110th ACNW Meeting U.S. Nuclear Regulatory Commission, June 29, 1999
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
ADVISORY COMMITTEE ON NUCLEAR WASTE
MEETING: 110TH ADVISORY COMMITTEE ON
NUCLEAR WASTE (ACNW)
Southwest Research Center
6220 Culebra Road
San Antonio, Texas
Tuesday, June 29, 1999
The Committee met, pursuant to notice, at 8:30 a.m.
B. JOHN GARRICK, ACNW Chairman
GEORGE HORNBERGER, ACNW Vice Chairman
RAYMOND WYMER, ACNW Member
CHARLES FAIRHURST, ACNW Member
. P R O C E E D I N G S
MR. GARRICK: Good morning. The meeting will now come to
order. This is the second day of the 110th meeting of the Advisory
Committee on Nuclear Waste.
My name is John Garrick, Chairman of the ACNW. Other
members of the committee include George Hornberger, Ray Wymer, and
This meeting will be open to the public.
Today the committee will continue its review of center
activities and also later on today we're going to be touring several of
the laboratories of the institute.
Howard Larson is the Designated Federal Official for the
initial portion of today's meeting.
This meeting is being conducted in accordance with the
provisions of the Federal Advisory Committee Act. I don't think we've
received any written statements or requests to make oral statements from
members of the public regarding today's session. However, should anyone
wish to address the committee, please make your wishes known to one of
the committee's staff.
As usually, it is requested that each speaker use a
microphone, introduce themselves, and speak with sufficient clarity and
volume so that they can be readily heard.
Today we're going to talk about design issues, for the most
part, and the committee member that's going to -- the committee members
that will be leading these discussions, for the most part, will be
Charles Fairhurst for the first presentation and George Hornberger for
the second presentation.
So with that, I guess our first speaker is Raj and our
second one will be Simon Hsiung. So with that, we will proceed. Raj,
MR. NATARAJA: Good morning. The presentation this morning
will consist of three parts; a short introductory part that I will be
doing, and a more detailed presentation by Dr. Hsiung, Simon Hsiung,
from the center, at your end, and we'll try and finish this presentation
in about half the time that is allotted to us, so that there will be
enough time for discussion at the end.
My name is Simon Nataraja. I am a Senior Engineer, working
in the engineering section, and I am also the group manager for this
particular KTI, repository design and thermal-mechanical effects. And
Dr. Hasad Choli is my counterpart at the center. He is the group
manager for this KTI. And Simon is the main presenter. However, he is
being supported by several others; Ray Chen, Mr. Ghosh, you probably
have seen him and probably have talked to him before. They are all rock
mechanics experts, experienced in design and modeling.
And a new addition to this group is Doug Gute, who is a
mechanical engineer, and we are looking forward to his participation in
the pre-closure aspects of this particular KTI, which have been under
development for a while now.
In fact, I used to be a single man team and thanks to the
new organization, our budget was doubled. And another person working on
this end is Bernard, and since we have two people now, he is able to
attend the NWPR review and, at the same time, attend this meeting here.
The outline of this morning's presentation includes a
statement of the issue, and we will also go into statements of
sub-issues within this main issue, and we will talk about pre-closure
design aspects of the repository. Then we will talk about our
understanding of -- our limited understanding of integrated safety
analysis. As you know, Part 63 requires that DOE perform an ISA to
demonstrate that the pre-closure performance objectives have been met.
We are going to start on a fast-track to learn, this group
at least, because there is no experience applying ISA to repository
design, either at the center or on the staff here.
Then we will touch on some of the post-closure aspects of
thermal-mechanical effects on the repository performance, and we will
list a few insights that we have developed by our exercises of providing
input to the system performance assessment.
And we will discuss how we are reviewing the fast-changing
repository designs, the various options being considered by the
Department of Energy, and we're looking at the analysis using
engineering models, how we are developing capabilities to be able to
review their designs, and we will summarize the progress that we have
made in Rev. 0 and Rev. 1, which you probably have there, and we'll
discuss the path forward for resolving the many issues before license
You have all received this, I believe, in Budhi's
presentation yesterday. Our particular KTI, the main contribution comes
in the box under the engineered barriers, specifically EBS-2, which is
the mechanical disruption of waste packages.
We have done some preliminary work using the rock fall model and the
effect of rock fall in repository performance, and we also are
continuing to review those models.
We also have some contribution to the other boxes, EBS-1,
perhaps even four, but mostly in a secondary fashion, not in a primary
fashion, and we have some contributions to make under GS-1, which is the
spatial and temporal distribution of flow.
During Simon's presentation, you will hear some details
about the changes of geometry due to excavation, and the effects on flow
and so forth.
This particular KTI, the main issue statement is the
adequacy of design, construction and operation of the geologic
repository operations area, GROA for short, to meet both pre-closure and
post-closure performance objectives, taking into account long-term
thermal-mechanical processes. We had another portion which we dropped
because of the Part 63 changes. We will go through some of those
The sub-issues that I have listed here, you may not find all
of these in the current versions IRSR, the zero and one, because the
definition of KTI was focusing mostly on the post-closure and because of
the descriptions, we had no application to work on the pre-closure
The next revision, we hope to revise it and include some of
the sub-issues with pre-closure.
The first three items that I have listed there under
pre-closure sub-issues, mainly effectiveness of design control process,
design of seismic events, direct fault disruption, and
thermal-mechanical effects, all those three items affect both pre and
post-closure attributes of the repository design.
The last item there, the item dealing with the pre-closure
performance objectives, those are next under the two design basis
events. And the design meeting the options -- the ability to retrieve
and the ability of the design to accommodate the performance
confirmation program, all those three aspects are pre-closure issues, as
far as design is concerned, and also the implications for post-closure
performance confirmation program.
And as far as the post-closure issues are concerned, it is
the effects of thermal-mechanical events, including seismic, on
post-closure performance. And as I mentioned, we dropped the issue that
used to be part of the post-closure sub-issues.
With this brief introduction, I will get into some details
and I'd request Simon to pick up from here.
MR. HSIUNG: Raj, you can take it off now. Good morning.
My name is Simon Hsiung. My background is mining engineering. I'm
going to give this second part of the presentation. I'm going to talk
about the pre-closure related issues first, and then touch upon some of
the mechanical effects on post-closure performance, then discuss a
little bit about the path forward.
For the repository design, in the past, like Raj alluded to,
we put our main focus essentially in the underground facilities. But the
actuality is that the repository design includes two major parts; the
first one is the surface facilities and the second one is subsurface
For the surface facility, we have a -- the consensus is that
we have many experiences in the past within the agency in terms of
design review, all the different facilities like reactors and temporary
storage. Also, in the past two years, the center has been tasked to do
some design review for the interim storage by private fuel storage
facilities and essential interim storage facilities and even the
So there is some review right now in the licensing process.
So in that sense, we can say that we have a considerable amount of
experience in this aspect.
So in the design review of surface facilities for the high
level waste disposal, and we draw on those experiences and apply them
directly into our program at this time.
But the only thing right now that we don't have is that Part
63 has specifically required DOE to conduct the integrated safety
analysis, called ISA, and to use that to demonstrate the compliance of
the pre-closure performance objectives.
That's relatively new. I don't think that it is my opinion
that DOE, from what I see, has the experience that they apply in these
type of techniques in the regulatory framework.
So there should be some amount of learning curve that would
be applied for us to gather the understanding in terms of what's the
exactly that ISA, that they include in it, and how can go about to
develop the review plan to actually do a meaningful review in terms of
DOE's license application.
Then for the subsurface facilities, there is also a bundle
of experience in the design of the tunnels, in the power stations, and
even in the mining access. But what makes this program unique is that
the underground facility, like, say, in the emplacement area,
emplacement drifts, that they will be put under tremendous temperature
conditions. The temperature will have, because of the decay of the
waste, will gradually build up the temperature and that will, in turn,
cause stress buildup in the emplacement drift area.
By doing that, also, the other difference is the long-term
surfaces that will be required for this emplacement drifts and the
current thinking is the design will be out for a hundred years, with
possible extension to 150 years or so.
So because of the relatively high temperature and also the
high stress induced by the thermal load, then the two phenomena that we
need to take into consideration, which perhaps other facilities may not
have to consider too much about this. That is, the mountain degradation
of the rock mass, the properties and the strengths, because of the
intensive temperature and also the stresses associated with that.
And the experiments have shown that for the rock under a
tremendous stress condition, they undergo so-called stress fatigue and
gradually the strength will go down.
The other concern that is associated with that is under this
high temperature, what will be the longevity in the service life and
effectiveness on the ground support system, which will be required to
maintain this relatively stable condition of underground emplacement
And DOE recognizes the complexity of the design for a
repository underground facility. So it proposed a dual approach for the
underground facility design. The approach, including two methods, one
is that they're using an empirical method, which directly links the rock
mass quality to the support system that will be required, and there is
some sort of empirical type of relationship being built up in the mining
industry that it can be used in here.
DOE also proposed using the design by using the numerical
analysis. So they're trying to compliment the two methods with each
other, so build a better confidence in terms of the design efficiency.
And in the design, using the numerical approach, there are
two methods that can be applied, actually used. One is the continuum,
also like Raj alluded to earlier, and also the other one is using the
discontinuum analysis, that considers the presence of the fractures,
For the continuum analysis, there is a considerable amount
of data that will be required that will be regarding to the rock mass
properties. So the rock mass strength properties that will be required
to be able to determine the behavior of the rock mass. Of course, there
are some properties related to the rock mass properties that will be
needed, as well, in this context, also.
Now, to determine the rock mass property is a very difficult
process actually and in the literature, they have developed some
relationship between the rock mass rating and also rock mass modules.
That's a sort of empirical equation that has been developed. The figure
on the right, showing in the horizontal axis, the rock mass modules, and
the horizontal axis is showing rock mass rating, and a rock mass rating
is sort of subjective type of index value to show the relativeness
between the rock mass conditions.
And in this sense here, the relative value, we're showing
that the rock is relatively stronger and there was a smaller value,
showing the rock is either softer or may be in a poor condition.
The curve here presented is available in the literature.
You can see the rock mass modules increase rapidly or in the
accelerating rate with the rock mass rating.
Now, for DOE, just using this data and encompass the
uncertainty associated with this particular curve, they will find a sort
of behavior of the rock mass. But DOE has done some amount of data
collection in terms of rock mass modules, what they found, saying that
there is a smaller amount of data in this region, showing DOE's
measurement. It looks like substantially smaller than the one being
presented and available in the literature.
And this point here is the intact rock mass, intact rock,
the modules, and they still have a debate as to how can we put in the
intact rock modules into this rock mass classification system.
I guess what we can do is assume everything is at the
highest value. That's what DOE did in this case. But I suspect that
that may not be a good way to do that.
MR. FAIRHURST: Simon, what you've got there is precisely
the problem that you mentioned earlier, that there has been little or no
experience of high temperature situations, and you've got ambient rough
temperatures. I think those low modulus values are what DOE has
measured at ambient temperatures. In the heater drift experiment,
they're showing that a the fractures close, then the modulus does go up.
So what you're going to have to do is put in a
temperature-dependent modulus, right?
MR. HSIUNG: Right.
MR. FAIRHURST: That's precisely the problem, is that all
the data that exists in the literature is for very different situations
than we're seeing.
MR. HSIUNG: Right.
MR. FAIRHURST: I don't think the Yucca Mountain thing is --
it's not that big of a real problem. It just means you have to deal
with it correctly.
MR. HSIUNG: Right. I think that's essentially the issue.
Now, if we don't use the original data set, the risk data set, then the
result come up with is sort of irrelevant to the design.
Now, another problem is to determine the rock strength
properties. Now, in the literature, the available equation being often
used is the Hook Bronze criteria. So they use it to determine the
cohesion of weakening angle. I can see the smiles.
MR. FAIRHURST: I'm smiling because, as Lynn Deering knows,
that when Dr. Hook gave the keynote address a couple weeks ago, he said
he was flabbergasted at how people were using that criterion and that he
would like to withdraw it from the literature.
MR. HSIUNG: Okay. That's the first time I've heard this
one. But that's some of the problem that I'd like to talk about in
here, although it's not in this viewgraph at this time, is that when we
-- we're doing that. We actually extract the strength property out and
they're using this empirical equation that's available right now. We
found out, okay, the study is showing that if -- for the -- a so-called
good quality rock, actually we're experiencing more failure using this
combination of approaches to get at rock mass modules and also the
So I don't know that something was happening there, because
in the past, when I joined, before I joined the center, I built a lot of
underground designs for the coal mines and roof supports, everything
else, and deal with a different type of rocks, and we also considered
The only thing different is that we did not use the Hook
Bronze equation and we found out that even though that the stronger rock
has induced stresses, but it is often more stable.
So this is just to reverse the trend. The thermal loading
has caused a different behavior in this context, or there is certain
incompatibility between the rock mass determination and the rock stress
So those are the things that will need to be sorted out at
some point in time by DOE, because the DOE is committed in its seismic
design methodology, saying that they're going to verify all those
equations. At this time, they haven't done it yet.
And, also, that in dealing with this repository underground
facility design, there are several issues that will be need to be --
that's unique for this program that will be need to be taken into
One is the possible effect of lithophysae that will be
present in 75 to 80 percent of the emplacement area. So it's quite big
and in the range maybe as large as two centimeters or even larger in the
So the presence of that, that will certainly affect the
behavior of the rock mass itself. I think Dr. Fairhurst did a simple
calculation and assuming a rock block with a circular opening in there,
then he found out that the presence of that can reduce the rock strength
to 15 percent. Is that the correct number that I've quoted?
