corrective actions
for unreliable DC batteries for some of the
emergency lightings at one plant, and in one case
inadequate lighting for an operator to implement a
manual action to open the service water valve which
supplied reactor equipment cooling to the
hypercooling injection system, and that was at
Cooper Nuclear Station.
Transient combustibles include either
not being on the permit or not being in the program,
just overlooked it totally, and then fire watch --
one instance in which members were conducting fire
watch duties and they hadn't been trained.
All these issues were green per the fire
protection SDPs. We had two that were borderline
white and were ultimately dispositioned before we
got to a regulatory conference. One of those
involved cable separation issues at Fort Calhoun
Station.
They -- in one particular fire area they
had only about three feet of separation in between
redundant trains and safe shutdown, and in this
particular fire area it had cabling that fed almost
all their accident mitigation motor-driven pumps,
and this was a case that was complicated by the fact
that they had submitted an exemption request in the
mid-80s indicating that we don't meet the 20 feet
but we have ten feet and we have suppression and
detection. So the staff granted the exemption based
on having a little bit of basically 3GA and B and C.
And when we went out Rebecca was the
team leader, went out, did the inspection. We found
that no, they didn't even meet the ten feet that
they said they had in the exemption request. They
had three feet in some cases. That one was
borderline. it ultimately -- correct me if I'm
wrong -- it hinged on whether or not automatic
suppression would extinguish the fire before the
cabling that fed the fire water pumps was in fact
damaged, which also went through the same room.
That one took a lot of time because it
wasn't real clear to us that the licensee had a good
handle on what cabling powered what equipment. It
took quite a while to identify the equipment list,
which also complicates exercising the fire
protection SDP, and so it took a number of weeks
before we dispositioned that issue.
The other one was more straightforward.
It was actually a three-hour rated fire door at ANO
separated, both violates the switch gear rooms and
it turns out in that case they had a -- although it
wasn't really documented and they weren't taking
logs they did have an ineffective roving fire watch
that was going through there, so they had a comp
measure in place.
Again, no obvious trends or patterns
with the exception that most of the findings or
certainly the significant portion of the findings
are in the separation area, which is not unexpected
given the focus of the inspection. But again,
what's a little bit troubling is that there are
three or four examples here in which exemptions were
granted and either the original plant configuration
that formed the basis for the exemption was never
met, or it was changed as a result of modifications
that occurred and was not detected over the years.
That's a summary of the findings. We
touched on the challenges. I mentioned some of
them. One is -- this was a new area for us, and it
was a significant increase in level of effort, and
so we enter this new program with some trepidation
in the fire protection area, and through the use of
the short-term formal training with Brookhaven and
the reliance on contractors in part and OJT we've
been able to implement the program, and quite
successfully I think.
We're finding issues that we clearly
would not have found under the old program, but then
the question is how significant are they given the
tools that we have? There's still some questions
about implementing the fire SDP, which we talked
about earlier.
MEMBER POWERS: Several of the findings
come out as green based on ignition frequency
arguments. How do you make those arguments?
MS. NEECE: Several of the findings come
out green because of the ignition frequencies are so
low. What we found in running the SDP from the site
you run a phase two SDP, and it's a simplified
version. You're not taking into account the
probability of a spurious actuation, a probability
of fire affecting this area. If they don't provide
the requisite level of protection you assume a
credible fire but you -- because we don't take into
account the probability of spurious actuations or
the probability the fire might not reach to a
certain point you basically assume everything in
there that's not protected is consumed by the fire.
We have found that if we have a
degradation in suppression that seems to be more
significant than the ignition frequencies. The
ignition frequencies for the areas that we choose
usually run around 1e to the minus three, 1e to the
minus four. They're all about the same. The
differences in the ones that are borderline white
and ones that are clearly green have to do with the
degradation we give them for suppression, and that
makes sense if the fire can be suppressed to the
point that they fire brigade can respond in 15
minutes and there not that much damage. Then it
makes sense for it to be a green issue rather than a
white issue, so that's been my experience so far.
Again, another concern that I would have
is relying on the ignition frequencies we get from
the IPEEE. The IPEEE is not required to be revised
or a control document, and as changes go along in
the plant I'm -- we have to use that ignition
frequency in the phase two, and it's developed by
the licensees and it's not revised as the plant is
modified or changed.
MR. HOWELL: Isn't it our experience
that some of those are actually conservative because
if one considers a credible fire as opposed to any
and all fires in a particular fire area the
frequency may be less?
MS. NEECE: Yes. That's correct.
Did I answer your question?
MEMBER POWERS: Maybe.
MS. NEECE: Do you have another one?
MR. HOWELL: You made an earlier point
earlier in the day about it all boils down to how
much credit one gets for automatic suppression, et
cetera. Yes, and there's a lot of latitude there,
and ultimately that has affected some outcomes. I
believe Forth Calhoun was initially three greens
next to a white, or was a white --
MS. NEECE: It was a white.
