ACNW Audit Review of Chemistry Issues for the Yucca Mountain Site Considerations Report
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: ACNW Audit Review of Chemistry Issues for
the Yucca Mountain Site Recommendation
Considerations Report
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Thursday, February 22, 2001
Work Order No.: NRC-078 Pages 1-38
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 COMMITTEE
+ + + + +
ACNW AUDIT REVIEW OF CHEMISTRY ISSUES
FOR THE YUCCA MOUNTAIN SITE RECOMMENDATION
CONSIDERATIONS REPORT
(ACNW)
+ + + + +
THURSDAY
FEBRUARY 22, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The ACNW Audit Review Committee met at the
Nuclear Regulatory Commission, Two White Flint North,
Room T2B1, 11545 Rockville Pike, at 8:30 a.m.,
Dr. Martin Steindler, Acting Chairman, presiding.
COMMITTEE MEMBERS:
DR. MARTIN STEINDLER, Acting Chairman
DR. JAMES CLARKE, Member
DR. PAUL SHEWMON, Member
. ACRS STAFF PRESENT:
DR. ANDREW C. CAMPBELL
DR. TAE AHN
DR. JOHN BRADBURY
DR. RICHARD CODELL
DR. GUSTAVO CRAGNOLINO, CNWRA
DR. BILL DAM
DR. CARL DIBELLA, NWTRB
DR. DAVID ESCHE
DR. BRET LESLIE
DR. TIM MCCARTIN
. A-G-E-N-D-A
AGENDA ITEM PAGE
Recap and Discussion of Issues . . . . . . . . . . 5
Presentation by David Esche. . . . . . . . . . . . 8
Adjournment. . . . . . . . . . . . . . . . . . . .38
. P-R-O-C-E-E-D-I-N-G-S
(8:30 a.m.)
DR. CAMPBELL: All right. The situation
is that Dr. Steindler here is going to be the Chair
for this morning's session, such as it is. All of the
group has gotten early flights out to avoid getting
caught in the storm that is eminent.
And Ray actually had to catch a flight
very early. So he had already had to go. Otherwise,
he was going to get stuck here. So, Dr. Steindler
will chair the meeting, and what we are going to do is
we are going to have to adjourn before 10 o'clock, is
that right, Gentlemen?
ACTING CHAIRMAN STEINDLER: Yes, I believe
so.
DR. CAMPBELL: And we are going to go
ahead and have Dave Esche from the NRC staff discuss
some of the aspects of the TPA code. We will be
adjourning about 10 o'clock this morning because of
the situation with the snow coming and people having
to catch flights. So with that, Marty, the floor is
yours.
ACTING CHAIRMAN STEINDLER: Thank you very
much. Unaccustomed as I am to public speaking, if I
can find my agenda, I would be in good shape. There
will be no introductory remarks by Ray Wymer.
And there will be a little bit of a
discussion as to what we learned yesterday, or what
transpired yesterday. The focus of the meeting was or
is chemistry, and specifically chemistry as it relates
to the waste package, and its role as a source term
for using and transport.
We did hear some of the issues, but
certainly not all of them, dealing with the chemical
background built into the models. We heard
considerable discussion on corrosion and were told in
no uncertain terms that the excessive use of
conservative assumptions could easily lead to a
nonsense output.
However, we do I think have to change the
role of our normal protocols, and this is not a
scientific discussion. This is a practical analysis
of what needs to be done to satisfy the Commission
through the staff, and provide them with reasonable
assurance that the models that DOE is using that are
checked by the staff are appropriate for in our case
ultimately the license application.
I think that is the focus, and our focus
is to determine whether or not the staff, the NRC
staff, is able to satisfy that role and specifically
how are they doing it.
We heard a considerable amount of
information and handed a considerable amount of
information on the methodology, the process that the
staff is using to resolve what they call issues, which
are really questions to the Department of Energy on
the source, the nature, and the implications of some
of the assumptions and activities in the models and
their abstractions.
It isn't very clear at this point whether
the staff believes that they are in satisfactory
condition, considering that they don't have a whole
lot more time between now and the time that, one,
somebody is going to ask them about the TSPA site
recommendation.