But the problem is more complicated than that, because that
the -- in those type of rock condition, is the elliptical type, with the
longer axis parallel to the horizontal axis.
Now, if you have a number of these type of Y-axis stacking
with each other, one of the things that I can think of is that in the
underground mining, that's a similar situation, that in the coal mine,
there are several layers of seams that need to be mined, there is a
thickness of intra -- what we call inadvertent, in between the seams,
and when you go and mine the minerals, that's similar to this situation.
And the stress of the situation is quite different in some
areas, depending upon the relative location of those voids.
In some areas, you will experience extensive shearing
behavior and in some areas, actually, there is tremendous tension. So
those behaviors -- that will affect certainly the behavior in terms of
these analysis, the whole thing, the rock mass, that those type of
impacts should also be included in the analysis to provide a reasonable
The point I'd like to mention here is the -- as I mentioned
earlier, is the degradation of the rock mass because of the extent of --
long time under stress, and 100 years surface life, and also another
thing that needs to be addressed is the longevity of the support system
and the effectiveness of the support system under that kind of
condition, as well.
Some of the concerns that I raise for the continuum analysis
that should also apply here, like the effect of lithophysae and also
like the effect of degradation of rock mass should be also included in
this continuum analysis.
And for this case here, other than the inter-rock properties
that they would need to have that to simulate rock flux and we also
would need the fracture characteristics and the fracture properties.
DOE has done quite a bit of the underground mapping to
determine the depth and maybe the length of the cracks. So they have
bundled data in that area.
And other properties that will be needed for the numerical
modeling is the contact stiffness and also the contact strength, the
joint strength, those things that will be needed. So far, DOE doesn't
have a coherent program to collect those information. The available
information they're using is that the single shared task that we did
about eight or nine years ago, they just brought all those information
in and they are doing their own analysis at this time.
Even with that -- we know enough information in terms of
joint and if we do not apply it properly, that may raise some problem.
In this case, I'd like to show a case study here, is that the figure on
the right is the so-called irregular joint patterns, which means that
the model right now, the elaboration of the joint patterns in this case,
and this approach is being used by DOE and even right now, and I also
I've asked in the past, trying to understand the behavior with this type
Okay. The outcome of this one is that the resulting
situation is relatively more stable compared to the one on the right.
That's what we call the irregular joint pattern. Actually, we just add
a little amount of variation in terms of the joint in there. You can
see tremendous differences in responses in here.
What this one implies is that we're using the approach to
the left of the -- the regular pattern will require a different set of
one support system to stabilize the ground. But looking at the parts on
the right and the support system that will be required will be
So it doesn't matter which we use, the two things that I
think that's vitally important is that we need to use representative
data to do the analysis and a different type of analysis that will be
used and the other one is using the proper phenomenon that supports
involving that particular situation.
MR. FAIRHURST: Simon, just a comment on this. Again, these
analyses were done under ambient stress conditions. They were not --
there was no temperature involved.
MR. HSIUNG: No, for these two --
MR. FAIRHURST: You will be up to a high temperature well
into the pre-closure period.
MR. HSIUNG: Right. From these two, we did it assuming
that's the 100 years of thermal heating and also that we have the
seismicity also put in there for these two set of analysis that we have
And as of now, for the lack of representative data, that we
have been talking with DOE, I think that Dr. Fairhurst attended one of
the Appendix 7 meetings and talking about that particular item and we're
encouraging that DOE at least engage in some sort of dialogue with us in
trying to come up with a reasonable program to generate additional small
amount of data into that, so we can develop data and more site-specific
equations that we can use for the design purposes.
Now I'd like to shift the gear in talking about integrated
safety analysis, because that's one of the important parts of our
activity maybe for the next few years.
Because of this new Part 60 requirement, it requires
specifically for DOE to conduct this analysis, to demonstrate
So what is the ISA? ISA, the definition that I extracted is
from the NUREG, one of the NUREGs prepared by NRC. I forgot the NUREG
number at this time. It is a systematic examination of a facility's
processes, equipment, structures, and personnel activities to ensure
that all relevant hazards that could result in unacceptable consequences
have been adequately evaluated and appropriate protective measures have
By looking at this definition, that is essentially telling
us that if we apply this approach, it will give us a risk-informed
approach that will be relevant to the regulation of Part 63.
Also, this one sort of link the composition all the -- even
including the systems, equipment that will be required to do the
operation and also breaking down the flowchart of the activities
associated with a certain function. That's a combined analysis of
looking at that. So this is definitely a systematic systems engineering
approach in doing this type of analysis.
So in conducting an integrated safety analysis, there are
two major steps. The first step is that given the processes, the
equipment that will be required and the systems that will be housing the
particular function, we can do a systematic identification of potential
hazards and accident sequences based on the information being provided
And once that the potential hazards have been identified, we
can use it to actually do the estimation of the consequences related to
So given these consequences, then we can -- the next step is
to identify the safety systems and the controls that can be used to
prevent certain accident or mitigate the consequences to bring the risk
level down to meet the regulatory requirement.
The last step is a very important step, to identify the
conflict in between the safety controls being identified.
For example, if there is some sort of release or cause a
fire hazard, some of the immediate control come into mind, for the fire
control, will be the water and the sprinkler system. But if, in the
neighboring area, that you have some sort of a process in the nuclear
waste, then that will increase the likelihood because of the water to
the criticality concern.
So that's the things that by going through this process,
we'll be able to identify those conflicts. So we'll make the overall
process more reliable and safer.
And the benefit for conducting this ISA now, in essence, is
that we've thoroughly identified the hazards and we evaluate the
consequences, and that will provide reasonable assurance that the
overall risk that is acceptable -- you know, in this sense, it will be
the pre-closure performance objective for Part 60.
And another benefit in terms of using this analysis is that
it will streamline the design. Our concern is that all the safety
system has to be designed very rigorously so they can function the way
it's supposed to. And by going through this analysis to identify the
functions, we can focus priority and including those type of design, and
also it's possible, during the process of ISA, we may identify some of
the design originally been put before us is not adequate.
So they need to go back to the drawing board and redesign
it, or maybe that the certain design in the field, with certain function
across that, and so it will require a changing of design. So that's
another benefit coming out from the ISA analysis.
So this whole process is a really iterative process. So by
doing this rigorous integrated safety analysis, that we'll understand
the risk better, so we'll increase the confidence for management to
manage the risk in the facility, because of that. They know what will
be the consequences of what could happen, so they will put more effort
and attention to those areas.
Also, that because of this comprehensive analysis and the
extensive documentation, provide a rational as to why that those things
are being done, being conducted, it will provide a more better assurance
to the general public to accept the design.
MR. HSIUNG: Yes, sir.
MR. NATARAJA: I think you will probably have to speed up
your presentation considerably.
MR. HSIUNG: Okay. Thank you. Now, one other thing I'd
like to point out is that conduct a very good integrated safety analysis
is not enough. One thing is that it cannot ensure an effective design
and it cannot ensure effective implementation of the control functions.
So those we would like for other elements of the program to
do that and one way to do that is that we make sure that they have a
very effective design control process, and that will require the DOE or
the licensee to provide reasonable basis as to how the design basis and
the design criteria are being developed based on the intended function
of particular systems and based on the performance objective, and, also,
will ensure that the design being done using the proper standards.
Another aspect is that we need to have proper training in
operations, so we can eliminate some amount of human error.
The next topic that I'd like to briefly touch upon is the
thermal-mechanical effects on post-closure performance. The way we see
it right now is that there are two aspects of the thermal-mechanical
effects can have to the overall performance.
One is seismically-induced rock fall in the thermal environment. This
one Raj has mentioned has something do with -- is part of EBS-2. The
second effect will be the flow into the drift, and that comes from
geometry change that could affect the seepage into the emplacement
drift. That will be related to EBS-3 and GS-1 in terms of integrated
I will touch upon each one of them briefly in the next few
In order to deal with the mechanical disruption of waste packages, we
have developed a very simple rock fall model. It is essentially that
assuming that the waste package is a beam, elastic beam, a simple
supported beam with elastic supports.
So by doing that, we will convert the potential energy to
kinetic energy and can calculate the impact forces and the rock size is
determined by joint spacing and also the height of the rock size is
determined by the joint spacing.
Actually, that randomly sampled between the joint spacing
and the height of the yield zone.
And I would like to emphasize in here is the failure
criterion we use in the key areas, the so-called two percent total
strain at a point of the contact. That's the failure criterion used.
This one gets very conservative in the sense that we're looking at waste
package here, at the time that we're doing this analysis, about 12
centimeter. So there is a considerable amount of conservatism
associated with that.
We understand the committee letter saying that we need to
use a more realistic model instead of this building so much conservatism
in there and we're in the process to revise these failure criteria by
doing the numerical modeling. Hopefully, at the end of this fiscal
year, we can come up with something that can be used included in the new
So this model, at this time, the waste package degradation
is not included in the model itself. So based on this current model,
what is the risk insight into the waste package performance and you can
see the -- in the figure on the right-hand side, vertical axis, is the
total dose and the horizontal axis is the time after emplacement, and
the two curves are shown in there. And after conducting 250
realizations, it will pick out one of the worst case, because of rock
fall effects, and the other one, it is same realization, without rock
You can see the differences here. There is about a 28
percent increase in terms of total dose if you have the rock fall and
without rock fall, considering the worst case situation. But out of 250
realizations, the average amount of rock fall in total is only two.
So this so-called mean total dose come out is relatively
small due to the rock fall impact in this case here right now.
And the thermal-mechanical effects on the hydrological flow,
I would talk about fairly quickly in here is that -- the hydrological
analysis, these two cases shown in here, and one is using the so-called
minimum quality of rock that show up in the ESF and this one is the
maximum rock mass quality that show up in the ESF, as well.
This horizontal bar is showing the permeability ratio after
the deformation, after thermal heating, compared to the original value
here. So you can see that it's more than one order of magnitude in
there in terms of increase in the permeability here.
One of the things that I mentioned earlier is that because
of the rock mass property and the strength being used, that you see, for
the stronger rock, the impact is much more. In this case here, in terms
of changing the permeability in this case.
And in terms of changing the shape, because we recognize
that no matter how you design the underground facility, essentially,
eventually it will deteriorate given time and some rock fall may have
taken place, that it will change the shape of the underground openings.
And I think Debra Hughson, who is in the audience, she did
the limited calculation that assuming that the emplacement drift is 5.5
meter in diameter and also if the roof of the opening changed the rock
mass to maybe 20 centimeter, the so-called threshold percolation for
dripping will decrease one order of magnitude. So it's quite a
So we are in the process of working with other KTIs and
trying to sort out what will be the impact of this type of change is
going to be.
Now, for the DOE has done, in the past six or maybe more
than six months now, that in terms of coming up with a new alternative
design. On April 15, they looking at about five different type of
alternative designs, with a different combination of design features,
and they think that alternative number two is the one that they need to
propose to the DOE for consideration.
The main feature in here is that they're going to go with
the 60 MTU in here compared to 85 for the VA design. They're doing that
by increase the drift spacing from 28 meter to 81 meter, and also reduce
the gap between the waste package from originally about six meter to .1
And also they're going to propose to have continued
ventilation, two cubic meter to five cubic meter for 50 years. The idea
for this one is trying to preserve the so-called shedding pillar design.
What the idea is that hopefully by using this type of design, the
temperature between the two neighboring emplacement drifts will now
coalesce, which means that somewhere in the middle, the pillar
temperature will be always below the boiling point.
So sort of provide some shield to provide access for the
water to flow through the pillar and go into the ground water level. So
it will reduce the amount of water that possibly will go into the
The other design related to the waste package is they intend
to add drip shield and also the waste package design is being changed
somewhat and also the total thickness has been reduced by five
Now, in response to DOE's new design, we conduct a
preliminary analysis using UDEC and trying to see the potential change
in terms of rock mass behavior around the openings. For lacking of the
data for those emplacement area, TW, Topopah Spring number one unit,
we'll do this analysis, we don't know how appropriate that's going to
be, but that's the best we have.
I am not going to show the results for the stress analysis,
just trying to touch upon the temperature distribution here. This one
essentially is about -- the temperature distribution at 150 years.
Using the new design and without considering the ventilation. You can
see the 100 degree contour is in here and in there.
And for the 50 year temperature distribution, the 100 degree
contour is approximately in here, I think, between the maroon color and
the yellow color.
So in essence, if they don't use ventilation, this shedding
into pillar mechanism cannot be realized. That's one thing. And DOE
has done preliminary analysis assuming that they have the ventilation
and assuming that the 50 percent flux will be taken out, and do an
analysis -- they found out -- I haven't seen the report yet, been talked
in different presentations, and the temperature in this pillar, 80
percent of them will be below the boiling point. That's something that
we need to verify, because based on this plot right now, we just don't
The progress to date, and Raj mentioned that in our IRSR
Rev. 1 and Rev. 0, that we have four major sub-issues. One is design
control process. For the design control process, we're using it and
continue talking with DOE and we also review their design. We found
out, at that time, the ESF design is acceptable based on that the
effective design control process.
But we are right now trying to review DOE's GROA design
using the same design control process to see whether it's still
effective or not.
The second sub-issue that we have is design for seismic
events and direct fault disruption. In there, DOE proposes three
topical reports. The first one is dealing with hazard assessment
methodology, that's being reviewed by our structural deformation and
seismicity group, and that's been accepted.
And the second topical report is a design methodology, and
reviewed and accept that. And we are waiting for DOE to submit the
topical report number three and we're going to review that and determine
whether we can accept that or not.