MR. HOWELL: -- until we came to the
conclusion that suppression would extinguish this
fire before the fire water pumps were put out of
commission.
MEMBER POWERS: How do you decide on the
response time of the fire brigades?
MS. NEECE: How do we decide on the
response time of the fire brigade?
MEMBER POWERS: Right. You've got a
fire in a particular fire area. I may or may not
have automatic suppression. I certainly can't count
on that to put the fire out, so I need the
firefighters to get to that to respond and put out
the fire. How do I estimate how long it takes them
to do that?
MR. HOWELL: To the extent that that
information is available, which it may not be in
every case, we would consider it, but clearly it
isn't available and so we have to fall back to
what's been our experience in observing the fire
brigades over time and have we identified
performance problems. And essentially -- correct me
if I'm wrong -- if there have been no documented
issues and there's no time line that we can verify
in terms of response time we default to giving them
maximum credit under the fire protection SDP.
MEMBER POWERS: You give them maximum
credit?
MR. HOWELL: Yes. Right.
MS. NEECE: Normal operating --
MR. HOWELL: Yes. If there's no
performance deficiencies based on observations of
the drills and absent any other negative information
they get credit.
MEMBER POWERS: So you really don't have
a database to draw upon in general?
MR. HOWELL: True. And -- but as you
noted or as I noted, we do now have at least
provisions to monitor fire brigades, although it's
not particularly frequent. We have an opportunity
to build that database with time through
observation.
MEMBER POWERS: We'll certainly discuss
San Onofre, discuss the barriers to effective
firefighter response and communications with the
control room.
MR. GWYNN: And Clyde Osterholtz is
here. He's the senior resident inspector at San
Onofre. He led the team that responded to the fire
at San Onofre. At that time he was not yet assigned
at San Onofre, but if you'd like I'd like to ask
Clyde to go ahead and make his presentation on the
San Onofre fire.
MR. OSTERHOLTZ: It's a great lead in.
Thank you, Pat. I'm going to try to make this as
brief as possible because I know we're a little bit
behind.
What we had here at San Onofre is
essentially a secondary breaker failure that had
complications which made a resultant reactor trip
and a complicated recovery. DC lube oil pump for
the turbine didn't start when it was supposed to, so
the turbine had to -- had grinded down in about two
minutes when it should have gone down in about 2.5
hours, so they were down for a significant period of
time preparing that turbine work. I think everybody
is aware that I had some pictures but I think you
mentioned that most folks have seen those.
So just briefly, the plant was at 39
percent power on February 3 when they were going to
switch from the reserve auxiliary transformers to
the unit auxiliary transformers, and as most of you
are aware this is a normal practice to get your
house loads on your turbine generator instead of
depending on offsite power. When that happened in
bus 3A07 breaker 12 developed a fault where the
phase Charlie portion of it partially closed but
didn't fully close, and that was determined to be
caused by increased resistance in the breaker
contacting mechanism.
That was one possible explanation, or
the other likely explanation was that there's a
fiberglass pusher in side that breaker that may have
had a crack and failed.
MEMBER POWERS: Now, that one I didn't
know about.
MR. OSTERHOLTZ: That's not in the
report. That is in their root cause analysis and
it's something they're still looking at now.
They'll never I don't believe --
MEMBER POWERS: Well, it's fried.
You'll never find out --
MR. OSTERHOLTZ: Right. A definitive
root cause analysis to this problem.
MEMBER POWERS: Is the manufacturer
looking at it?
MR. OSTERHOLTZ: Vendors are involved,
and they're looking at it as well. They're also
looking at increasing the frequency of how often
they look at these breakers, do refurbishments, and
perform inspections on them.
The big complication here was the
breaker that attaches to the reserve auxiliary
transformers is only two cubicles down, and although
its mechanism to not reshut back onto the reserve
auxiliary transformers functioned correctly, it
arced from the ionizing gases developed from the
fire in the 12 breaker. That subsequently forced
the reserve auxiliary transformers to trip, and as
you can see at the bottom of your handout I've
divided out into the 6.9KV reactor coolant pump
buses, the vital buses, and the non-vital.
So at this stage of the game, since
you've got the RATs seeing the fault, you've got the
unit AT seeing the fault, everything goes to unit 3
as far as the 6.9 and the 4KV vital are concerned,
and you lose the secondary buses, and that
subsequently meant that you lost your AC lube oil
pump for the turbine and the DC lube oil pump didn't
start.
CHAIRMAN SIEBER: Someplace I either
remember or am mistaken that -- were the dividing
metal shields from one cubicle to another in place
when this failure occurred or were they missing?
You know how you put metal-clad switchgears broken
up into cubicles? There's metal shields between
them.
MR. OSTERHOLTZ: The metal -- as far
as -- our inspection determined that everything was
in place that should have been there in between
those two breakers.
CHAIRMAN SIEBER: I'm probably mistaken
then.
MR. OSTERHOLTZ: Okay.