And not a whole lot of time when the TSPA
licensing application is going to come bouncing on
somebody's desk. With that, I would suggest that one
of the important issues that we are going to try and
at least address today is whether or not the staff can
produce estimates on an independent basis of the
behavior of the Yucca Mountain system as it is
described in gross terms by the Department, but as it
is described in detail models by the staff themselves.
The issue will be, one, is that doable,
and has that been done, and two, and perhaps most
important, how independent is that exercise done by
the staff. It is at the moment to us here, but
obviously not to other people, somewhat immaterial
what the answers are, unless the answers get to be
drastically different than what the corresponding
Department of Energy results are.
And then somebody needs to at least look
at the reasons for that difference, and try to unravel
those. With that, my suggestion is let's hear what we
can about the staff's efforts to model abstract and
calculate. Can we do that?
DR. ESCHE: Yes.
ACTING CHAIRMAN STEINDLER: All right.
State your name, rank, and serial number, and sit
down.
DR. MCCARTIN: Can I just add one thought?
You talk about the differences, and I think whatever
calculation the staff does, we will have to know why
we got our numbers, and why DOE got their numbers, and
whatever comparisons are appropriate.
And I will give you the best example I
have. To date, generally our calculations show
similar doses and similar release rates. DOE has a
much higher release rate, but takes a very large
credit for cladding.
We take no credit for cladding and have a
lower release rate. We get a similar answer, but for
drastically different assumptions, and I think that is
the healthy part of the process, is that we will
understand not only our own results, but DOE's.
And whatever assumptions affects ours,
versus theirs, I think -- and whether we have to walk
through them all in the licensing area is another
issue, but I think we have to understand all those.
And we get the same results, but like I said, for very
different reasons.
ACTING CHAIRMAN STEINDLER: That's quite
right. It isn't just focused on the number at the
end.
DR. ESCHE: All right. I am Dave Esche,
and I am in the Performance Assessment Section, and I
am in charge of the TPA5.0 code development. I was in
charge of the TPA4.0 code development, because Tim was
responsible for that in the past, but he got
overwhelmed with rule makings. So the activity was
shifted to me.
One of the things that we are looking at
for TPA5.0, and working pretty hard on developing, is
a revision to the gas and seepage models. And I think
this picture up on the screen is a pretty good
representation of the methodology to use to try to
evaluate that chemistry.
It is the connections and flow paths that
the DOE is using, and I think it is a pretty good
picture of how things will be moving, and in what
locations you may need to evaluate the conditions.
And our focus right now is on location one
and location two for the time being. What is the gas
and seepage chemistry coming in, and then what happens
to that once it gets on these engineered systems.
So we have a team made up of Gustavo, and
Tae Ahn, and the CLST folks, the corrosion people,
working with the near-field environment people, and
the TEF, thermal effects on flow people, and all of
those are put together to try and evaluate this
problem.
The hard thing is that you have
uncertainties coming in each area, and so what does
that mean. What we are trying to define is what is
the window of environmental conditions that we think
you may have there.
The answer may be that the engineered
system is still robust after you have defined that
window, but for right now that window is pretty ill-
defined, and so we are working to better define what
that window is.
And at location two, we have a group at
the Center that is using a code called OLE, I believe,
which is -- well, it is to predict the concentrations
of salts under very extreme conditions.
So EQ36 breaks down at a certain ionic
strength, and it is no longer applicable, but this OLE
code, I guess, is used or has been used at Hanford to
evaluate what is the chemistry of these extreme
conditions, and that's what we are trying to use at
location two to evaluate what happens to the
chemistry.
So our idea is that we are going to
propagate the differences in the chemistry and water
coming in from location one, and if we have high
temperatures, and a hot repository, then that water
coming in boils and leaves something there.
So we try to evaluate what is there, and
then based on what is there, then that determines what
happens to the chemistry from that point. There is
some key uncertainties, like what happens when you go
from completely dry to you first start wetting.
What is the chemistry that that surface is
seeing, and is it aggressive to those metallic
materials. Well, it is hopeful that the corrosion
people can get data.
Once we define what the window is, then
maybe they will be able to do some tests and evaluate
it to see if there are any corrosion problems in that
window o susceptibility.
ACTING CHAIRMAN STEINDLER: When you say
wetting, you mean gross wetting?
DR. ESCHE: Well, we have both. Like we
have considered that some locations may have seepage,
but other locations may just have an increase in
relative humidity.