For the repository design, we just started looking at the
new design and see how that will impact the stability and in the past,
we also studied the fact of using different type of rock properties to
see how that will impact the repository behavior.
Raj has mentioned that the CO sub-issue has been dropped
because of the new regulation. So we're going to remove that particular
sub-issue from our IRSR.
Okay. The path forward, for the next two years, we're going
to focus our work in the pre-closure safety concerns, like Bill Reamer
mentioned yesterday that this particular activity has been rated as high
priority. So we're going to put the majority of our effort in this
Now, several activities coming to mind is that we're going
to continue to develop the Yucca Mountain review plan. That will
include the ISA, retrievability, and review the performance
Also, in order to supply ISA, we're going to develop a
capability to apply in the ISA in the regulatory framework. And one
important thing is that we may want to develop an evaluation tool that
can be used to conduct selected independent analysis of DOE's submittal
in this regard.
For the pre-closure related concern, we're going to upgrade
the SEISMO module in the TPA code. The basis is that the base case in
the TPA code and they engage in new design and the drip shield adding --
added and also that the waste package design being changed. So that's
what we're going to do in here.
Also, we're going to support the activities associated with
other KTIs to provide information regarding to the thermal-mechanical
effects in there.
That pretty much concludes my presentation. Any questions?
MR. FAIRHURST: Yes, I have a lot of questions, I'm sorry,
but we only have a few minutes. One, I was particularly interested to
see that in the EBS-2 that Raj showed, this layout diagram of what
you're doing, that you include consideration of dyke intrusion, which I
think is quite interesting, because we heard a very good -- had a very
good discussion yesterday of some of those issues.
I think if the rock mechanic group could link with people
there to look at the consequences of dyke intrusion, and particularly
consider the fact that the repository is likely to be elevated
temperatures during that -- for most of that period of regulation, and
that could result in a rather different situation, maybe, I don't know,
as far as a dyke being attracted to or diverted from a repository.
So it's just a comment, because I know you didn't discuss it
in the rest of your talk. So you probably have not done a lot with it
as yet. The second --
MR. HSIUNG: Not a whole lot, but --
MR. FAIRHURST: I'm not saying you should have. I'm just
saying it's nice to see that there is a linkage being discussed there.
The second one is that you mentioned a number of cases about
rock fall, et cetera, and the -- and your need to know certain rock
properties and modulus and so on.
But I really would encourage that you do sensitivity studies
to try to identify how much do you really need in data. For example, if
you look in the lithophysae zone, as you said, it's not as simple as
spheres. There are actually ellipsoidal cavities.
Well, I think if you were to look at how you would compare a
spherical cavity with an ellipsoidal cavity and taking the larger, the
major axis of the ellipsoidal cavity instead of the -- and, you know, I
think you will find -- and the range of influence of each of those
cavities is quite small because it's a three-dimensional opening.
And it's true that the modulus goes up -- or, should I say,
goes down the bigger the holes in the thing, but so does the strength.
And what is really determining the behavior is the ratio of the strength
and modulus, because the modulus is the thing that determines the
So if you start to do sensitivity studies, you will have a
better definition of what you feel that DOE -- I mean, maybe DOE should
be doing this, but rather than just saying we need to know the modulus,
we need to know this, maybe you don't need to know them all in a great
deal of detail.
And it's a different question between the pre-closure and
the post-closure. And as far as the pre-closure game is concerned, make
the analysis with supports installed, because the rock falls that one is
showing for post-closure would not occur, would not be allowed to
Finally, let me say that within NRC, you correctly raised
the issue that the way DOE has to design now, that the drifts would not
be accessible over the pre-closure, but accessible at considerable
difficulty because of the unshielded canisters.
So are you going to insist to DOE that they have some
monitoring facilities to assess whether failures are occurring inside
the drift and when? Because presumably, in your -- it's NRC's role to
say what is an acceptable safety strategy, and I don't know whether that
falls in it, but monitoring -- you know, remote monitoring for maybe a
drift above or along the side is a way to do that without risk to your
-- to the occupational people.
MR. HSIUNG: I think that I'd like to answer that you have
MR. FAIRHURST: Yes, I'm finished now.
MR. HSIUNG: The first question in terms of sensitivity
analysis, how much data is necessary, I think that our role is
determining that what is important that will govern the design. I
believe, in a sense, in here, I'm trying to present that they go with
the empirical equation to determine the rock mass and rock strength, and
that's fine. We don't have any problem with that.
That's why we accept -- I know you do, but we accept the
approach. But there is a caveat that they need to verify those
equations. Now, by doing that, however, much of data will be necessary,
that's at their discretion to decide. The representative case to us, is
this good enough, and then we'll exercise expert judgment to see, okay,
that's sufficient amount provided to us and we'll accept the approach.
So that's the reason that in here, that would come up, this
discussion in here is trying to show we are not really pushing DOE to do
And because some of the work they already promised us, they
haven't done it yet, we accept their design methodology based on that
premise. That's something that needs to be done. That's one thing.
And to answer that your second question, in terms of
monitoring, I think DOE is already has some program in the license
application report right now saying they are going to monitor certain
So right now, it's a matter of what is going to be adequate
or not. That's for us to decide. We haven't go into detail into that.
MR. McCARTIN: I guess one thing I'd like to add. Tim
McCartin, NRC. Part 63 does require them to monitor the repository the
entire time period and even provide a monitoring program after closure
to ensure that things are behaving as expected.
MR. FAIRHURST: Tim, my question is, if you look at the
design now, do you see how that monitoring is going to be done, because
if you have a -- you know, the drift radiation is sort of hazardous.
One would presumably have to do it through drifts that are not exposed
MR. HSIUNG: What they have is they have the Gantry design.
People can stay in there. So they just move the entry around for people
to look at the roof or anywhere, the support systems.
MR. FAIRHURST: Okay. But that's part of worker exposure
and issues that presumably that analysis concern.
MR. HSIUNG: Yes. Those things would bring into the
consideration in the ISA. Now, if they can demonstrate that people are
in there, that will be no problem, that's fine. They also proposed
another methodology using the photo, the camera, periodically during the
monitoring. I don't know how that's been done, but it is possible.
MR. McCARTIN: We haven't seen DOE's plans for it, but
beyond the worker safety, there is a requirement that they have to
monitor the behavior and we'll have to see what that entails.
Most likely, in a high radiation area, it would be done
remotely somehow, but there will have to be something in place to
MR. GARRICK: Any other questions from the committee?
MR. GARRICK: Thanks, Raj and Simon. Very good
MR. HSIUNG: Thank you.
MR. NATARAJA: Thank you.
MR. GARRICK: Our next presentation is on the thermal
effects on flow, key technical issues, Ron Green and Jeff Pohle will do
this and the committee member George Hornberger will lead the
MR. GREEN: My name is Ron Green. I am a hydrologist at the
center. My counterpart is Jeff Pohle at NRC. And there are a number of
participants at the center working on this; most notably, Deb Hughson,
who is working at a high level on this KTI, in addition to a couple
other hydrologists, Randy Fedors and Melissa Hill.
Just in response to some discussion yesterday, we have
several other consultants; namely, Ross Bgtzoglou and Sue Stothoff, who
were hydrologists at the center for five years, approximately, each, and
they are contributing to this. Danny Or is another consultant and then
Frank Dodge is institute staff senior technical person contributing to
The outline for my talk will follow as posted here. I will
first talk about TEF and its relationship with performance. I will
discuss briefly the abstraction into TPA, along with some sensitivity
analyses and results, some observations.
I will briefly go over some confirmatory lab tests that we
are doing and I will go over those briefly, because we will have an
opportunity after lunch to visit the lab and it will probably be a
little easier to do it at the lab.
I'll then discuss some comparisons with our results and DOE
results and then talk about the DOE design modifications, progress to
date, and the path forward.
I'd like to spend just a moment on this figure, just to
point out two things. TEF is involved in these two columns. In the
first column, where it discusses the engineered barrier, the importance
of TEF to this category is in terms of the integrity of the waste
package and corrosion, and then the second column, that's more with the
Once a canister has failed, radionuclides are released into
the geosphere, TEF can have some impact on transport. There are two
secondary areas where TEF may be important and that's in coupling with
chemistry and with the mechanical effects. That would be in changing
the permeability porosity of the medium.
So this next slide illustrates that a little bit. The
second bullet talks about -- addresses the impact of TEF on canister
integrity and basically that's in terms of water, the solutes contained
in the water, and the humidity and their impact on the waste package.
The third bullet addresses the issue of release once the
canisters have failed, and the first bullet just illustrates the
importance of corrosion of the canisters to TEF.
Just a general comment. I think as we move along in this
program, we notice that the KTIs have more overlap. That's very true in
TEF. We work more closely with the canister people, near field people,
and some of our objectives are overlapping. So some of the testing, the
analyses are performed jointly between the KTIs and I think you will see
that through this presentation.
The next slide is a conceptual schematic that illustrates
DOE's understanding or the way they view the effect of heat on the
repository environment. There are two important areas. One is the
condensate shedding, and that's one of the driving forces behind the new
DOE design. The intention is that they reduce the heat load, so that
water can be shut off in the pillars between the drifts. That's
something that, as Simon mentioned, we're going to be looking at
The other area is with the potential for REFLUX, and I think
it's interesting that in this DOE drawing, that they have shedding and
fracture flow going between the drifts, emplacement drifts, even though
in this schematic they have it above boiling and they have no indication
of that occurring above the drifts. That's one of the contentious
issues we have with their conceptual model and it's something that
you're going to see that we look at quite closely.
Just a little bit of discussion on the TEF in terms of
performance. Once again, we're kind of back to the waste package
performance and corrosion. Corrosion is dependent mostly on
temperature, relative humidity and liquid water. That's relative to
I'd like to also point out that it's also very important on
water chemistry, corrosion potential, and those other things. But in
terms of TEF, it's temperature, relative humidity, and liquid water.
For the most part, corrosion is highest in the range of
about 80 to 100 degrees. That upper end can change with the amount of
solutes that are in the water and some other things like that. The
lower end can be changed somewhat. This 80 to 100 is mostly for aqueous
corrosion. You could have humid air corrosion that's lower and you
could also have liquid water entering the drift with relative humidities
lower than that.
That's one area that we're looking at closely.
The VA addresses TEF directly. However, in their analysis,
they assume that no water contacts the waste package for the first 5,000
years. That's what the 85 MTU per acre heat loading. That's something
that they've mentioned that they are not going to assume in the next
TSPA and the license application.
However, neither in the VA nor in the TSPA/LA are they, at
this point, going to address the issue of water entering into the drift
prior to the boiling isotherm getting back to the drift; meaning, they
will not consider water getting back into the drift until the point that
the drift falls below 100 degrees. So they don't consider any
penetration of the boiling isotherm dripping down fractures, and that's
a major concern to us.
The TPA, which is the NRC version of the performance code,
uses MULTIFLO. You've heard mention of our internal code for both heat
mass transfer and site transport. It's a one-dimensional analysis.
It's based on the equivalent continuum. We're using that in order to
determine what essentially the chloride content is present in the water
when it gets back to the canister.
By being an equivalent continuum, we're not considering
fracture flow in this particular approach.
The TEF is directly addressed in the source term and with
regard to getting water available at the canister for transport away
once canisters have failed.
So at this point, we have identified technical needs that
we're addressing, that we're working on, and principal among these is
the arrival time of water on the waste packages and we feel
uncomfortable using a continuum model for this.
This may be a little more consistent with comments you made
about Bradehoff and Shlomo, using these idealized models. We don't feel
we can incorporate all these process level mechanisms in our TPA code,
but we feel that we have to address it, understand it correctly enough
so our abstractions do capture the physics of it.
So we have a considerable effort towards looking at seepage,
refluxing, and the dripping processes. And just for organizational
purposes, we're conducting this under the TEF, even though some of the
analysis is under isothermal conditions.
MR. FAIRHURST: Ron, could I just ask? When they assume no
contact for 5,000 years, is there a drip shield included in that?
MR. GREEN: No. That was just based on the dry-out.
MR. FAIRHURST: All right.
MR. GREEN: And that's in chapter three of their technical
basis document, and chapter five, which is the container, they do
sensitivity analysis where they do consider seepage onto it. But that
doesn't -- from my understanding, that doesn't enter into their final
So we also understand the importance of getting the spatial
distribution of seepage into the drift and the chemistry of this water.
So the abstraction process is, first, based upon accurate
process level modeling. We're using the MULTIFLO for parts of this.
Some recent enhancements in MULTIFLO provide us with the ability to look
at dual continua. DOE has been doing this for a couple years, we've
been doing it for about a year, year and a half.
It's my personal opinion that this has really enhanced our
ability to look at flow through fractured porous media, particularly
under partially saturated conditions.
We recently have included non-structured gridding, and that
allows us to have a little better rendering close to the circular drift.
There are some other general provisions provided by this
analysis. However, the MULTIFLO analyses don't give us a good
representation of episodic fracture flow, nor does it give us a good
feel for focused fracture flow into the drift. We're doing some other
mechanistic type analysis of this and I will discuss that in a little
So we abstract this information into REFLUX sub-modules.
MR. HORNBERGER: Ron, if I could just interrupt for a
second. It's always struck me that this reflux problem is fundamentally
at least a two-dimensional problem. How do you justify your treatment
of it using a one-dimensional code? Do you not see it as a --
MR. GREEN: No. I see it more actually as a
MR. HORNBERGER: As I said, at least two.