MEMBER POWERS: The magnitude of the
fire is such that the shields wouldn't have made any
difference.
MR. OSTERHOLTZ: Right. And I know you
are all interested in automatic fire mitigation
equipment. This secondary switch gear room had
none, but it did have fire detection equipment. I
offer that out for you as well.
In addition to all of those problems
this fault caused DC grounds about 800 amps worth
between the secondary battery and ground, which was
just enough to give you a significant problem but we
not enough to trip open your protective breakers, so
therefore they lost control room annunciators.
MEMBER POWERS: There must be some sort
of rule that that's what's going to happen.
MR. OSTERHOLTZ: So they had a
distribution panel in the control room that fed
power to the control room annunciators and tried to
reset that breaker. It retripped, so they
subsequently just stripped the bus, shut that
distribution panel breaker and were able to restore
control room annunciators in about 14 minutes. So
we gave them a thumbs up for that, because we
thought that was above average.
They did enter an unusual event based on
a fire that could have or was adjacent to areas that
had safety-related equipment in it. In retrospect
they believe they never really had to enter the
emergency plant at all because of the location of
the fire and the fact that it didn't affect any
safety-related equipment. And their emergency plan
is structured such that they have a specific list of
what is the definition of when you have to enter an
unusual event what equipment is affected, and none
of it was subsequently involved.
MR. LARKINS: Would the loss of the
control room annunciators have driven them to
that -- to an emergency plan --
MR. OSTERHOLTZ: We looked very closely
at that. Loss of control room annunciators gets you
into an unusual event if you lose them for 15
minutes --
MR. LARKINS: Okay.
MR. OSTERHOLTZ: -- and their logs had
them down --
MR. BROCKMAN: Now, this is an
interesting point because you're getting into a lot
of legalistic things here with respect to do you
have a violation? Do I have to make an appropriate
report within X amount of time, and not into the
aspect of is the right thing to do to utilize some
of the facilities for marshaling people and
controlling and what have you, and that's why they
get so particular on some of these issues, and
you're really getting into the legalisms of
enforcement.
CHAIRMAN SIEBER: The more interesting
thing comes later.
MR. OSTERHOLTZ: I wasn't really clear
on that though. The answer to your question is the
loss of annunciators automatically gets you into an
unusual event, so that did apply. If you get into
loss of annunciators for more than 15 minutes it
goes to an alert from an unusual event. I just want
to make that clarification.
And subsequent recovery -- we had a five
man fire department team show up at the scene. San
Onofre is different than every other plant I've seen
where they don't have a dedicated fire brigade made
out of control room or licensed operators, security
personnel, et cetera. This is a dedicated fire
department, and it showed up. Had a fire chief who
is the fire chief for the site who happened to be
there on time. In fact, most of their senior folks
happened to be there because they were in this
evolution of starting the plant up after an outage.
MEMBER POWERS: You realize that they
were about to host the fire protection forum in San
Diego the next morning.
MR. OSTERHOLTZ: In any case, when the
firefighters got to the cubicle in question there
was heavy, thick smoke. They began ventilating.
They used haylon PKP portable fire extinguishers. I
think they exhausted between 22 and 24 total
canisters. They had the fire under control. There
was some communications problems between the shift
manager and the fire chief at the scene because they
had a liaison who was an operator -- licensed
reactor operator transferring information.
There was a little bit of confusion.
The fire chief reported no flames visible. That was
translated to the control room as the fire was out,
when actually the fire was under control but the
cubicle door was still closed and they just kept
flashing it with powder, and then every time they
opened the door it would reflash. They'd hit it
with more powder and keep the door shut.
And we estimated there was about a 16
minute delay in getting water put on the fire
because the shift manager was reluctant to give that
authorization even though the fire chief -- once the
fire chief spoke to him personally the shift manager
was convinced, the door was opened, and the fire was
completely extinguished using water.
MEMBER APOSTOLAKIS: I thought that
issue of using water had been settled after Browns
Ferry. We still have this hesitation?
MR. GWYNN: We saw the exact same
characteristics at Waterford during a significant
fire very similar to this --
MEMBER APOSTOLAKIS: So there's still a
reluctance to use water?
MR. GWYNN: Yes. And it depends on how
the people have been trained, and in particular the
control room folks, whether they came through the
Navy program, whether they've been trained
subsequent to that. The Navy trains people you
never put water on an electrical fire, but in fact
the industry knows that you can safely use water on
an electrical fire under controlled circumstances,
and so it was a training issue at Waterford. We saw
remnants of that here --
MEMBER POWERS: It's a training issue
here as well.
MR. GWYNN: -- where the fire brigade
knew the criteria and knew the approaches, but the
person who was in charge in the control room was
reluctant.
MEMBER APOSTOLAKIS: He was from the
Navy?
MR. OSTERHOLTZ: As most of their
control room operators are.
MEMBER POWERS: I think it's a training
issue here as well. I think these people were just
not familiar with the process.