So the relative humidity may be low where
it is effectively dry, but then the relative humidity
increases, and depending on the compositions of the
salts that are there on the surface, that will define
when you start having moisture present on that
surface, and at what temperature effectively.
But the key problem is propagating the
uncertainties, because you have so many permutations
that you can only do a subset of those permutations,
and how do you choose what the right subset is.
So we are doing the best we can in trying
to calculate the edges of that window, but we are not
going to be able to do every point within that window
of different chemistries.
But in performance assessment -- and I
have heard you guys talk about this, but we always
want our most realistic number, and then our range of
uncertainty. We don't like the conservative bounding
effects, because when you propagate it through 10, or
15, or 20 perimeter distributions, you start ending up
with a ridiculous result, which has been said.
So we try our best at, okay, this is what
we expect to happen, and this is our uncertainty,
rather than setting things towards the ends of
distributions.
Lauren Browning would be good to talk
about this, but she is on an audit out at Las Vegas.
She is the one who is coordinating the work down at
the Center.
And it is likely that for TPA5.0 that we
will have at least some revision to the chemistry
model, but we might not get the whole way to where we
want to be, because it is resource consuming, and it
requires cooperation between five different
disciplines basically. It is a difficult problem, a
really difficult problem.
But I think this picture that is put on
the board is a pretty good framework. You run into
all sorts of complexities though, like what happens
when you go from dry to wet, and you don't really have
data.
What happens if the salt on the surface
has some porosity to it that causes vapor pressure
lowering, or if you form an aggressive condition on
the top of the drip shield, and the drip shield fails,
what is the mass transfer of that material then to the
waste package surface.
And if you don't have dripping is there
any mass transfer, and if you have intermittent
dripping, and so we are going to try and model it with
basically equilibrium calculations at each step or at
each point in time.
And maybe that is a bad assumption, but
considering that we probably won't have much data
about the kinetics of some of these mineral phased
dissolutions, et cetera, that is probably the best
that we can do to get a rough answer at what is
happening.
DR. SHEWMON: You sort of dropped your
voice as if you are finished. Are you going to get to
location three, or is that somebody else?
DR. ESCHE: Well, as to location three,
what we imagine is that when we have an intact drip
shield, we will evaluate the chemistry at location
two.
And then if we open a gap in the drip
shield due to either general corrosion, et cetera, we
will propagate the environment at location two with
some mast transfer mechanism to location three. When
the drip shield is intact, we don't imagine that we
will have progressive chemistries at location three.
DOE has collected information on dust, and
it is going to evaluate what the chemical composition
of that dust is. So maybe the dust interacting with
relative humidity gives you some sort of environment
at location three.
But it certainly is not going to be as
aggressive as the test conditions that we have seen so
far. So maybe it is of minor concern. And maybe that
is --
ACTING CHAIRMAN STEINDLER: It looks like
your temperature is going to be higher.
DR. ESCHE: Yes, your temperature will be
higher. So at location three, it is conditional on
what happens to that barrier above it. We won't have
an aggressive chemistry at location three until we
have a hole at location two.
Or if our mechanical folks said that you
can form gaps in the drip shield due to seismicity,
then that would provide the opportunity for the
environment, either the seepage water from location
one going directly to location three, or seepage water
that lands on location two.
DR. SHEWMON: And currently the seismicity
people have decided that the seismic won't break up
the drip shield. Isn't that what I read someplace?
DR. ESCHE: Yes, DOE has calculated both
that it won't fail the drip shield and that it won't
open gaps in the drip shield, because they designed
the drip shield with, I think, a lip on each side that
kind of locks over the top of it.
So the seismic forces just aren't large
enough to pop those off and make a space. But I don't
know what we have done what the NRC feels or the
Center feels about those calculations, but that is
DOE's results.
ACTING CHAIRMAN STEINDLER: Okay. Other
than the attention to conservative estimates, how does
this differ from the TSPA?
DR. ESCHE: How does it differ from DOE's
TSPA?
ACTING CHAIRMAN STEINDLER: Yes, DOE's
TSPA.
DR. ESCHE: I think they aren't using --
they aren't evaluating -- they are modeling what
happens to the chemical conditions at location two
when you go from completely dry to start wetting.
So we are trying to better get a handle on
what is happening in that time regime, and what is
happening to the chemical condition at location two
when you go from dry to wet.