MR. GREEN: That was some early work that we did in order
just to get that calculation, the reaction path along one dimension. We
are not looking at the seepage or the dripping or the refluxing in one
MR. HORNBERGER: Okay. So MULTIFLO is not a one-dimensional
MR. GREEN: No. It goes up to three dimensions.
MR. HORNBERGER: Fine. Good.
MR. GREEN: We've had an evolution of REFLUX sub-modules.
We're currently on three and there may be enhancements to this as move
along. Basically, what we're doing in these sub-modules, we recognize
the fact that there are two sources of water. One is the infiltration
water from the natural infiltration and the second source is the water
from the rock dry-out. That will vary considerably between the
different heat loadings.
Back with the earlier heat loading, there could have been,
with 100 meter dry-out, there could have been as much as 8,000 cubic
meters per canister. That should be reduced significantly with the
reduced heat load.
As we progress into the REFLUX3, we have also included an
ability to calculate penetration of the boiling isotherm. So we can
have water advance in front of the boiling isotherm, but that mostly
goes into the transport away from the canister after release.
We're still working on how to incorporate that into the
first part, into the degradation of the canister.
So on conducting sensitivity analyses, we also varied some
of the basic UZFLOW parameters and what we found was that the UZFLOW
parameters dominated performance. So we're not comfortable that we're
capturing everything on REFLUX into our TPA code at this time, and
that's something we're working on.
I indicate as a bullet item what we varied in the first two
REFLUX sub-modules. So this next slide summarizes what we found. As I
mentioned, we're not comfortable with incorporation of REFLUX into TPA
at this point. We're going to try different ways of incorporating our
understanding into TPA, plus we're still trying to enhance our
understanding of the basic mechanisms.
So those are -- you know, we're refocusing our efforts along
those two lines.
We are somewhat comfortable with the way we handle it in
terms of radionuclide transport after waste package failure. However,
if the canisters last as long as they do, the thermal pulse is gone and
TEF doesn't really enter into it.
Another comment along those same lines, if these canisters
last as long as they do, if the heat load is reduced to what they're
considering now, TEF may not come into the equation in terms of canister
degradation. If that's the case, it's going to make our part of the
problem a lot easier.
MR. HORNBERGER: Ron, just out of curiosity, when you do
this calculation, is there any way, in an ad hoc -- at least in an ad
hoc way, that you'd take into account the thermal-mechanical effects?
MR. GREEN: In terms of reduced --
MR. HORNBERGER: Well, you're calculating reflux and clearly
during the heating period, you're changing -- the mechanical changes
affect the permeability.
MR. GREEN: We haven't done that yet.
MR. FAIRHURST: Sure, right. And the stress distribution is
changed between the drifts.
MR. GREEN: Yes.
MR. FAIRHURST: So you've got a potential for -- if there is
any -- put in a stress-dependent permeability, is what I'm saying.
MR. GREEN: Yes.
MR. FAIRHURST: And I think it would be very interesting.
MR. GREEN: I agree. It could -- and that goes back to that
spatial variability point, because it could prohibit -- you know, if
it's -- Simon has a comment.
MR. HSIUNG: I think that in the past, when we are in the
phase one, we did do some of this type of calculation in terms of the
thermal loading that would actually change the fracture permeability,
the stress level, and see how that --
MS. WASHINGTON: Simon, are you there? We can't hear you at
MR. HSIUNG: Okay. When we, several years back, when we the
program that we did do that kind of analysis, consider thermal load
impact to the stress level in the emplacement area and see how that
affect the permeability change, and we found out that normally when you
have this thermal load, the elastic type of solution being looking at,
and one of the things, the vertical joints were closed and that will
reduce the water in-flow.
But you're going to -- I think that increase the
permeability for the horizontal joints, so the water will come into
that. I think the percentage-wise, it's about 30 percent, also, in
terms of differences.
MR. GREEN: I think it's good to keep in mind that the
repository is not going to be uniformly heated. The line load is going
to help smooth out within the drifts, but there's still going to be
hotter and colder areas, and those hotter areas are going to incur
different stresses than the cooler areas and you can have some
I think that's something that we're trying to look at.
MR. FAIRHURST: I think this goes to your point of what you
were saying you will see overlap between the various KTIs and areas.
MR. GREEN: Yes.
MR. FAIRHURST: Because it's very definitely that's what is
MR. GREEN: Okay. I'm going to just comment on -- briefly
introduce some lab studies we've done, without going into any detail at
this point. I would just note that we have a test assembly. You will
see the framework of this in the lab this afternoon.
But we infiltrate water up here, we have a heater down here.
We've conducted two tests, we're setting up a third. In the first two,
we had concrete blocks and the idea was to infiltrate water and then
monitor it, as best possible, and see the nature of flow along the
fractures back to the heater.
MR. GARRICK: Can you make that heterogeneous?
MR. GREEN: Well, by nature, it is heterogeneous. Even
though the blocks were uniformly shaped, we did have some
heterogeneities. Our third experiment is going to be with crushed tuft,
and that will be more heterogeneous. We have different objectives for
I will discuss those in the lab when we're there. It will
be a little easier.
But just to note, briefly, that the -- we monitored the
movement of moisture. The best way we could do it was through
thermocouples based throughout the test assembly and looking at along
the fracture immediately co-planar with the heat source, we find that
the boiling isotherm, we had no indication that the boiling isotherm
penetrated the drift during the experiment.
This is important because DOE is using thermocouples in
their drift scale tests to give them an indication whether they have
reflux into the drift.
However, in the drift, we placed these drip sensors and both
test one and test two and in both cases we had dripping into the drift.
In test one, we were able to look through the Plexiglas side, so we know
that the dripping occurred throughout the experiment when the boiling
isotherm was at some distance.
I'm not going to try to say that we can scale these up.
However, you can use some empirical analyses, like Philips that we use,
in order to get an idea of that. But what we found here is that there
was dripping throughout about 75 percent of the drift in the first
experiment. This damage down here was caused by the heater.
The heater was at about 500 degrees C, but it was only one
inch in diameter, and that pushed the boiling isotherm about ten
centimeters into the rock.
But what we found was that we had these drips of this real
nasty mud-like material and there has always been discussion about what
kind of water will drip back into the canister and there has been some
feeling that it will be a condensate, so it will be relatively pure, it
will come down, it's not going to have much impact on the canister's
performance, because water, in itself, isn't going to degrade the
canister. It's going to be the solutes in the water.
What we found here was this stuff was oozy mud and that's
what we found here. And in test two, we found the same thing. In the
first one, I don't want to get into the details now, because that's a
discussion itself, but in the first one, the drift air temperature was
less than 100. But in the second one, it was 200 degrees C and that was
-- there are some reason for that, and we can get into it later.
But the performance of dripping into the drift was the same
over -- it was a little harder to tell what percentage dripped in the
second one, but it was for the most part of the drift anyway, and you
can see how this material completely degraded the drift sensor in parts
of both one and two.
MR. FAIRHURST: Are you saying this is some sort of
counter-current flow that is going up?
MR. GREEN: Yes.
MR. FAIRHURST: And water is coming down.
MR. GREEN: Yes, exactly.
MR. CAMPBELL: Ron, it would be very interesting if you
could actually collect drips and do an analysis of the water chemistry,
design an experiment to do just that.
MR. GREEN: Well, we actually put ports on the side of the
first one to collect it, but that water evaporates almost immediately.
But we did do XRD analysis of the precipitate that formed on the
fracture and on the drip sensors. We did a chemical analysis of the
water we infiltrated, and in the second experiment, we did have some
drainage through the bottom of the cell and we did a chemical analysis
of that, and we have all that information.
Now, just briefly, I think this next slide is really
interesting, because this is water quality analysis of water that was
collected from the drift at the DOE drift scale test. We only have one
analysis at this time, so we have a working hypothesis, and we'll have
to revisit this when we get more results.
But you have two samples that were taken below the drift.
These are from the hydrology holes that are directly below the drift,
and you have one sample collected from above. This one happens to be
right above this one.
And if you notice, the concentration of the ions in these
two are considerably lower than what you find above the drift. In
particular, look at the sulfates and the chloride. This is 600 times
above compared to below.
Now, I had this hypothesis and one of the first places I
went with it was Bill Murphy, because he shoots down every chemical
concept that I have, and he didn't disappoint me on this one. What he
said was that we can't go much further than a hypothesis at this time
because we don't know where we are in the gradient, the thermal gradient
and the potential gradients. We just might be taking that sample at a
different place in the gradient than from above from below.
And I have posted in the lab some chemical analyses from DOE
and we'll see that they're symmetric, that they show that the chemistry
is the same above as below. So we're going to watch very closely what
they're doing and what they're collecting at the drift scale test, and
it has profound effect upon our conceptual model of the refluxing.
MR. GARRICK: So what is your hypothesis?
MR. GREEN: The hypothesis is that the water below, as it
moves down, keeps getting more condensate, which has no solutes in it,
but above, the closer you get to the drift, you're nearing the end of
the evaporation pathway. So when that first water gets to the drift, it
just has a little bit of moisture in it and it's almost all solute, and
that's what goes into the drift.
So it's a highly mineralized water that goes into the drift,
and this is consistent with it. That's a hypothesis, but until we get a
little more data, we won't know for certain.
MR. WYMER: And you would expect that to change with time,
MR. GREEN: Yes, you will. But through the thermal period,
I think it will stay this way. However, there is another discussion I
had with Bill, he feels that later on, after the thermal pulse has
cleared the drift, that you're going to have the clean water come
through and flush things out.
MR. WYMER: That was my point.
MR. GREEN: Yes. And that's something we'll have to see.
And we can use some of these tests to support or strike down our
MR. CAMPBELL: Can I ask? To what extent, Bill, do you
think this is due simply to evaporation or also to water/rock
interaction? That could make a big difference over the long term.
MR. MURPHY: This is Bill Murphy. I was glad to see
chemistry brought up. I can only speculate at this point about the
significance of these particular chemical analyses. The high
concentration waters that Ron showed from above the drift are calcium
chloride waters. Their origin is something of a mystery to me. I
wonder about the potential influence of draught.
The notion that very concentrated waters may form during a
refluxing epic in the history of the repository is a very reasonable
hypothesis. But in the absence of that refluxing, one might expect a
return gradually to ambient water chemistry. There certainly will be
water/rock reactions and in my talk, which is next in this series, I
will be talking about some of our water/rock interaction calculations to
illustrate what we think or where we stand on that subject.
MR. GREEN: Just to reiterate two points. One is reflux is
not detected with the thermocouples, and that's due to the spatial and
the temporal collection of data.
You have to have that thermocouple right up at the right
point, taking the reading at the right time in order to detect it.
They did detect that at the large block test. So that's
sort of a fortuitous thing if you do detect it and I don't think you can
state that you're not going to see it if you don't get that information.
DOE doesn't have any other confirmatory way of letting them
know whether they have dripping into the drift scale test, other than a
camera, which they have not been using.
Then the last point is issue we just discussed about the
concentration of water above the drift relative to below.
Moving ahead to the DOE design, I will not spend a lot of
time on this. Simon went over the high points of this. Other than to
mention that ventilation will have a profound effect on the thermal or
the hydrologic conditions in the drift.
That raises a point that we've identified with the drift
scale test. The way they're conducting the drift scale test is they
have a leaky bulkhead, they're not monitoring the moisture or the heat
loss through that bulkhead, and it's hard for us to model it. It's hard
for us to understand everything going on in the absence of that lack of
Ideally, they would have conducted the test by putting a
very large bulkhead in there in order to allow the drift to come up to
perhaps some moderate pressures. However, it may be just beneficial to
them or fortunate on their regard because now they're going to consider
ventilation. But they're not monitoring it and that was a point that we
brought up at an Appendix 7 meeting a couple months ago that Dr.
Fairhurst participated in.
MR. GARRICK: Ron, can you give us a hint of your opinion of
the profoundness of the effect?
MR. GREEN: They -- yes. In the DOE preliminary
calculations, this was mentioned by Simon, is that they're assuming 50
percent heat loss during ventilation. They're not putting ventilation
in their calculation. They're putting the assumption of heat loss. So
the mechanical effect of ventilation isn't in there.
We conducted a laboratory scale ventilation test about a
year ago and we only had about 20 percent heat loss through ventilation
and we're still looking at our results. We're a little -- we're not
completely through that, but we don't buy off on their 50 percent at
this point. So it's somewhere in that range.
If you put in a 50 percent heat loss in your heat load,
you're going to -- you're not going to experience the coalescence of the
boiling isotherm between the drifts. That's one of the reasons they
opted to go with this design. But it's all predicated on that one
So that is a profound effect. Plus, it removes moisture and
that's another reason I don't think they will see dripping in the drift
scale test. They have the ventilation going out, and so they're not
going to see dripping.
We've told them that we don't -- we're uncomfortable with
them saying they're not going to experience dripping in emplacement
drift based upon the results of this drift scale test, in the absence of
ventilation. If they add ventilation, I don't think they'll see the
dripping during the ventilation period.
The progress to date. I mentioned that we've conducted two
laboratory tests and we've learned an awful lot from these. We have a
third one in preparation. That will be -- that's jointly conducted lab
tests with the CLST group and the primary objectives are actually geared
towards them. We're going to use crushed tuft and we're going to
infiltrate water again.
We're only going to get the temperature up into the 80 to
100 degree range, because that's where corrosion is of interest. And
we're going to monitor it as best as possible to understand the
And one of the criticisms that our group has had of the
drift scale test is that they're putting coupons in their drift scale
test and they're going to take these out at the end of the eight-year
test and they're going to say, well, we've experienced this much
corrosion or we haven't, but they haven't been monitoring the corrosion
So Sridhar's group is actively trying to identify the
sensors that they want to put into this drift and we can talk about that
a little bit this afternoon.