MR. OSTERHOLTZ: That was one of the
things that we brought up to them. The licensed
operators since they are not involved in fire
brigade activities don't receive training on
advanced firefighting techniques such as using water
on energized equipment. I think that added to some
of the confusion. The licensee saw it more as a
command and control issue where they're going to
make sure the shift manager understands the fire
chief is the expert. He's the one in charge. Take
his advice when you're in these situations.
MEMBER POWERS: What do you do when the
fire chief isn't -- just doesn't happen to be there?
MR. OSTERHOLTZ: Then there's a
designated incident commander assigned to the fire
department to perform that function.
MEMBER APOSTOLAKIS: Now, this
reflushing when the portable fire extinguishers were
used, is that something that's common?
VOICE: Yes. Any time you've got a fire
in a cabinet --
MR. OSTERHOLTZ: Yes. You have a medium
that takes the fire and puts it out, but then when
it's dispersed as oxygen you come back into the
area.
MEMBER APOSTOLAKIS: So it just tries to
starve the fire?
MR. OSTERHOLTZ: That's correct.
MEMBER APOSTOLAKIS: So why are we using
them at all?
MEMBER POWERS: These dry chemicals are
simply oxygen displacement devices, and they in fact
what they act is a nice insulator to assure the
stuff is nice and hot, so as soon as oxygen comes
back to it it flashes. It happens all the time
in --
MR. OSTERHOLTZ: And although we noted
that there was those 16 minutes delay in using the
water we did conclude that it really didn't have any
effect on the outcome of the event because they had
the fire totally under control, isolated, and it was
completely away from any of the safety --
CHAIRMAN SIEBER: Completely away may be
a little strong, isn't it?
MR. OSTERHOLTZ: When I say completely
away we felt that it was far enough away where the
fire could not affect safety related equipment.
CHAIRMAN SIEBER: With that amount of
control on it. Had the activities been delayed
substantially then it would have been a worse fire.
I don't fault your report. I think the report's
right, but -- in fact I enjoyed your report.
MR. OSTERHOLTZ: Thank you.
And that's it in brief. It took them
some time to recover because of the significant
turbine damage. However, just in ending I'll tell
you we were very impressed on their startup because
of this -- there's eleven journal bearings in this
turbine. They're all different sized now because
they had to lay the thing down because of the damage
done to the shaft, but when they started up they
expected to have to come back down to do rebalancing
work, and they started up and went completely up
without having to do any of that, and their
vibrations are consistent with what they have on
unit 2.
So that was -- we were pleased with the
quality of the work that went into that turbine.
CHAIRMAN SIEBER: Now, they changed out
the old English electric turbines there? They
replaced their turbines. Right?
MR. OSTERHOLTZ: They're still the
English electric design.
CHAIRMAN SIEBER: Are they?
MR. OSTERHOLTZ: Yes. In fact, they're
the only ones left. I think Fermi was the last
other plant that --
CHAIRMAN SIEBER: They aren't too
smooth.
MR. BROCKMAN: In fact, the station
management as part of their recovery operations
visited England to see some of the work that was
being done and wasn't happy that their plant wasn't
operating and the Brits were taking the weekend off.
MEMBER LEITCH: The loss of the DC lube
oil pumps to the turbine I guess because that's not
safety related you didn't go down that road?
MR. OSTERHOLTZ: We looked at it. It
was not something that we spent significant time on,
because although the destruction of the turbine was
a significant financial loss for them it really
didn't impact the event safety wise -- rector safety
wise on our end.
However, I will let you know that part
of that corrective action is they're now going to
have two redundant DC lube oil pumps for each
turbine so if this ever happened again they would
have a backup.
MEMBER LEITCH: But the loss of that DC
I believe was related to the miscalibration of the
DC breaker.
MR. OSTERHOLTZ: An over current
breaker. It was more of a mispositioning after
calibration. You have a low to high. They did a
bunch of testing in the lower range and they thought
they were leaving it in the high range but they
actually went too far around and now we're at the
bottom of the low range again.
MEMBER LEITCH: Did you take a look at
whether that was generic? Although that was the
balance of the plant did that -- could that kind of
an error have occurred in the safety related
equipment?
MR. OSTERHOLTZ: Yes. The breaker
specialist did look at that and determined that it
was an isolated problem to that maintenance
activity.
MEMBER LEITCH: Because it sounds as
though it may be generic to that type of breaker I
guess, and that was not the case?
MR. OSTERHOLTZ: Not the case.
MEMBER LEITCH: Okay.
MR. OSTERHOLTZ: In fact, one of the
other things they're looking at is getting rid of
that overcurrent device completely for this
equipment, because their view is who cares if you
burn this pump up? You let it supply lube oil to
the turbine as long as possible.
MEMBER LEITCH: Right.
CHAIRMAN SIEBER: Save the pump and lose
the turbine?
MR. OSTERHOLTZ: That's what happened
unfortunately in the case in February --
CHAIRMAN SIEBER: Did they damage
anything else besides the bearings, the shaft, and
perhaps seals?