If you look at the ionic strengths that
they have in their model, the ionic strengths that
happened at location two are much lower than what
happens when you have a salt or precipitate layer on
the top of the drip shield, and you start adding a
little bit of water.
So the ionic strengths are much lower than
that. I don't know exactly what the assumptions were
in that regime, but I think we are evaluating location
two differently, and with a different code, and with
a different -- well, maybe the same geometric
framework, but maybe a little bit better consideration
of the time scale to the processes.
And whether that is -- and I don't know
about the process model, but in DOE's TSPA they used
500 year time steps. Well, if you use the 500 year
time step and you were trying to look at chemistry
that is happening at location two, you skip right over
all those processes that happen when you go from dry
to wet.
How important are those? Well, maybe they
are not important at all, and maybe some simple
testing could identify that, and I think that it is
testing that DOE is doing.
They have samples that are halfway
submerged, and sickled wet-dry, and we can model it to
death, but the data is really the good way to put to
rest some of these things.
ACTING CHAIRMAN STEINDLER: Why did you
think that was an important thing to do?
DR. ESCHE: To look at the wet-dry
conditions?
ACTING CHAIRMAN STEINDLER: Yes.
DR. ESCHE: Just because the -- well,
whether you have, say, a localized corrosion
phenomena, for our model, it is dependent on the
temperature and the ionic strength. You need high
temperatures and high ionic strength.
So what is happening when you go from dry
to wet is that you should have the maximum -- at that
point in time, it should be the maximum of both ionic
strength and temperature.
So if you have certain deliquescent
mineral phases on that surface, they will start taking
water from the atmosphere before, say, if you had
sodium chloride there.
So depending on the mineral phases that
you form will define the temperature and ionic
strength that you have. And it might be a short
period of time, and maybe it is not important, but if
the corrosion rates are fast, that period of time may
be important to try to characterize, and so that is
what we are working towards.
ACTING CHAIRMAN STEINDLER: It still isn't
clear to me whether or not in the grand scheme of
things if it is worth the effort if it looks like what
is going to take to get a better handle on what is
going on.
DR. ESCHE: I think in the grand scheme of
things that it may not change the result at all, but
at least that we asked and answered the question would
provide confidence that the -- well, right now a lot
of people have uncertainty, yourselves or the NWTRB,
that you have this package that lasts 50,000 years or
a hundred-thousand years, a million years.
And actually at a hundred-thousand years,
only .2 percent of the surface area has failed. So
you have a few hole, but generally the thing is pretty
intact.
ACTING CHAIRMAN STEINDLER: The surface
area of what?
DR. ESCHE: The surface area of the waste
package. Only .2 percent has actually failed at a
hundred-thousand years.
ACTING CHAIRMAN STEINDLER: So that is
what I am getting at. If in fact that is true, I
guess I can argue that the amount of resources that
you are going to spend, and the folks that you are
going to try and pull into this team that you have
got, may not be worth the effort.
DR. ESCHE: Well, certainly --
ACTING CHAIRMAN STEINDLER: In other
words, how did you pick it?
DR. ESCHE: How did we pick what?
ACTING CHAIRMAN STEINDLER: Why did you
pick that?
DR. ESCHE: Because the waste package is
the most risk significant thing in the whole system.
If the waste package isn't working, then that is the
only way you start getting close to doses that would
violate the standard in the regulatory time period.
ACTING CHAIRMAN STEINDLER: So you must
have had some assumption that the approach that DOE is
using is unsatisfactory or uncertain.
DR. ESCHE: Well, it is uncertain, yes.
Unsatisfactory, I would say no. They have made leaps
and bounds in that area, and they continue to work
further. But in a lot of these areas, we find -- we
don't know the right questions to ask until we look at
the problem ourselves.
And when we look at the problem ourselves,
then we say, oh, we should have been asking about
this. So that's why we do a lot of the independent
analysis that we do, because when you are reading a
document and you look at it, and you go through it,
and you may find questions.
But when you try to solve the problem
yourself, you identify a lot of different things than
when you identify it by just looking at a document.
So that's what I see the utility of it
is, and the engineered barrier system is the most risk
significant system. And even if you take out the
engineered barrier system though, the natural system
still does the time.