The results from our lab scale tests are consistent with the
drift scale test. However, they both illustrate different aspects. We
observe penetration of the boiling isotherm in the lab test. I don't
know if they will see that in the drift scale test. I would be
surprised if they do see a drip into the drift itself.
And as I've mentioned, I think it's very important that we
both observed these elevated solute concentrations above the drift.
One critical point is once again, the penetration of the
boiling isotherm by flow down a fracture. DOE has not included that and
at this point does not intend to include it in their performance
assessment. I think that's a concern.
These last three items, as I alluded to earlier, address,
somewhat on a mechanistic level, analysis of seepage, dripping into a
cavity, flow back to a heat source. This past December, we had an
internal workshop, we had about 15 people participate, NRC, center staff
and some consultants, because we feel this issue is very important.
We have identified some tasks out of that and we're working
actively on those.
So the path forward, we're going to assess the new
repository design. As I mentioned, the time and the flux of water
arrival at the waste package is of critical concern and we're doing an
awful lot of work in that area.
We're going to continue to scrutinize the fundamental
assumptions taken by DOE in their performance assessments, continue our
sensitivity analyses, using the process level models and TPA to continue
to explore what's important to TEF, and then finally, we've most recent
-- just recently completed our recent revision to the IRSR and we'll
continue working on that.
That concludes my presentation.
MR. HORNBERGER: I'll start. I just have a couple
questions. Why do you decide to use crushed tuft rather than try to
simulate more a fracture matrix thing, like you did with your concrete
MR. GREEN: The main reason is it would be very difficult to
have -- we have 520 concrete blocks in the earlier one. It cost 100,000
-- $200,000 to prepare a tuft like that and we learn an awful lot about
fracture matrix flow. We're not going to try to replicate that in this
I think what you're going to learn, the best things you can
learn about fracture flow is in the field in the drift scale test. It's
very hard to capture matrix fractures on a reasonable representative
elemental volume in the lab, and it's mostly because of that.
We're going to use crushed tuft, from gravel size to powder,
and the reason we're using the entire fraction of the crushed rock is so
that we have a very high surface area for the chemical interaction of
the infiltrating water, so to create the environment as close to what
they might expect in the emplacement drifts.
MR. HORNBERGER: Because of the backfill?
MR. GREEN: No. Just so that the water coming into it will
be close to water that's dripping through tuft.
MR. HORNBERGER: Of course, by crushing tuft, you're
creating lots and lots of fresh surfaces. So you'd probably get an
argument on that.
MR. GREEN: Yes. We got it from Bill. We already got that
argument from Bill. That's correct. And anytime you do a laboratory
experiment, you have to make compromises and that's one I thought we
This afternoon -- we're still in the preparation of this and
this afternoon, I would really like to hear your assessment of a
different way of doing this. We may go forward with this experiment,
but if you have some ideas on an alternative approach, I'd be glad to
MR. HORNBERGER: It strikes me as quite interesting, but it
also seems to me that it will be critical that you do get the
involvement of the geochemists, because a lot of the important data
you're going to get will depend upon making those measurements as well
as the physical measurements.
MR. GREEN: Yes. They are involved and we've met with them.
Our most recent discussions are on what chemistry should the water be
that we infiltrate.
MR. HORNBERGER: That you infiltrate, right.
MR. GREEN: Yes. And we're doing some just low scoping
experiments, where we're putting different waters in a bucket with the
crushed tuft and allowing that to equilibrate for a couple weeks and
look at it.
So the question is still open as to the chemistry of the
infiltration water, yes. Plus, we're going to look at this experiment
and try to use our models to replicate the reaction path. We've looked
at some -- we had it as a higher priority to look at the past
experiments, but that was a concrete environment. And it had some
relevance with the former design, where they had concrete ground
support, but it has very little relevance now, other than giving a nice
example against which to test your code.
So we've dropped out -- we've dropped the priority on that
analysis, but at this point, we're going to go forward with the chemical
analysis of the tuft, the reaction path.
MR. HORNBERGER: That's good. The other query that I have
has to do with your modeling, your MULTIFLO modeling. You said you're
doing 3-D modeling and you're particularly interested in this reflux
It strikes me that that reflux problem must be an
instability problem; that is, you're talking about flow fingering. How
do you model this in a continuum sense?
MR. GREEN: Well, we're not. Some of those analyses are
mechanistic. They're looking at single fractures.
MR. HORNBERGER: This is the Philips solution.
MR. GREEN: That's one of them, yes. We're looking at
others. We're coming at it from about three different directions to
look at it. We're looking at continuum, but we're also looking at these
mechanistic that don't look at it as a continuum. It's a single
MR. HORNBERGER: But that single fracture, that's a
one-dimensional analysis, again, isn't it? So how do you determine the
frequency of the fingering?
MR. GREEN: No, it's a two-dimensional, actually.
MR. HORNBERGER: Oh, it is?
MR. GREEN: There's different approaches. I didn't have an
opportunity to go into detail. We're looking at groove, groove flow.
MR. HORNBERGER: Okay. Yes.
MR. GREEN: That's one approach, along with fracture flow,
and at different potentials, matrix potentials, you're going to have
different contributions. And your biggest contributions are from the
grooves and that's -- you know, my feeling of flow down a fracture in
this environment, the best image I can get is water flowing down a
windshield or a window during rain.
And what you see is that water will be held up until it
reaches a threshold potential and release and go very rapidly over a
short distance. I think that's how the fractures operate. And a groove
fill flow is one of the -- it's about one of the better ways that we've
come up with for looking at that mechanism, along with the fill-ups, and
we have a couple other approaches that aren't as --
MR. HORNBERGER: And I gather you're doing percolation
thresholds from one of your slides, right? So you're looking at
percolation models, as well.
MR. GREEN: Yes. Yes. And I think Debra is back, she's
looking at some other -- the effects of capillary diversion and the
geometry of the drift.
If you look at the VA, everything has a nice uniform
geometry, with a continuum, and if you put water down there, it's going
to go around every time. Well, that's not how -- if you notice, the DOE
folks, the people doing the process level are essentially all physicists
and the people doing all their performance assessment are all mechanical
If you're an earth scientist, you look at things
differently. You know, rocks are very heterogeneous. I mean, even as
uniform as we try to make them in the lab, they're heterogeneous.
The fractures are very heterogeneous. The fluxes are
heterogeneous. The medium is going to act very differently than these
nice continuum. If you have any protrudence off the top of a drift
wall, that's where dripping is going to go. You can see that anywhere.
You can go in a mine and if there is a rock there and dripping is coming
off that, or just any jagged edge, and that's not captured in their
models and they have very high -- in their base cases, they have very
high diversions, and I'm not comfortable with that at all.
MR. HORNBERGER: But one last question.
MR. GREEN: Yes.
MR. HORNBERGER: With a C-22 outer shell on the canister,
does dripping matter? What do your corrosion people tell you, does it
MR. GREEN: I think the vulnerability -- I had it on a
slide, I didn't mention it, but the vulnerability will be in the closure
seams -- the closure welds. So I know I've talked to Darrell Dunn and
his -- and their group and that's an area that they're now focusing a
lot of their efforts.
MR. HORNBERGER: Charles, do you have anything?
MR. FAIRHURST: I notice you're talking about putting the
steel ground support in.
MR. GREEN: Yes.
MR. FAIRHURST: You know, the recent drift stability panel
advised against that and I suspect that is going out. But what is the
significance of that, as far as you're concerned?
MR. GREEN: Well, two things. One, with the type of medium
we have, with this next test, that's crushed, we're going to have to
have something to keep it open.
And because of the EBA-2 recommended steel ground support,
we're going to use steel similar to that to replicate the environment.
MR. FAIRHURST: So it has significance from the chemical
point of view.
MR. GREEN: Yes. I mean, to some degree, yes. It would be
better to use than a ceramic -- we used ceramics to hold up our
instrumentation in the second test, for instance.
MR. FAIRHURST: The idea is to use lock bolts and
reinforcement, because steel sets. It may look like a nice -- but it
doesn't work that way in practice. It's not easy to put those in.
MR. GREEN: I'd like to discuss this a little more when we
go to the lab and you can give us some ideas.
MR. FAIRHURST: Fine. Okay.
MR. GREEN: That measure closer to what they're going to go
MR. FAIRHURST: I'm not sure they know what they're going to
go for. At least the last train around, that got low marks.
MR. GREEN: I went to the thermal abstraction meeting, which
was on a Tuesday, people prepared their slides on Friday, and they were
out of date because they made changes on Monday, and we have to
So in our acceptance criteria, we had this discussion last
week, they're changing things, and so the discussion was should we
remove these acceptance criteria. The feeling was no, we'll leave the
acceptance criteria in, we'll close them till some point where they may
bring these concretes back in, and then it will be reopened.
So there is a moving target out there and that's a challenge
to keep the crosshairs on the correct rabbit that's running around.
MR. WYMER: I just have one comment. I thought that the
viewgraph you showed where the water composition changed so dramatically
above and below the drifts was interesting.
But I wonder if that is really very important as far as
corrosion is concerned, because the period during which you get that
concentration is when it's hot and that's not going to last very long.
Then you get the more dilute water coming down and you'll flush the
system, and does it matter?
MR. GREEN: That's very possible and our challenge is to
find out -- to determine how much water, when it gets there, and how
long it comes in before it's diluted, and whether the canisters will be
vulnerable to that exposure, and, if they aren't, our problem is half
done, we can go on half days, sleep until noon, and come in and then
just worry about source term in the afternoon. But that's exactly
MR. FAIRHURST: Ron, even with the backfill scenario, the
idea, I think, is not to actually fill right up to the roof.
MR. GREEN: Right.
MR. FAIRHURST: And to maintain even with natural
MR. GREEN: A gap.
MR. FAIRHURST: -- a gap with air.
MR. GREEN: Yes.
MR. FAIRHURST: Presumably, at least if the climate doesn't
change that much over the next 10,000 years, is generally low humidity.
So how much will the fact of having a constant ventilation eliminate or
reduce the dripping?
MR. GREEN: One of the -- there are a lot of analyses we'd
like to do and we've put on our to-do list and we've never gotten around
to them. One is them is natural convection that would form, even in the
case where the emplacement drifts are closed off and you can get natural
ventilation, along the lines of Ed Weeks' work, that you can get
Especially if you have a variable heat load and you have
heterogeneities in the medium, if you have an area of lower permeability
and higher permeability, variable heat load, you can get natural
ventilation. If it's closed off, the humidity is going to be
essentially 100 percent.
If you have a source of air from the outside, it's going to
be low enough that you're going to remove all the moisture.
MR. FAIRHURST: That's the key issue. If you have a source
of air from the outside, it's going to eventually remove all the
MR. GREEN: Right. So it's not been tempered by the
moisture. The mountain is very moist. The matrix saturation is over
90. They've changed it a little bit in light of recent measurements.
MR. GARRICK: I have no questions. Staff? Andy? All
right. Thank you very much. That was very interesting. Okay. I think
we'll take our break, a 15-minute break. That will allow us to get a
five-minute jump on the restart. So by that clock, we want everybody in
here by 20 minutes to.
MR. GARRICK: We're now going to get to the chemical side of
things, among other things, evolution of the near field environment and
committee member Ray Wymer will lead us in this discussion, and Bill
Murphy will be our presenter.
MR. MURPHY: Thank you very much. I would first like to
acknowledge, in my first slide, there are quite a lot of people involved
in this project. It's very multi-disciplinary field.
I would like in particular to acknowledge English Pearcy, my
manager, and Brett Lesley, NRC, the program manager, have both
contributed a lot, and a large number of people on both the NRC and
CNWRA staff, and I will note some of them as I go along. They have all
contributed to this.
In my next slide -- my next slide is the presentation
outline, which is not the slide I'm looking at. But the presentation
outline is essentially the same as in the agenda. It's essentially the
information that's in your handouts and in the presentation outline, you
will see it in your agenda.
I will talk first about our view of thermo-hydro-chemical
effects on performance and that's a major theme of the near field key
technical issue; some of the technical bases; some risk insights from
performance assessments; some of my personal views on design changes
significance, this is all relatively new; some of our progress, and some
ideas for the future.
This diagram illustrates one set of systems for organizing
the various aspects of performance for the waste repository, the
proposed waste repository. The near field key technical issue has
direct connections with, I think, seven of these 14 key integrated
sub-issues. I won't go through them in detail. But once again, I will
emphasize the fact that this is a very multi-disciplinary and integrated
We've alluded to that already this morning, as you've heard
in the talks on the thermo-hydrology work and the RETME work and also
we're very close integrated with the container life and source term
work, and also with performance assessment people and so forth.
So Wes mentioned yesterday the advantages of the kind of
integration that can occur in a place like the center and I appreciate
that myself, as well, that there are interactions among -- close
interactions among many of us and many of us are sufficiently good
friends that we can make off-the-cuff remarks about one another up here.
MR. GARRICK: You've been the victim of a few of those
MR. MURPHY: The thermo-hydro-chemical effects on
performance are organized here in regard to the five key -- of the five
sub-issues of this key technical issue, being all focused on
thermo-hydro-chemical effects or processes; first, on seepage and flow
of ground water; on the waste package chemical environment; on the
chemical environment of the waste forms; on the chemical environment
affecting radionuclide transport in the near field; and, on criticality
in the near field.