MR. OSTERHOLTZ: Exciter had significant
damage, had to be shipped --
CHAIRMAN SIEBER: Okay.
MR. OSTERHOLTZ: -- by airplane to
Virginia I think. It was a horrendous expense.
MR. BROCKMAN: The front thrust bearing
when I was out there was really something to see.
Imagine stopping your car from 90 miles an hour with
no brake pads and what your discs would look like
and the coloring and the striations and everything.
That's exactly what happened. From 1,800 RPMs the
front thrust bearing was the spindle brake, and it
looked it.
CHAIRMAN SIEBER: Okay. Any other
questions?
MR. LARKINS: So the violation here was
one non-cited green?
MR. OSTERHOLTZ: One non-cited green. I
didn't get into that because it really didn't affect
the fire, but they did overfill a condensate storage
tank inadvertently because of a difference between
unit 2 and unit 3. A fill value for unit 3 fails
open on a loss of power. The fill valve on unit 2
fails shut. So the operators were thinking unit 2
and inadvertently left water going to the condensate
storage tanks.
It overflowed and it's in a vault that's
seismically qualified. It got up to about 12 feet
in the vault and at the bottom of this vault are
valves that will cross connect the main tank to its
backup tank in case that tank empties to give it
seismically qualified water, and that tank was
effectively rendered inoperable because you couldn't
get to the valves because they were 12 feet
underwater.
CHAIRMAN SIEBER: Did it float the tank?
MR. OSTERHOLTZ: There was a nitrogen
blanket at the top of the tank that did burst. Yes.
CHAIRMAN SIEBER: But it didn't float
the tank off its structure?
MR. OSTERHOLTZ: No. It did not.
CHAIRMAN SIEBER: Okay.
MEMBER LEITCH: Absent that event there
would have been no violation at all then? Is that
correct?
MR. OSTERHOLTZ: Had they realized that
valve was opened and shut it and controlled the
condensate storage tank level there would have been
no findings of color.
MR. GWYNN: So I stand corrected on my
statement earlier. There were some safety
implications --
MEMBER POWERS: The most significant
thing is just this communication from control room
to fire brigade issue, and I guess you feel like
they've handled that issue?
MR. OSTERHOLTZ: They've embraced it.
It's in their corrective action program. It may be
too early to say definitively that they have
completely resolved that problem.
MR. SINGH: How do you correct it?
CHAIRMAN SIEBER: It's a fact that fog
nozzles can be used on electrical fires provided
it's not saltwater.
VOICE: It sounds like they've gone
policy wise here.
MS. NEECE: Yes. Can I make a comment?
VOICE: Policy training, and we'll
observe it during drills and what have you and see
if they test it, and other people say, Yes. I
understand you're in control. You say water, go
with water. That's the way it will have to be --
MR. SINGH: I was going to make comment.
After the Waterford fire, when they pour water on
the -- there was a counterpart meeting and the
gentlemen from SONGS were there, he was sitting next
to me at NEI conference when this happened. Anyway,
they have already administrative procedures in place
to tell the fire brigade what to do or what not to
do, so I don't know if he was familiar with the
procedures or not or what happened. I have no idea.
But they were in place at that time.
MR. GWYNN: Yes. It was a matter of
this one individual in the control room who was in
charge who was reluctant to have the fire brigade do
what it knew it was supposed to do.
MR. OSTERHOLTZ: We've got to be careful
about getting into the mode of calling it a fire
brigade, because I got into that -- was making that
mistake and I was confusing some folks because it's
not -- when you say fire brigade people think of
operators and security people. It's a dedicated
fire department.
CHAIRMAN SIEBER: All right. I think we
have time to finish our last topic here.
VOICE: And in fact if Mr. Andrews and
Mr. Pellet would come up and --
MR. GWYNN: And while they're doing that
there was a question that was asked earlier
concerning the Callaway capacitors. Those
capacitors are in fact connected at all times.
They're basically a UPS. You'd expect them to be.
They can take them off for maintenance if need be,
but they can take them off for testing and
maintenance, but normally they are engaged and on at
all times.
For this topic, the Region IV
responsibilities under continuity of operations and
continuity of government we have Mr. Tom Andrews,
who's our emergency response coordinator here in
Region IV. Tom, would you hold up your hand. And
Mr. John Pellet, who's the chief of our information
resources management branch in the division of
resource management and administration, and they're
going to share some information with you about
Region IV's unique role as the backup to
headquarters for continuity of government,
continuity of operations.
Tom.
MR. ANDREWS: Good afternoon. I want to
make sure your understanding of our continuity of
operations plan, the idea that we have implemented
ties back to some time ago somebody realized that we
might lost headquarters. An event occurred up the
road here in Oklahoma City, for example.