You can see just by some of the
neutralization analyses that DOE Did that a lot of
these things are working in combination to take that
hazard down. So it is not just the engineered system,
but in the regulatory time period, or in the 10,000
year time period, the engineered system is buying you
compliance.
DR. MCCARTIN: If I could add one point.
I think the regulatory question that we have been
pushing DOE on -- and I will say maybe about a year-
and-a-half ago that we raised the question, and they
said we will have no waste package failures in 10,000
years.
They have made that statement, and we can
defend it, which is fine. The question we asked is
have you considered an appropriate range of conditions
that would affect the waste package in a deleterious
way, and they said we believe we have.
And that is where the staff is now, okay,
are we certain, and we are just pushing to make
certain they consider an appropriate range of
conditions for that 10,000 life time.
And then not too long afterwards, you guys
are aware of the State coming in with the trace
metals. And I think it is just the ZLST people at the
center and NRC, and the near-field people, let's get
together and make sure how we found the conditions
that we think would capture an appropriate range for
defending that 10,000 year lifetime.
ACTING CHAIRMAN STEINDLER: Okay.
DR. CAMPBELL: Just for the record, that
was Tim McCartin from the NRC staff, and Dave Esche is
the speaker at the table of the NRC staff.
Given, Dave, what you said about the 0.2
percent failure at a hundred-thousand years, and the
doses associated with that, I think that emphasizes
the importance of establishing with some high degree
of certainty that in terms of reasonable assurance
that the 10,000 plus lifetime for the waste packages
is defendable in a hearing type of process.
And not only on every technical aspect of
it, because 0.2 percent is a pretty small segment of
-- I mean, there are a lot of waste packages in there,
and that is still a lot of surface area. But when you
start thinking about failure rates of normal stuff
that people have in their every day lives, you are
looking at 5 percent being an acceptable failure rate.
ACTING CHAIRMAN STEINDLER: Are you
talking about washing machines?
DR. CAMPBELL: Right.
(Laughter.)
DR. CAMPBELL: But you see what I am
saying, is that is what most people are dealing with
in their normal lives; is that 5 percent of the
products that I get are a lemon, and they are worried
about that.
And then you say, okay, now 2/10s of a
percent at a hundred-thousand years, and we are
getting doses that far exceed the limit, even though
we are way out in time, a lot of people are gong to
say, hey, it doesn't take much failure earlier on to
get you into trouble in terms of compliance.
DR. ESCHE: That was -- the .2 percent is
surface area failed, and the actual percent failed,
the ones that get very small holes, or cracks, I think
it is like 50 percent have cracks in them, and 26
percent have patches. But it is only .2 percent of
the area.
DR. CAMPBELL: Right.
DR. ESCHE: It is kind of number smithing
or something, but in general the amount of surface
area failed is very small at a hundred-thousand years.
DR. SHEWMON: Failure here is significant
corrosion or penetration?
DR. ESCHE: Well, the larger numbers are
penetration.
DR. SHEWMON: No, the .2 you said was --
DR. ESCHE: The .2 is a 6-inch-by-6-inch
patch, I think, is roughly the size if a hole failed
that big.
DR. SHEWMON: So that is penetration.
DR. ESCHE: So it is an opening, yes.
DR. SHEWMON: Thank you.
DR. CAMPBELL: But isn't that an
assumption of the size of that? It assumes --
DR. ESCHE: It is just the size selected
to model the problem.
DR. CAMPBELL: Based on general corrosion
rates?
DR. ESCHE: Yes.
DR. MCCARTIN: But DOE gets releases from
cracks, no matter how small, or early, or how much
water.
DR. ESCHE: And the other thing you have
to remember is that if they have a patch that fails
and let's say a 6-inch-by-6-inch patch, they have
diffusive releases over that whole area. So that
whole 36 square inches or whatever.
In reality, you are going to have a water
film around the outside of that hole, which is the
diffusive area. So they might be greatly
overestimating diffusive releases which are making
those numbers a hundred-thousand years much larger
than they may be in reality.
So some of those conservatisms you have to
keep in mind when you are looking at what the curves
are telling you at the later times, and I think they
are working on evaluating that, and you may see a
revision to that in the future.
DR. CODELL: Dick Codell, NRC. And in
fact almost across the board the DOE puts this
diffusion model in, and it seems wherever they have
that that they have exaggerated it.