There are potential THC processes on all of these areas of
the physical evolution of the near field environment and they all have
potential effects on performance. Once again, I'll note integration
with many other key technical issues and a lot of that integration
occurs through the integrated sub-issues structure that's been
The summary of our current technical basis is a rather
generic survey of the kinds of information that we make use of; site
characterization data, lab data, et cetera. You can read through this
list. Performance assessment in general and sensitivity studies and
abstractions for performance assessment in general. I refer not only to
the center's work, but to the DOE's work and work outside of the NRC and
the DOE, where we make an effort to have a broad view of sources of our
technical basis for the conclusions and analyses we're doing.
In the next couple of slides, I'm going to show rather in
some detail, as much detail as we need, some examples of results of the
technical work that we've done that illustrate a variety of uses of
In the next slide, this slide is -- I'll give you the jargon
first. It's a MULTIFLO calculation of one component, dual continuum,
two-phase fluid flow, about nine kinetically reacting minerals, about 20
aqueous and gas species, 60 metric tons of uranium per acre of thermal
loading, which is the new DOE design. These are results of the MULTIFLO
calculations to separate -- somehow I went forward. Can I go back a
To put this in more general terms, this is a plot of one
particular chemical characteristic, the pH of the aqueous solution as a
function of depth in the repository, calculated for both the ambient
geochemical conditions, on which there is quite a bit of uncertainty at
present still, and for thermally perturbed environment.
One sees that depending on how one interprets measurements
of the ambient chemistry in the near field environment or in the natural
system, one can have various boundary conditions for this kind of
coupled reactive transport modeling. One sees that the thermally
perturbed case illustrated here is for a time of 50 years, which is
approximately the maximum temperature given by their dotted curve,
reaching near the boiling point of water. One sees an excursion, a
significant excursion in the pH.
That's not the only variation that occurs due to coupled
thermo-hydro-chemical effects. There are changes in chloride, there are
changes in oxygen, there is mineral dissolution and precipitation.
Our development and utilization of the MULTIFLO code to
model these processes is represented in this figure. This, I must
attribute much of the work here to Loren Browning and Debra Hughson, who
presented this graph in part of an AGU presentation last fall.
I can go into -- or they could go into a great deal more
detail, if you're interested. But one of the points that I care to make
is that we are making an effort to integrate aspects of our
characterization of the site, our knowledge of the site design, our
knowledge of the properties of the materials and the environment, are
modeling capabilities that were developed to couple
thermo-hydro-chemical effects to generate models relevant for the
MR. WYMER: What did the temperature go to there?
MR. MURPHY: Well, the temperatures on the upper scale, the
maximum temperature is about 100 degrees C.
MR. WYMER: I'm sorry, I see it. I didn't see it.
MR. FAIRHURST: What is the black bar?
MR. MURPHY: The black bar represents roughly the variation
in pH due to -- during the thermally perturbed condition due to
differing assumptions about the initial conditions of the infiltrating
solution, depending upon differing interpretations of the analytical
data for the unsaturated zone ground water chemistry.
MR. FAIRHURST: Is it just essentially covering the three
MR. MURPHY: Yes.
MR. FAIRHURST: It's just the span.
MR. MURPHY: Yes. The span of the three different models
illustrated in this graph for the thermally perturbed system and for the
-- and the various -- this is something that needs to be clarified.
There are different models here. APPS is a reinterpretation
of the ambient ground water chemistry. Yang's data are the literal
analytical values for the ground water chemistry which can be
demonstrated to be wrong, and this is not a criticism of those data,
they're very -- it's a very hard thing to know, very hard to sample, but
it -- I think it requires a reinterpretation in order to get more
realistic view of what the ground water chemistry is.
APPS has done that, we've done that. These various
approaches have been used for the modeling only to set the initial
conditions at the top of the model. Everything below that depends on
those initial conditions and then the subsequent reactive transport
calculation of gas and water flow and mineral, gas, water, rock
So there is one complicated slide that illustrates many
different aspects of our studies. The next slide, which I can't get to
either, it's that one, the next slide shows another example of some of
our relatively recent work supported by the evolution of the near field
environment KTI, and once again, the jargon, this is a CCDF for
potential Yucca Mountain repository modeling the probability of
exceedence of peak annual doses, given on the horizontal axis, for a
critical group located 20 kilometers from the emplacement area, for time
periods of 10,000 and 50,000 years.
And we have three different curves shown in this diagram.
The curves labeled by the lower cased symbols, which are a little
difficult to see, but they're the three on the left, are for a
10,000-year period of performance in which corrosion doesn't influence
performance at 20 kilometers. The three curves on the right, indicated
by the upper case symbols, are three alternate models for this CCDF, for
a 50,000 year performance period.
And the three models are the results of complete performance
assessments, sampling, all the variables generally in the way the 3.2
TPA code samples variables and calculates performance, varying only the
source term for the various comparisons.
The curves labeled B are the base case NRC source term
model, which perhaps you'll hear about or have heard about. I won't go
into that in detail. The curves labeled N are based on an
interpretation of the maximum average rate of oxidation of urananite at
the Pena Blanca natural analog site, which has been interpreted based on
our field data and radiometric dating of the site, and scaled to the
Yucca Mountain system and applied to this performance assessment
And the curves labeled S are based on our secondary
schoepite, secondary uranium mineral, solubility model. Dick Codell
spoke yesterday about these alternate source term models and this is an
example of the work that Dick and I have worked on together, along with
One sees that both the natural analog and schoepite
alternate source term models give substantially lower doses than the NRC
base case model. I'd prefer not to make the case that either of these
is a conservative model in every respect, in many regards, particularly
the schoepite model can be criticized for being non-conservative in many
But I think there are very large uncertainties in the NRC
base case model and in the natural analog and schoepite solubility
models and that an appropriate approach to dealing with this uncertainty
through sensitivity analyses, such as these, in which we consider the
system from various points of view and take alternate conceptual and
numerical models to evaluating the significance of the source term.
Now, this is work sponsored by the near field environment
key technical issue, because, to a large extent, the evolution of the
waste forms depends on the chemistry of the near field environment, the
oxidation of the waste forms in particular, and the potential
incorporation of secondary -- of radionuclides in secondary phases.
I can talk about this in a great bit more detail, if you'd
care, but once again, my main point is to illustrate that we are
coupling our work with performance assessment. We're drawing from our
natural analog studies, from our understanding of the Yucca Mountain
system, and we have a fairly integrated approach here.
The next slide illustrates my view of some risk insights
from performance assessment and for those of you who saw Dick Codell's
talk yesterday, you will recognize that it's not necessarily an easy
thing to do to establish a unique way to identify where the potential
risk is and where it's not.
He showed a variety of ways of testing sensitivities to the
model. In one of the sets of slides he showed, he illustrated the
quantity and chemistry of water contacting the waste package has a major
effect on predicted performance.
Among the integrated sub-issues, this one repeatedly comes
out as being a dominant issue. It's strongly affected by the evolution
of the near field environment to the extent that the evolution of the
near field environment affects permeability and porosity of the
surrounding rocks and it affects flow, it affects the chemistry of the
water that may interact with the waste package and the waste form.
I showed in the previous slide how alternate source term
models can have a large effect on predicted performance, and so one of
the risk insights I drew is that alternate source term release rates can
have a major effect on predicted performance assessments, even though
that wasn't illustrated by the particular judgment of relative
significance of key technical or integrated sub-issues in Dick's talk.
That showed up very clearly in his ranking of alternate
models and the sensitivity studies, in which he showed that these
particular natural analog and secondary phase-based source term models
showed the very lowest releases in all of the alternate models he
In contrast, there are a number of risk uncertainties that
still exist in performance assessments as they have been performed, and
in my view, part of the reason for this is the exceptional complications
associated with coupled thermo-hydro-chemical and mechanical effects,
and I know this is an issue that the ACNW is aware of and has made a
point to pay attention to the development of the way and the manner in
which we address these coupled processes.
Some things are simply not in NRC's performance assessments
at present. Near field criticality is not considered at all. I think
that's a very reasonable thing. It's one of the sub-issues in the near
field key technical issue that I believe will be resolved or closed
eminently. I think that's a reasonable thing not to include in our
performance assessments, based on auxiliary analyses.
Our near field chemical effects on the flow regime are not
presently included in NRC's performance assessment. Now, there as some
discussion in the previous talks about the potential effects of
thermal-mechanical effects on the flow regime. The potential effects of
chemistry on the flow regime in the near field are really very enormous
and there are a lot of field data, there are a lot of lab data, and
there are emerging modeling data to sustain this idea from our work, as
well as in work supporting the DOE TSPA.
In the near field altered zone models report, there is a
sensitivity study presented in which there is a one chemical component
reactive transport model presented, and this is the most sophisticated
model I've seen applied in performance assessment at that level, and
they showed enormous variations in permeability and porosity due to
formation of a silica cap above the emplacement horizon.
None of this is taken into account in anyone's performance
assessments strictly at present, and this is a big uncertainty, I think,
that still needs to be addressed and it's one that we're working toward
and the DOE is working toward. This is not intended to be a criticism
of anyone's work. It's just a statement of where we stand at present.
Temporal changes in near field chemistry are potentially
included in our TPA code. It's designed to be flexible to adapt to
various conditions, but presently, chemical environments for each
realization in the NRC TPA code are constant, regardless of the thermal
Infiltration changes, temperature changes, chemistry is
constant for every realization. This is not very realistic at all. I
think that DOE recognizes this, as well. In TSPA/VA, they've taken a
step, what I think is an important step forward in recognizing the
temporal variation in chemistry during the thermal period and in their
TSPA/VA, they have, I think, five periods of different chemical
environments that may interact with the waste package and the waste
form, and I think this first attempt in TSPA/VA is a laudable or
important advance and I think we need to be moving in the same direction
in our own performance assessments.
MR. FAIRHURST: Is there any general statement you can make
about the effect of chemistry, such as it will it increase the
permeability and then decrease it further out?
MR. MURPHY: I think that the general statement that can be
said is that, yes, it will increase the permeability in porosity in some
places and decrease the permeability in porosity in other places.
MR. FAIRHURST: In some places and other places -- I asked
more specifically, is it going to be in the immediate vicinity of the
excavation, it's likely to be dissolution, and, therefore, probably
precipitation further out.
MR. MURPHY: In general, precipitation will occur in areas
of evaporation, where water is evaporating or water is becoming more
concentrated, minerals will be precipitate, precipitation will occur
where water exists, that's predominantly in the matrix. The fractures
will be predominantly gas fill. Dissolution will occur where water
condenses, because the water vapor is essentially distilled water. It
may have elevated C02 pressures, it may be slightly acidic. Dissolution
will occur predominantly where water condenses. This may be on fracture
linings, where there may be particularly susceptible minerals to
dissolution, such as calcite.
So that's a general answer to your question. Maybe I can be
MR. FAIRHURST: We'll talk about it later, Bill.
MR. MURPHY: I'd be happy to.
MR. FAIRHURST: Yes, I'd like to.
MR. MURPHY: I think that with regard to the extent of these
changes, it will -- it depends on time and it will depend on the thermal
loading very much.
So in my following slide, that's one of the points that I
address explicitly right up front. I think that there are potentially a
lot of responses of the evolution of the near field environment to the
proposed DOE design changes, the absence of concrete liners affects near
field chemistry, the introduction of backfill affects the near field
chemistry, addition of titanium drip shields, I don't think anyone has
looked very carefully about the extent to which this might affect near
People have considered it in regard to flow, maybe it's
inert, maybe it's not, I'm not sure. The lower thermal loading will
affect the chemical regime in the near field environment. The line
loading, as well, could affect the distribution of flow and the
distribution of chemical effects due to a couple of
thermo-hydro-chemical effects in the near field.
So I think that there are potentially important consequences
of the DOE design changes to the evolution of the near field environment
and I think that for the most part, these changes make the problem
somewhat easier, make the technical aspects of the problem somewhat
The non-coalescing boiling isotherms between the drifts
allows the system to remain relatively more like the ambient system in
the pillars. That's much easier to model than more or less continuous
silica cap that might cover square kilometers.
The absence of concrete greatly simplifies the chemistry in
the near field. The lower temperatures greatly simplify the
calculations of the evolution of the water chemistry and the mineral
chemistry. So for the most part, I think the DOE design changes make
the technical aspects of the evolution of the near field environment and
particularly the chemical environment somewhat easier to address.
On the other hand, there are some new issues introduced in
the design changes, the effect of titanium I touched on, the effect of
backfill on seepage is an issue that hasn't been very well studied, to
my knowledge. I've heard people speculate about certainly asparities in
the drift walls, but how about the flow field of the backfill material
and its evolution with time due to thermo-hydro-chemical effects.
The steep thermal gradients, steeper thermal gradients as a
consequence of backfill may have some impact that I don't think has been
studied very seriously at present, and there may be interactions newly
introduced materials that haven't been thoroughly considered.
So I think we need to respond to DOE design changes, we need
to be aware of it, and we need to be flexible in adapting to their
In the next slide, I'd like to report on some recent
progress and accomplishments. We've revised the evolution of the near
field environment IRSR with applications of the acceptance criteria
developed in previous revisions of the IRSR, specifically to a review of
DOE's TSPA/VA with some very specific comments about whether or not we
believe that their VA performance assessment meets the acceptance
criteria, as we have posed them.
We have done a number of sensitivity studies of alternate
source term models based on natural analog data and secondary mineral
solubility. Some of those results I showed you previously. We're
planning for laboratory and theoretical studies on cementitious
materials and for the role of secondary uranium minerals in the
evolution of the near field environment and on performance.