Several years ago there was an incident
in Oklahoma City that received a lot of notoriety
and demonstrated that an act of terrorism could
adversely affect a large structure, and from that
time on there's been a lot of focus on continuity of
operations and continuity of government. When you
hear the term COOP and COG you can now know that
COOP stands for continuity of operations. COG
stands for continuity of government. Under the NRC
continuity of operations plan we view continuity of
government as a type of continuity of operations
event.
In our plan we talk about our critical
functions. Each federal agency had to go through
and identify what they consider to be their critical
functions and describe what mechanisms they were
going to put into place to protect them. In the NRC
we have one thing, and if we only do one thing in
life then we will survive as an agency. If we
respond to events we're going to protect the health
and safety of the public. Everything else we do
that makes sure that we don't ever have to get into
the situation of having to respond to events. It
helps to make sure that licensees are doing the
right things up front, but things still happen.
Emergency response covers a lot of
territory, and when you try to decide what covers
the emergency response function -- I'll give you an
example of what that includes. That includes the
receipt of the event notification, whether it be
from the licensee or resident inspector, member of
the public, or another federal agency. Performing a
screening type assessment of the information
provided and then determining what forms of internal
notifications need to be made, and then going from
that to determine if there needs to be some elevated
form of response. Do we need to activate our
instant response plan?
We might need to call in people, staff a
center to perform assessments and monitoring of the
conditions, licensees' actions, et cetera,
communicating with state and other federal agencies
regarding the event, assessing licensees' ongoing
actions and any protective action recommendations
that they may be giving to state and local agencies
so that they can protect the public, and
coordinating the technical response from the federal
government.
The primary resources that we use to do
this is communication. The NRC does not have a lot
of physical resources that we take to the field for
emergency response. We don't have -- like other
agencies we don't have trucks and helicopters and
satellites and things to deploy. The thing we bring
to any emergency response is brainpower, and the way
we engage that is you have to feed it, and that's
through communications.
So we have a lot of very diverse means
of communication paths. We have our federal
telecommunications service system. We have
commercial telephone systems. We have satellite
telephones at the reactor sites as well as in our
response centers. We've got cell phones. We've got
network for e-mail, et cetera. So we have a very
diverse set of communication path that we can use.
We have evolved our response process to
the point where we use a lot of this equipment in
focused centers. Many of you have probably toured
the headquarters operations center and realized that
is a very robust response center. It has a lot of
capability. But if something were to happen such
that headquarters could not operate or could not be
used or it was no longer there how would the NRC
deal with some form of event like that? So we want
to protect our critical function, and that's why we
have continuity of operations plan.
Region IV was selected as the backup for
headquarters. In the continuity of operations plan
we're referred to as the default region.
Technically any of the regions can stand in the role
of headquarters as far as event response and being
able to staff a center and coordinate how we're
responding to events and communicating with other
agencies. The difference for Region IV is at
headquarters we have a headquarters operations
officer, a person that's on shift 24 hours a day to
receive that first phone call to initiate the
response. They would call the appropriate region,
get decision makers on the phone, and kick off the
response.
If something were to happen to
headquarters we would be picking up that role.
Why was Region IV selected as a backup?
Well, it ties into some lessons that we learned from
Y2K. In preparing for Y2K which in itself could
have been a continuity of operations type of event,
we selected Region IV as the backup for headquarters
for that purpose, and the reason being was we're a
long ways from DC. It takes a real big event if
it's a weather type of event or some other type of
disaster that affects DC to also us. We're
typically in a different weather pattern and we're
on a different electrical grid.
I know you've been talking about grids,
and in the case of the United States there's three
main electrical interties. There's the eastern
interconnection, which you can see would cover
Regions I, II, and III, as well as headquarters, so
if you had a massive blackout that cascaded across
the whole interconnection it would take out all of
those offices, or it would impact all of those
offices.
Texas is pretty much its own grid to
itself. Not only is it entirely within Texas, it
doesn't go outside of Texas, but there's different
types of connections between the ERCOT grid and the
interties and what you find inside. It has to go
across a DC connection to get into the ERCOT grid or
out of ERCOT grid.
MR. GWYNN: That situation with Texas is
consistent with the state constitution that says
that they can secede from the union at any time
without prior notification, so the ability to
disconnect from the electrical grid in the rest of
the United States is an important part of that.
MR. ANDREWS: And the advantage of
having that DC type --
MEMBER POWERS: How does that impact
your ability to serve as a backup for headquarters
if they decide to secede?
MR. ANDREWS: It promotes international
relations. We'll have a field office in their
country.
MEMBER POWERS: Will you be citizens of
their country?
VOICE: If they secede does the
president have to resign?
VOICE: He's left. He don't live here
no more.
MR. PELLET: It was actually a FERC
issue and the two major Texas utilities prefer not
to cross interstate boundaries in transmission of
electricity.
MEMBER UHRIG: They refuse to serve
beyond their boundaries until something is settled.
MR. ANDREWS: The good thing about the
DC type interties is that if there is a disturbance
on either the western grid or the eastern grid it
doesn't propagate into the ERCOT grid.