For example, they assume that the waste
package is sitting directly on the inverts, and so
there would be a direct pathway, and without taking
any credit for the fact that it is sitting on a metal
stand right above the invert.
And that like the example that I gave
yesterday where the waste package just failed and that
it would lead to the maximum diffusion. I mean, it is
not our job to tell them to be less conservative, but
it just struck us all that way. That the diffusion
models in their TSPA are exaggerated.
DR. ESCHE: I guess the only thing we were
worried about was the conservatism, as if it is
masking what we should be worrying about. So like
maybe the amount of advective flow is really what we
should be concerned about, and the uncertainties
associated with that.
But if we have a conservative diffusive
model clouding that, then we might not be focusing as
much attention on what are the uncertainties in the
seepage model, you know.
So that's where we -- at least in PA
space, we try to communicate what we think and that
you may need to look at your conservatism here.
I mean, when uncertainties are large, in
some instances you have to be conservative. But you
have to be careful in risk assessment that you are not
doing that all over the place, and that you aren't
generating some goofy number, I guess.
DR. CAMPBELL: Well, we did talk about
this issue of what appeared to be -- and I will use
the phrase "ultraconservatism" in the way that they
are implementing diffusion in their models.
And not only setting a boundary condition
at zero concentration, and never altering that, even
though obviously if a species is diffusing along that
gradient, then the gradient is going to be attenuated
with time.
DR. ESCHE: Sure.
DR. CAMPBELL: Otherwise, things would
diffuse away from everything, and we would all be in
some sort of equilibrium state that doesn't exist in
the real world.
But that even raises even more questions
about the value of these neutralization analyses,
because the way that they implement neutralization of
the waste package is to impose a 300 centimeter square
patch open.
And if you are saying that then have 300
square centimeters of diffusion in an area, and you
multiple that by the 8,000 or so waste packages in
there, you can see where these high doses come out on
the calculation.
But it has no relationship to a real world
type of situation of failure, and that creates a lot
of issues, I think. That maybe it isn't the place of
the NRC to comment about the conservatism built into
that, but it sure can mask a lot of other -- you know,
as you say, potentially important issues, because that
just dominates all the release.
DR. MCCARTIN: And you don't need water.
You assume there are monolayers of water all the time
and that's the fusional --
DR. CAMPBELL: Right. Well, you need
water, but all you need is a few monolayers of it.
DR. MCCARTIN: Well, they make no
calculation with respect to water, in terms of how
much is there, et cetera, especially for the
neutralization that it has failed at T zero. That is
pretty hot.
And in terms of perimeters, for the
diffusion coefficient for source term in the waste
package is 10 orders of magnitude larger than the
diffusion coefficient they use in the unsaturated
zone; 10 orders of magnitude larger.
DR. CAMPBELL: Are they using pure water
diffusion?
DR. MCCARTIN: Oh, yes, for the source.
But in the unsaturated zone and the matrix diffusion,
they have a number that is 10 orders of magnitude
less. And that's fine, as it certainly can't be any
higher.
DR. CAMPBELL: No, it's not fine.
DR. MCCARTIN: Well, it is their
calculation, and they need to put forward what they
think they can defend and support. And as long as we
know what they are doing, we can evaluate. And it is
not our calculation.
DR. ESCHE: Usually when you are doing
something conservative, there was a reason for it.
But in some instances maybe it is a little too far,
you know.
DR. CAMPBELL: Well, I guess what disturbs
me is that diffusional ungradience is not a new
phenomena, and it is 19th Century science. This is
not a new phenomena, and for DOE to take the position
that this stuff is too uncertain to deal with in a
more realistic fashion I find incredible.
DR. MCCARTIN: Sure. And the only thing
-- and Dick may remember this better than I -- and Tae
-- but there are some experiments where DOE has some
information that the experiments were flawed, but I
thought they suspended the spent fuel particle, and
they got some concentration above a pool of water, and
they got some concentrations in the water.
And you back out some numbers, and we have
measured this, and we can't say that something isn't
going on here. And this is about 4 or 5 years ago.
DR. CODELL: Those were the experiments at
Oregon on spent fuel.
DR. MCCARTIN: And Dave is absolutely
right. We look at this and we want to understand this
assumption of what it does to the calculation, because
it can cloud some of your results.