We're developing, we're continuing to develop our coupled
reactive transport model, MULTIFLO. We've done some serious
benchmarking exercises with other well established codes and analytical
solutions recently, and we're making applications, as I showed in a
previous slide, to the new design, thermal design for the Yucca
Mountain, proposed Yucca Mountain repository.
Also, we are trying to maintain close interactions with the
DOE. We've attended the DOE workshops on how they plan to deal with
near field and waste form issues. Particularly, Brett Lesley is a
regular participant in the thermal testing workshops. We participate in
the natural analog working group, which is an international organization
that deals largely with -- not largely, but among other things, with
evolution of waste forms in a repository environment.
I'll mention incidentally that our natural analog work,
which is supported by the near field environment project, has progressed
quite far and this week, the Department of Energy is making their first
research trip to Pena Blanca.
We also attend other public meetings and make publications
and presentations and so forth. So we're quite active in trying to keep
up with the moving target and with developments in our own abilities and
Then the path forward, we're interested in closing KTI
sub-issues. I think near field criticality is one likely to be closed
in the near future. We're looking forward to a revision of the near
field IRSR with the movement of acceptance criteria into the license
application review plan, to which we're contributing.
We're not closing interactions. We have close interactions
with other KTIs to address integrated sub-issues. We're continuing to
do and contemplate sensitivity studies regarding design changes and our
own enhanced modeling capabilities, and enhanced and enlarged database
coming from the thermal testing regime in particular.
We have laboratory and theoretical studies of cementitious
materials and secondary minerals in the planning and underway. We plan
to continue making use of our natural analog studies at Pena Blanca to
address issues of alteration rates, as an alternate source term
We're reviewing DOE FEP, feature-event-process data list for
coupled thermo, hydro, chemical and mechanical issues specific to Yucca
Mountain. It's not clear that their feature, event and process list is
necessarily comprehensive with regard to the things that we think need
to be considered.
And furthermore, and finally, we're planning for performance
confirmation period, partly because this is a difficult issue. Not
every question is going to be answered to everyone's satisfaction. We
anticipate that the DOE will state that in their license application and
will have a plan for future studies and a plan for confirmation, and we
need to be aware in planning for that as well.
That finalizes my formal presentation and I'll entertain
questions. Also, I'll be in the lab this afternoon.
MR. WYMER: I think you've done a very good job of pointing
out the complexity of the chemical system and, in particular, the way it
couples to all these other areas. You said seven out of the 14, and
maybe there's even more than seven.
MR. MURPHY: It may be more, right.
MR. WYMER: So it's an extraordinarily complicated system.
I made a couple of notes as we went along. One is, I'm always concerned
about what abstraction of the model does in the way of losing important
things. This is, I think, particularly important in the case of coupled
In general, when you talk about coupled processes, in
general, here at the center, you're talking about coupling among the
various systems and subsystems that you show on your flow-down chart,
but there also is a refinement, as you know better than I do. The
coupling of the chemical processes themselves, which are so
inter-dependent one upon the other as things change, which is a good
deal more subtle than coupling these other processes.
MR. MURPHY: It appears you understand that very well, too.
MR. WYMER: So you worry about the abstraction sort of
losing those sorts of things. How comfortable are you about that?
MR. MURPHY: I will address that in the way I perceive that
you intend it. There are a lot of coupled chemical interactions simply
regarding the system as a chemical system, and the first order way in
which that is addressed is to regard parts of the system to be at
equilibrium, to be representable by chemical equilibrium relations in
which we have mass action relations. They can be solved sometimes
analytically, in general, numerically, and where we have data to support
those equilibrium relations, that's a very nice abstraction for a
Now, in other cases, at the other extreme, there are
reactions that simply don't occur, even though there is a chemical
potential for those reactions to occur, they don't occur over very long
This goes not -- this ranges not only from the natural
system where there is metastable glass at Yucca Mountain that's 12
million years old to metals that are predicted to persist for -- in an
oxidizing environment at Yucca Mountain for tens or hundreds of
thousands of years. There are some reactions that are so slow that
chemical equilibrium is an inappropriate approach and one needs to
address various aspects of the kinetics.
At one limit of that, one assumes that certain parts of the
system are inert. In our ambient system models, I think it's reasonable
to consider that quartz is inert in the ambient system. The reaction
rates are -- it's a dominant component of the natural system, but the
reaction rates are just too slow for it to be an important component of
the chemical evolution of the ambient system.
It's persisted there for 12 million years, but the waters
are ten times super-saturated with respect to quartz.
Other parts of the system do react kinetically. The glass
in the ambient system at Yucca Mountain has altered substantially to a
secondary suite of minerals, zeolites and smectites. There is secondary
calcite in the system.
The introduction of a thermal field, the introduction of
coupled thermo-hydro-chemical processes in the repository will induce
these kinds of alterations to the environment in which some kinetics
will have to be regarded.
MR. WYMER: Can I --
MR. FAIRHURST: Had enough?
MR. WYMER: Can I assume from that that you are --
MR. HORNBERGER: You deserved it.
MR. WYMER: -- that you are comfortable -- my question is,
are you comfortable.
MR. MURPHY: Well, there are -- I'm comfortable with some
things and very uncomfortable with other things.
MR. WYMER: That's what I would have said from what you just
MR. MURPHY: And I will tell you that some of the things I'm
very uncomfortable with are the properties of the secondary phases that
ultimately will control much of the source term at the repository.
MR. WYMER: I couldn't agree more.
MR. MURPHY: I'm very uncomfortable with that at the
present, and we have an experimental program that was initiated under
the NRC research program many years ago and terminated when research
support for this program was terminated.
It's now reinitiated, and so we're trying to make progress
there again. I'm uncomfortable with the virtual complete neglect of the
changes to the hydraulic system due to chemical processes in the near
field in performance assessments, and that's an area in which we're
making an effort through our coupled thermo-hydro-chemical modeling to
address those things we're most uncomfortable with.
MR. WYMER: Great. Okay, fine. That takes care of that.
That's right. That's what I would expect as an answer.
One of your viewgraphs says the near field transport
phenomena have small effects on performance. Of course, that's in the
context of the current design and everything else.
MR. MURPHY: And in the context of the current performance
MR. WYMER: There are ways that things could be changed
where that would not be true.
MR. MURPHY: That's true. In fact, DOE has done some
sensitivity analyses in which they have increased the porosity of the
invert materials to a substantial degree, such that the flow transit
times coupled with the sorption on the invert materials actually can
lead to a significant retardation in the near field.
One can imagine some cases where near field chemical effects
could effect transport potentially, but in the present generation of
performance assessments, particularly in NRC's present generation of
performance assessment, near field transport issues are not.
MR. WYMER: Which is an artifact of their particular --
MR. MURPHY: Well, it's part of the process of abstraction.
MR. WYMER: Yes. Another viewgraph, you say some near
field, new near field coupled technical issues are relatively unstudied,
and you say the effects of backfill on seepage. I just wondered why you
didn't add chemistry.
MR. MURPHY: I think that that's a -- I should add it. I
think that the effects of the backfill on the chemistry will be -- and
I'm speaking very, very spontaneously now.
MR. WYMER: Please do.
MR. MURPHY: I think that the chemical effects will be
similar or that we'll be able to handle them in a manner very similar to
the way in which we are addressing the chemical effects in the ambient
system, because the backfill is envisaged to be crushed tuft or maybe
silica material, and I think that in general, the tools that we've
developed over the years to characterize the natural rock system will be
very appropriate for addressing questions about the chemistry of the
backfill, unless someone comes up with some exotic backfill, which I
MR. WYMER: And on the progress and accomplishments
viewgraph, you said planning for laboratory and theoretical studies of
cementitious material for secondary alteration phases.
That means the cementitious materials are the kinds of
things that would form due to the stuff that's coming in and the things
that dissolve and the reactions. It's not cement contained in the
drift. These are cementitious materials.
MR. MURPHY: In fact, what I intended by that slide was to
address two related, but somewhat different sets of studies, one in
which we're looking at the stability of uranium minerals, and one in
which we're looking at the properties of the phases that make up what
would potentially be a concrete or cement, as a function of temperature.
MR. WYMER: That's a little bit broader than I picked out of
that, but that's fine.
MR. MURPHY: I put those both in the same line, but there
really are separate studies involved.
MR. HORNBERGER: Is this because you anticipate that the DOE
design of a month may return to concrete liner or is it because you
already had the studies started and you want to see it through?
MR. MURPHY: Both. We already had those studies started. I
think it's reasonable to see them through and that I am not convinced
that concrete won't come back.
MR. WYMER: That's all I had.
MR. LESLEY: This is Brett Lesley, from the NRC staff. The
other aspect of it is if they're going to go for a rock fall, as
suggested, they would build a fully grouted rock fall. So one of the
actions I've taken is to make sure that the center's work on the
cementitious materials is relevant both not just to the cement liner,
but also to grout material.
MR. WYMER: Yes. One of the questions that I had was just
how much cement is there in the drift in the new design, but that's
another issue. We won't talk about that here.
MR. LESLEY: Actually, I would like to follow up on that.
In fact, what we have noticed, that in terms of alcove seven and also in
the drift scale heater test, in fact, Debra Hughson made this
observation, is that the dripping that occurred even under ventilation
was only under the rock fall.
So when she was wheeling along on the floor, she noticed
that there was concrete type of deposits --
MR. WYMER: Stalagmites.
MR. LESLEY: -- on the floor and those were associated with
the dripping of a cementitiously grouted rock fall.
MR. WYMER: I see. Okay.
MR. MURPHY: This is Brett Lesley speaking, by the way.
MR. WYMER: Yes, thanks.
MR. GARRICK: Bill, this committee visited Europe last week
or last year and had discussions with the Germans and the French and the
Swiss and what have you, and they made the frequent observation that the
United States is living up to its reputation of continuing its search
for the most complicated possible design to dispose of high level waste.
Of course, we've always been preaching that one way to
decrease the uncertainty in the performance assessment would be to move,
wherever possible, in the direction of simplicity.
One way to eliminate the uncertainty of a specific phenomena
or reaction is to eliminate it.
Are you convinced that the design changes that are being
proposed -- and you have delineated some advantages and disadvantages
here. Are you convinced that these design changes are indeed moving in
the direction of simplicity as far as the overall design is concerned?
MR. MURPHY: I wouldn't say that I'm absolutely convinced,
but I'd say I'm persuaded in that direction. I believe that the design
changes have simplified things with regard to the thermal loading in
particular. I'm not sure that the backfill or the drip shield
simplifies things in that regard.
MR. GARRICK: Are you -- one of the members of this
committee, of course, is very high on changing the natural setting in
such a way, engineering the natural setting in such a way that we get
some real advantages from it in terms of eliminating some of the
What is your perspective on that?
MR. MURPHY: I'll be frank. My perspective is that Yucca
Mountain is a mountain and it's hard to change a mountain.
MR. SAGAR: As simple as that.
MR. GARRICK: You can't make a mole hill out of a mountain?
MR. SRIDHAR: Dr. Garrick, can I take a shot at your
question regarding simplicity and complexity?
MR. GARRICK: Yes.
MR. SRIDHAR: Perhaps give a corrosion engineer's
perspective. Again, just for the record, Narasi Sridhar, from the
I also have had discussions on various MRS meetings with
Europeans and Japanese regarding complexity of design and simplicity of
design, and most people, when they refer to simplicity of design in
terms of waste package only and not addressing other issues in terms of
repository, would say, well, copper and steel are very simple materials
and very simple designs, we know archeological artifacts of steel and
copper have existed, so why can't we design.
And it is not, in my opinion, that simple. From a corrosion
and electro-chemical perspective, copper and steel, especially in an
oxidizing environment, like Yucca Mountain, are the most complex
materials to understand. When we initially did have copper in the
program, and when we started looking at it, we felt that as opposed to
nickel-based alloy, like C-22, the number of environmental parameters,
chemistry-wise, that we need to understand to characterize the behavior
of copper is much, much more complex than that of something like alloy
C-22. Because copper, for example, if you have five different species,
bicarbonate chloride, sulfate, they all have an interactive effect.
So you have a five-term interaction to characterize the
copper behavior, but as for an alloy like C-22, the one thing you need
to most necessarily understand is chloride concentrations.
So it is actually more -- maybe metallurgically more complex
than copper, but electro-chemically, it would be actually a simpler
So what some may consider to be simple and complex design
perhaps from a geochemical perspective may not be the same from some
other perspective. That's all I want to say.
MR. GARRICK: Stay here a minute. Do you think that there
is opportunity for simplicity with respect to something like, for
example, welding? As one speaker of a meeting several times ago
indicated, there's 30,000 feet of welds in this.
Based on what we heard earlier today, the welds could be one of the
vulnerable points of the integrity of the containers.
MR. SRIDHAR: Yes. I think this is something, maybe I'm
stealing Gustavo's thunder, but basically we are addressing that. The
main concern is not with all the 30,000 meters or whatever it is of
weldments, because most of the weldments that are premanufactured will
be annealed, and we are not concerned as much about those weldments.
What we are concerned about are the closure welds, which
cannot be annealed because of the temperature limitations on the
So those are the ones we are focusing on. Having said that,
what we are finding, and this is something we are continuing our
experimental work to confirm, that the weld corrosion resistance may be
lower than that of the unwelded material, but it may not be sufficiently
low to have a big difference in the performance. That is something we
need to complete our determination.
So I think we are simplifying the problem a little bit in
terms of focusing on only the welds that are of importance, and we feel
that we have a way to handle the effect of weldment on performance. I
think Gustavo will talk about that this afternoon.
MR. GARRICK: Yes.