Just like I pointed out earlier, if
something happened on the eastern grid where it
affected Regions I, II, and III as well as
headquarters but not Region IV, likewise if
something happened on the ERCOT grid it wouldn't
propagate out. The reason for that statement is to
tell you that being a backup doesn't mean we're
bullet proof.
Now, I've told you about the
communications that we have and we use, and they've
focused in our response centers. We've got some
other equipment that we use that is considered to
support our critical function, but not necessarily
critical that we operate. If the NRC had to operate
for a longer period of time, more than just a couple
of days, we would need to have means to communicate
internally as well as externally and have means of
accessing the internet. The internet has become a
very important part of being able to conduct
business.
Now, I'll mention that the local and
wide area network is primarily our internal computer
system. The e-mail is between the offices, between
various regions, the sites, and headquarters,
whereas the external internet access is our ability
to go out and look for information on the internet.
One of the things that we may lose for a period of
time is we may lose our web page, but that's not
necessarily considered to be vital for our
operation.
CHAIRMAN SIEBER: Is Adams vital to your
operation?
MR. PELLET: The answer to that is no.
CHAIRMAN SIEBER: I'm not surprised.
MR. PELLET: Adams was not a required
support function under a COOP-COG activation.
CHAIRMAN SIEBER: Okay.
MR. PELLET: Neither is Star Fire, Pay
Pers --
CHAIRMAN SIEBER: I understand that.
MR. ANDREWS: I'm going to let John talk
about this diagram.
MR. PELLET: This is a busy slide.
For those of you who didn't hear, I'm
John Pellet, chief of the Information Resource
Management Branch in the region. If it has an
electron or a piece of paper attached to it it falls
within our purview.
And basically all this slide is
attempting to tell you is these little yellow lines
are what's being added to our computer
infrastructure for COOP. Right now all of the
agencies' infrastructure outside of the office,
outside of Region IV, outside of Region I, outside
of Region II goes through White Flint. If White
Flint goes away we can't talk to Region I today in
terms of computer support.
Under a COOP environment we need to have
redundancy to where we can bypass the headquarters
infrastructure. This is going to involve a series
of hardware-software changes to the agency's
network, a lot of -- from my perspective a lot of
money being spent. From any other federal agency's
perspective probably not a whole lot, but
essentially we're going to add several racks of
computer equipment into our space. We're going to
add considerably more network connections across
between offices. In essence, we're going to double
our existing bandwidth to offices by having a
redundant pipe that goes around headquarters.
Region I is actually the backup backup
facility. If we were to be lost and still have a
COOP scenario Region I will have some capability to
come back, but basically the wide area network
connection outside of each office probably would be
lost if we were to be lost with headquarters
infrastructure wise.
Of course, this is focusing on
computers. There's a lot of telephone
infrastructure changes required to support us being
able to redirect and handle emergency phone traffic
for the agency. As we've demonstrated before,
that's not something a region is normally prepared
or configured or has the infrastructure to do, but
that's something that's being added as part of COOP,
and it's something we're testing and implementing as
we go.
Basic time line for COOP -- of course
the agency is fully COOP functional now. The
computer infrastructures and the telephone
infrastructure stuff will be done across most likely
the remainder of this year. We're going to have
some facility changes in the region to better
support a COOP environment in a more smooth manner,
and all that's under current development.
But the thing I would say take away from
this is two things. One, we're operational in a
COOP context now. We hope to make it much easier
with infrastructure changes in the near future.
They're being worked between all the regions, OCIO,
IRO, Region IV. Tom and I are on a regular calls
every week about COOP infrastructure requirements.
Anything about infrastructure?
MEMBER LEITCH: Are these yellow lines
shown on your diagram -- I'm trying to visualize
what they represent. Is that hard wire or what is
that?
MR. PELLET: In today's infrastructure
world what that really -- this is not going to
involve a new wire coming into our building. We
actually have a piece of fiber cable that comes in
from MCI that's capable of carrying more than enough
to do all of this. It's actually called a DS3 cable
connection into our router but it's provisioned into
these separate virtual circuits.
MEMBER LEITCH: Okay.
MR. PELLET: And so we haven't fully
negotiated with MCI, the local carrier, which is
Southwestern Bell, and the building exactly how
we're going to bring this new data bandwidth into
the office. It could be a whole new pipe wire. It
could be just an additional speed down the wire we
have. It could actually be three new wires. That's
a contractual issue and a local telecom
infrastructure compatibility issue that we don't
have fully worked out and is somewhat dependent on
whether we end up renewing our lease in this
building or moving.
We can't quite finish all of that
negotiation until our lease negotiations are
complete and we know we're staying here, because
obviously there are capital investment requirements
to increase the size with the building and the local
carrier and the FTS 2001 carrier.
So the answer to your question is I
think it's going to be in the one pipe we've got.