But DOE has an experiment or two out there
that tends to support some larger releases, and it
could be their concern with that experimental evidence
that we will not be able to refute it.
DR. CAMPBELL: Is this documented in an
AMR somewhere?
DR. AHN: Well, originally in the early
'80s, Northwest Laboratory, supported by DOE, spent a
long time on spent fuel testing. In other words, in
emerged air or simulated ground water, and
periodically measured the dissolution rate over a long
period of time, from 5 to 7 years.
Then all of a sudden a new testing was
initiated in Livermore, and they attempted to study a
fundamental using only a single carbonated solution,
excluding all complicated species. The test method
was different and it was a so-called flaw testing to
determine the dissolution rate.
That rate was very, very high compared
with the emerging test results, which came from real
well water conditions. We hoped that they could
incorporate that in other species later on, but they
never did it.
They kept going on with the single
carbonate solution, because they wanted to understand
the very basic mechanism. Later on we questioned them
on why do you need to do that, and they said, well,
this testing is conservative. That was the only
reason.
ACTING CHAIRMAN STEINDLER: Okay.
Gustavo.
DR. CRAGNOLINO: Well, I know that it is
important about what Dave Esche mentioned and
understanding the chemistry and the drip shield, and
the waste package and the corrosion, and that the drip
shield was conceived by the DOE as a way to control
the flow of water, and this was a very strong
statement in number two.
And from the point of the engineered
barrier system, the two principal factors were the
performance of the waste package and the performance
of the drip shield. However, if you look in the
recent recital, and condition three, it is not the
drip shield alone.
Now it is the drip shield drift invert
system. Why? Because apparently the drip shield has
not played the significant role that was expected by
using the dose after 60,000 years. And I think this
is an important modification, and is something that we
have to explore in more detail.
It changed the emphasis, and the emphasis
is now more in the relief and leaving aside
performance of the waste package. This is where we
really have to refine our approach to the knowledge in
the same way that David has mentioned.
DR. ESCHE: Well, the way I look at the
drip shield is that it may not show up in the current
analyses as being extremely important because of the
large diffusional releases, and it is mainly
preventing water flow.
But it also minimizes rock fall and
aggressive chemistries if you have it. So like say
the titanium is only susceptible to fluoride, for
instance.
Well, the time that the drip shield lasts,
which is on the order of 10,000 years in our model,
and 20 to 40,000 years in DOE's model, I think, you
would expect that most of that aggressive chemistry
has been rinsed out of the system so to speak by the
time of the drip shield failure.
ACTING CHAIRMAN STEINDLER: Why would you
expect that?
DR. ESCHE: Well, in DOE's model, what
happens is when you have seepage, it is just as a
mixing cell. So the seepage comes in and mixes with
the salt and carries some of it out, and that is
enough to dilute the chemistry back to ambient pretty
quickly.
ACTING CHAIRMAN STEINDLER: But ambient
fluoride, which is the issue.
DR. ESCHE: Well, the ambient fluoride, I
think we heard, was 4 milligrams per liter. And from
what our people told me, the corrosion of the drip
shield requires a threshold fluoride concentration,
but the fluoride is consumed in the reaction.
So you need to get a higher fluoride
concentration, but then you need a certain mass flow
of fluoride into the system for the corrosion, because
it is consumed in the reaction.
But if the drip shield is preventing all
these chloride and other salts that are formed there
initially, those salts are probably rinsed out either
due to seepage water in, or relative humidity becoming
high again.
And by the time the drip shield fails the
aggressive chemistry is prevented on the waste
package. So we forget that whenever we are doing the
analysis, too, because right now the data suggests
that even if you have those aggressive chemistries,
the waste package only in DOE's model has general
corrosion and stress corrosion cracking.
Localized corrosion never happens for any
of the conditions that are generated, and so that is
the issue. If you have a different window and those
conditions were generated, would you have localized
corrosion, and that's what we are trying to find out.
And it would build credence to, oh, yeah,
the drip shield doesn't show up in your model as doing
much now, but there is maybe a good reason for why it
is there.
ACTING CHAIRMAN STEINDLER: Okay. Any
other comments or questions?
DR. MCCARTIN: One quick correction. It
is not a 10 order of magnitude. It is only a 2 order
of magnitude. I did the unit conversion in my head
wrong.
ACTING CHAIRMAN STEINDLER: So two orders
of magnitude.
DR. MCCARTIN: Two orders of magnitude.
ACTING CHAIRMAN STEINDLER: That's still
not bad.
DR. ESCHE: And my comments about what DOE
is doing is my understanding based on reading the SR.
So I may not be accurate in everything that I said,
but I hope that I am.
ACTING CHAIRMAN STEINDLER: All right.
Well, this chair is going to turn into a pumpkin not
too long from now because you folks have such
unfriendly weather.
DR. CAMPBELL: As opposed to the friendly
weather in Chicago.
(Laughter.)
ACTING CHAIRMAN STEINDLER: Well, we don't
shut the place down.
DR. CAMPBELL: John.
DR. BRADBURY: John Bradbury. We have
been talking about the ultraconservative and the
problems with that approach. I would like to pose a
possibility that there is a situation where DOE has
been nonconservative, and that relates to what was
raised yesterday concerning the use of hydrochemistry
in the saturated zone to delineate flow paths.
Again, to remind you that the line from
Yucca Mountain down to the critical group is at
constant chloride chemistry, chloride being a non-
conservative tracer similar to technetium and iodine.
Essentially this is saying that there is
no dilution along the saturated zone flow path. Also,
one should consider the evidence that perched water in
the UZ is similar to perched water in the saturated
zone.
So one could also think that there is no
dilution on the whole flow path for these conservative
species or elements. Now, why I am saying this
evidence hasn't been used that way is because one of
the key attributes is delay and dilution of
radionuclide concentrations provided by natural
barriers.
DOE's modeling assumes or part of the
model involves movement through the UZ and then there
is a separate leg along the SZ, and there is a
technique, a convolution technique, to essentially
allow source material from the UZ to be added to the
SZ.
So it is like throwing sticks on to
flowing streams. Now, the flowing streams flow at
different rates. So the faster flowing stream, if you
throw it in at the same rate, dilutes the
radionuclide. And that is what DOE models.
That is not what this type of chemical
evidence would suggest. So here is evidence that may
suggest that DOE's model is non-conservative.
DR. CLARKE: If I understand you, John,
you are saying that you can't have it both ways. Do
you have an opinion which way it should be?
DR. BRADBURY: What I want to make sure is
that all evidence is used to the point where you have
got alternative models being considered.
DR. CLARKE: Well, I guess my question may
be better phrased by saying do you have a problem with
the flow paths as they have been delineated?
DR. BRADBURY: Well, the flow paths, the
hydrologists say that those flow paths -- and
hydrochemistry aside, those flow paths are reasonable.
I take that evidence and say that seems
reasonable, and then when you couple that with the
hydrochemical evidence, you come up with a situation
where you say, well, it looks like hydrochemical
evidence, which is evidence that is not just short
term, but evidence that has been developed over
thousands of years, it is really kind of very good
evidence to say, yes, no dilution occurs.
And therefore I am seeing this discrepancy
here with regard to the evidence.
ACTING CHAIRMAN STEINDLER: Okay. Any
additional comments? If not, then my intention at the
moment is to suggest we all go home. What else do we
need to do here?
DR. CAMPBELL: I think the path forward
for this group is that all of us agreed in our
discussions yesterday among ourselves that we are
going to look harder at the IRSR and do a cross-walk
with some of the individual issues that we have
identified in our own look-see at the various DOE
documents and models of TSPA.
And that basically the group will be
writing a report to the main committee for its
consideration as a possible letter to the Commission.
So that is basically the path forward for us.
This meeting is not the end of our work,
but rather a halfway point, in terms of getting a
better handle on some of the issues and interacting
with people here. But we have some more work ahead of
us.
DR. MCCARTIN: I was going to offer that
Dick Codell will provide you with the paper that talks
about this experimental evidence that I think at least
at one time DOE was looking at and suggesting a high
release rate.
ACTING CHAIRMAN STEINDLER: All right.
Well, thank you very much for all of your work, and I
hope you all get home for those of you who have
someplace to go. And we will call the meeting
adjourned.
(Whereupon, the meeting was concluded at
9:27 a.m.)
Page Last Reviewed/Updated Monday, October 02, 2017
Page Last Reviewed/Updated Monday, October 02, 2017