COOP is not intended to be redundant. It was a
design decision made long ago. If you notice if we
lose this box right here, which is an actual box
sitting behind that wall, we can't activate the COOP
computer infrastructure part of COOP. COOP is not
intended to be single failure proof throughout the
industry.
It's not intended to be. It is very
robust. We have two potential paths. One's in 2
White Flint and one is in 1 White Flint. We can
lose one building in White Flint and not have a
computer infrastructure COOP problem. It will
automatically auctioneer back and forth. So with the
redundancy we have to harm our COOP function we
would have to lose both of these plus this.
MR. GWYNN: John, did you mention the
red boxes?
MR. PELLET: The red boxes are
essentially internet connections. One of the things
we're going to do -- obviously we think COOP
decision agency and staff decided being able to
access information on the internet was an essential
part of our event response. Therefore, since we
currently have one pipe to the internet through NIH
we're going to be adding a second pipe in standby
mode from here, and the red boxes are firewalls, and
we've got fire in this again.
So that's the function of these external
connections, and of course each region connects out
to each of its sites through its own equipment.
That's also -- in fact we have 14 sites, each with a
data pipe going out. They're all coming into that
same MCI Worldcomm FTS 2001 pipe. It looks like
individual pipes if you look at it from a schematic,
but from the electrical standpoint it's one piece of
fiber.
MR. GWYNN: Tom, we are essentially out
of time. Could you show the IRC plan very briefly
and then we'll conclude the presentation?
MR. ANDREWS: As John mentioned we are
spending a fair amount of money, at least for us.
We don't usually see that much money come through
here.
One of things we're doing is we're going
to be remodeling our instant response center. The
idea being when we responded for Y2K we put about
35-40 people here in the office to respond to Y2K.
Although our response to Y2K was quite successful it
was not pretty. We had to use offices outside of
the center in trying to keep in touch with everybody
and make sure everything stayed coordinated was not
easy.
So what we've done is we've talked with
admin and we're going to be ripping out the walls
and making a lot of changes to more efficiently use
the space that we have. So what you see here is
what we're looking at. To give you an idea, our
center has not been really upgraded since around the
1990 time frame. It's a very low tech center right
now, and we're going to be adding some things to it
to make it more usable and to help us more
efficiently use the space.
MR. GWYNN: And primarily the key COOP-
COG changes being made is to include in the design
two headquarters operations officer consuls so that
the WHO function can be transferred quickly to
Region IV. We would pick that up essentially
instantaneously. We'll have a computer here which
will be monitoring headquarters availability. If
the computer loses connection with headquarters for
more than a preset period of time then we
automatically go into COOP operations and our people
respond to the incident response center and initiate
COOP function until that -- the individuals from
headquarters could be restationed here.
MR. ANDREWS: The next slide just
basically tells how we would kick of the continuity
of operations process here in the region.
Do you have any questions about COOP or
COG?
(No response.)
MR. ANDREWS: Okay.
CHAIRMAN SIEBER: Thank you. We
appreciate the presentation, and it looks like we
made it all the way through the schedule.
I would like to express on behalf of the
ACRS and our staff our appreciation for the work
that you went through to put on these presentations.
I particularly liked the free flow of information
and our ability to ask questions and get a better
understanding of issues that we think are important
to us performing our function.
And so my congratulations to you, Pat,
and to all of the staff here at Region IV for your
hospitality and cooperation. I'm curious -- if this
ACRS meeting differs from what your expectations
that it would have been prior to our arrival. Did
you expect this kind of a meeting or interchange, or
did you expect something different?
MR. GWYNN: Based on my previous
experience observing ACRS meetings at headquarters
and the last meeting that we held here in Region IV
I think that it was pretty much what I expected.
CHAIRMAN SIEBER: Okay.
MR. GWYNN: I know that when the ACRS
asks you a question you need to get an answer, and
so this forum was perfect for that purpose. I
thought that it was extremely valuable to have this
dialogue.
CHAIRMAN SIEBER: Well, one of the
important things for us is that one of our roles is
to advise the commission or the executive director
as to the policies and the technical issues that
ought to be pursued and some prioritization and a
sense of direction, and you really can't do all that
stuff from Rockpit.
So these sessions with the regions and
with licensees are extremely valuable to us, and
that's why we want to come here from time to time,
and we consider this a very important part of our
function, and for that I offer you the thanks of the
ACRS and the members here.
And since we do have some airplanes to
catch I think -- John?
MR. LARKINS: I just wanted to thank the
administrative staff also for their outstanding
support.
CHAIRMAN SIEBER: I would remind the
members if they want to ship materials back as
opposed to using it as ballast in their suitcases
there's a box on this table. You can put your name
on it and put it in the box and it will --
guaranteed to go somewhere.
And with that, again, my thanks to the
staff in Region IV. We enjoyed our visit. It was
valuable to us. And with that I would adjourn this
meeting.
(Whereupon, at 3:15 p.m., the meeting
was adjourned.)
Page Last Reviewed/Updated Tuesday, August 16, 2016