MR. FAIRHURST: With regard to C-22, I think Shumaker has
suggested that there is a temperature below which a lot of these issues
will go away, right?
MR. SRIDHAR: Right.
MR. FAIRHURST: And the same with humidity. Now, as I
understand, with C-22, humidity is not that serious a problem.
MR. SRIDHAR: In the sense that in a --
MR. FAIRHURST: The other design, with the carbon steel on
MR. SRIDHAR: Well, the alternate wet and dry environment is
not as serious a problem with C-22 as with carbon steel. You have a
humid and then a dry environment.
MR. FAIRHURST: And as far as an engineering designer, would
you say that if we can keep the humidity down below X percent and the
temperature below X degrees, that will simplify these problems?
MR. SRIDHAR: Right.
MR. FAIRHURST: What are those numbers for carbon-22?
MR. SRIDHAR: Well, for the C-22, the current assumption
that DOE is making is 80 degrees Centigrade is the limit.
We are examining the technical basis for it. So far, our
experimental work has shown that that is a relatively conservative
temperature limit, lower, for localized corrosion.
MR. FAIRHURST: They bring it down to 80.
MR. SRIDHAR: Yes. I think they have a range between 80 and
100 that they sampled and our experimental work has indicated, again, I
may be talking in advance on what Gustavo is going to be saying, that
that is a relatively conservative temperature as far as what we have
It's something that we would like to continue to examine and
welding the C-22 will tend to affect that temperature, so that's
something we are examining, too.
MR. FAIRHURST: And in the absence of water, you greatly
simplify the problem. These are specifics. Like I want to come to Bill
and say give me some specifics that you say you can do that, and I don't
have a problem with it. That's, I think, what the intent of John
Garrick's question, to say what is the simplification.
MR. MURPHY: I think I would like to say that I answered the
last question rather glibly about effects on Yucca Mountain of
engineered materials and design and simplicity. I'm saying that the
question was directed in particular towards the oxidation state of the
Clearly, when you take a massive amount of highly reducing
materials, like spent nuclear fuel and the metals of the containers and
the steel sets and so forth, and you put it in a thoroughly oxidizing
mountain with a gas phase that's air and no reducing capacity at all, it
will generate huge potential gradients that will, over a very long time,
have to reequilibrate and will reequilibrate.
And I think that where circulation will occur in the system,
there will be oxidants reintroduced through that system and that is
where corrosion will occur and releases will occur and transport will be
So it's hard to change a mountain, but nevertheless, I think
putting a repository in there will lead to some big effects.
MR. HORNBERGER: Well, you just said that by putting the
stuff in, you're changing the mountain. You're changing the
MR. FAIRHURST: Bill, in an earlier part, one of the rock
mechanics concerns is the long-term degradation of the joint strengths.
Now, is it feasible, from what you're saying, that you could actually
have a long-term augmentation of joint strengths by precipitation along
those joints, so that degradation may not be as serious a problem?
MR. MURPHY: I think, absolutely, in parts of the system,
there could be strengthening of rock joints.
MR. FAIRHURST: That's what I would look for. I know that
this is -- my comment about your paper was complex as hell, because
that's the message I got. You're telling us what is the truth, of
course, but I'm saying out of that complexity, what can we find that's
either something we've got to design to avoid or something that we will
There are some codes right now, mechanical,
thermal-mechanical codes which will allow you to consider the
possibility of changing things like cohesion, joint stiffness, et
cetera, as a function of time, over extended periods of time. They're
not in a very advanced state, but the idea is to include chemical and we
need good chemists to come and sit with us and say, well what weight of
strengthening are we going to get or are we not going to get.
And this has considerable implications for doing it in a
thermal environment, you're clamping the joints at the same time, does
the process accelerate or decelerate under a -- it's like when you're
welding, slam them together and then weld them.
MR. MURPHY: Thank you for another complexity.
MR. FAIRHURST: It's a complexity which is either beneficial
or counter. So then you sit down and say, well, you design to include
it or you design to avoid it.
MR. MURPHY: That's a complication and it's one that I think
could be very interesting to pursue at some point. I have a counter
question. I still have five minutes for discussion, I think. I think
that the most vulnerable phase at Yucca Mountain to alteration is glass.
Much of the mountain is volcanic glass.
During its 12 million or 14 million year history, that is
the phase that has been altered. Much of it has been changed from glass
to zeolites. Is that going to have an effect on the mechanical behavior
of the system?
MR. FAIRHURST: How fast is it happening?
MR. MURPHY: That's a big -- in the natural system, that's
somewhat of an open question. There are certain people who say that it
happened very rapidly during a relative -- on a geologic time scale --
shortly following the eruption of the volcanic materials during a sort
of thermal cooling event, which may, in some regards, be comparable to
the thermal conditioning due to the repository.
MR. FAIRHURST: Well, I think it's a good thing to explore.
MR. HORNBERGER: Bill, in terms of your lab and theoretical
studies on secondary alteration phases, when you're looking at these
uraneal minerals, you're also looking at the incorporation of important
things like neptunium.
MR. MURPHY: Yes, indeed, and that's a very hot topic right
now, because of particular speculations that that may occur because
laboratory data from Argonne have showed that in vapor phase tests,
there is neptunium incorporated in secondary schoepite that may be ten
times its concentration in spent fuel relative to uranium.
There are crystallographic arguments that there's space in
these secondary uraneal phases for maybe a relatively broad suite of the
second -- of the minor radionuclides.
Our tests are designed ultimately to look at issues of what
we call co-precipitation. A dominant phase will be precipitated. That
will be a uraneal phase, it may be schoepite, it may be uranifane.
We're choosing to study uranifane. It is the dominant secondary uraneal
phase. The dominant secondary phases at Yucca Mountain are calcium
I think that uranifane is a likely dominant secondary phase
in the long-term for Yucca Mountain. So that's one of the reasons we
have chosen to study it.
We're considering and designing experiments to look in
particular at neptunium co-precipitation in uranifane.
MR. HORNBERGER: So you're looking at the co-precipitation
and then you're subsequently going to look at remobilization, kinetic
controls on dissolution?
MR. MURPHY: I'd love to, yes, and that sounds -- and people
-- that sounds like fun. People have talked about that, in fact.
That's discussed in the performance assessments, that in particular, in
regard to the Argonne experiments, which are very long-term experiments,
five and ten and 15 year experiments, and very short experiments
compared to the evolution of the near field environment at Yucca
What they see and what one can envisage is initial precipitation of a
relatively less stable phase, maybe schoepite, which may be less
crystalline, have less crystallinity, more capacity for
co-precipitation, a secondary species, a subsequent recrstyallization or
ripening, crystallization of more stable phases which may or may not
incorporate the secondary components as easily.
In some cases, there may be crystalline sites in the more
stable crystalline phases that better accommodate the trace
radionuclides. That's a problem that I don't think that short-term
experiments or crystallographic studies or natural analog studies alone
can address, but I think it's a good one to be looked at.
MR. WYMER: Do you have the capability here at the center to
do experimental work of things like neptunium and technetium?
MR. MURPHY: Yes. We have done and are doing experiments
with neptunium and plutonium and technetium and other things.
MR. WYMER: Just at the tracer level.
MR. MURPHY: At tracer levels. The institute has hot cell
facilities that we at present have not used for -- in support of the
Yucca Mountain work, to my knowledge.
MR. WYMER: I think we've wrung you out.
MR. MURPHY: I'm willing to stay longer.
MR. GARRICK: I think we're in very good shape. I want to
thank all the presenters for following the rule of allowing some of the
time for questions and discussion. It was an excellent job, both today
All right. I guess now it's time for us to adjourn for
lunch and as I understand it, we're going to have the tours of the labs
following lunch. Where do we rendezvous after lunch?
MR. SAGAR: We would be with you at lunchtime and would
direct you to the labs
MR. GARRICK: And we'll be back here at 3:15.
[Whereupon, at 11:45 a.m., the meeting was recessed, to
reconvene at 3:15 p.m., this same day.]. A F T E R N O O N S E S S I O N
MR. GARRICK: The meeting will come to order. We are now
going to hear about container life and source term. Please introduce
yourself before that.
I should note that Ray Wymer will leave the discussion on
this topic. Proceed.
MR. CRAGNOLINO: Good afternoon. My name is Gustavo
Cragnolino. I am a member of the corrosion science and process
engineering element. I am going to talk precisely about containment
life and source term.
Here I have listed the technical contributors in this part
of the program. You met, down in the laboratory, Mr. Brossia, Mr. Pan,
Mr. Pensado, who currently is a member of the PA group, as well as Mr.
Mohanto, Narasi Sridhar is the manager of our group, and this shows
again integration in our activities because Jay Weldy is a member of the
PA group, Dave Pickett is from the near field environment, and both
Charlie Greene and T. Ahn are members of the NRC staff.
Charlie Greene was visiting with us in the technical
exchange, something that was mentioned by Wes yesterday.
Now, the first slide is something that you have seen
already, and I refer to the flow-down diagram for the total system
performance assessment. I am going to concentrate on the issues related
to the subsystem, the engineering barrier system.
This is a description in terms of the analysis in total
system performance assessment, but the engineering barrier system, from
the physical point of view of the component, consists of a waste package
and other elements of the engineering barrier, such as the options that
DOE is planning, drip shield, backfill.
Here you have the so-called four elements that reflect the
integration in between differing aspects of the key technical issues in
the containment life and source term. You can see these much more
clearly in your handout.
I will show you in the next slide the correlation that
exists in between these keys that I've used in terms of the total system
performance assessment. One is related to the waste package corrosion
in terms of humidity, chemistry and temperature, and as related to the
KTI that deal with effect of corrosion on container lifetime.
The mechanical disruption of the waste package is related to
the effect of materials stability and mechanical failure on container
lifetime. That reflects the interaction with the near field chemistry,
the quantity and chemistry of water contacting the waste package and the
waste form is important in terms of the corrosion of the container
lifetime, but also in terms of the rate of degradation of the spent fuel
and the high level waste.
Last, the fourth of the key ISI included here is the
radionuclide release rate and solubility limits and these are in some
ways clearly related to two of the other key technical issues,
sub-issues that I mentioned before, the rate of degradation on the spent
fuel and radionuclide release from the spent fuel, and the rate of
degradation of high level waste glass and the release from the glass.
We have two additional sub-issues that are not included or
listed, and one refers to the criticality inside the waste package and
another one makes provisions for alternate material selection at the
site for the engineering barrier system, that in some way feeds within
this picture of the other pieces. It's a currently integrated sub-issue
because reflecting integration in between different KTIs.
From the point of view of performance assessment, we can
come out to what we consider insight in potential risk for the
performance of the repository. I listed here what we consider the most
important coming out from the performance assessment.
One is the importance of initial failures that has become
dominant as a consequence of the design change. The performance of the
container materials has been improved by the choice of alloy C-22.
Improved in this context means there is a low -- that it is less prone
to localized corrosion. Therefore, it's very important, the passive
corrosion rate in the life of the container, but DOE will still consider
variations to the VA design.
C-22 was introduced for the VA design as a way to improve
the performance, but they will consider an alternate design and the
first choice was to do the revision of the VA design, removing carbon
steel as the outer barrier and including, as an outer barrier, the alloy
But in order to provide integrity, structural integrity, the
new design that is currently considered, as you have probably seen in
design number two, design alternative number two, is the use of a
stainless steel, in order to provide the necessary mechanical
This design could be altered with additional features that
are part of the EBS, like the drip shield or even the option of the
One of the important things in this type of material is the
effect of fabrication process, that now we are starting to deal with,
since it's much more clearly defined that the DOE is considering it for
the license application.
In this context, the importance of the near field chemistry
becomes clearly up a range, and I think that near field chemistry is a
real narrow term, because here we consider also part of the environment
as the definition of the temperature that we are going to have in the
container itself or other environmental factors, like the application of
a stress or thermal loads.
It should go beyond purely the chemistry, even though the
chemistry is extremely important in terms of corrosion process.
The importance of penetration rate is related to release and
this is very much related with the way that water comes into contact
with the waste package. And I have included in this list the aspect of
cladding, because this affects the performance in the analysis that has
been done by the DOE, due to the relatively high rate of release that
they have from the spent fuel.
The important thing, and I am going to mention this later
on, is that we can deal with the corrosion problem of flooding material
with the same approach that we use for container material.
And, finally, the last point to stress is that once the
waste package are penetrated or breached, internal environment is very
important in terms of release.
These are important factors in terms of performance
assessment, and I want to emphasize.
In addition, you have to realize that this is going to be
affected by the relatively high variation field and, therefore, will
change the composition of internal -- it will happen on release, as well
as the high corrosion being emplaced inside the waste package.
Let's go now to some of the points that I mentioned before
and the specific aspect of initial failures. Initial failures, as was
mentioned before very briefly by Sridhar, is dealing a very different
way in DOE TSPA/VA and in the TPA code, because in the case of the DOE
-- essentially, in the case of the DOE, faulting and seismic effects are
included, while the TPA has a separate consideration of the faulting and
seismic effect on performance of the waste package.
There is a clear difference in the probability of failure
release effects, in which DOE is assuming one in 10,500 waste packages,
the order of ten-to-the-minus-four, and the initial failure time for
this type of initial failure of 1,000 years, while in the TPA code, the
average failure range from ten-to-the-minus-four to
ten-to-the-minus-two, and for subarea, and the assumed failure is at
If you are interested, I can go into more detail in the type
of associated problem related to initial failure definition later on.
But I will proceed and show you a compilation of performance
calculations by using, in the TPA 3.2 code. The initial DOE failure
If we take the value of the TPA code, the NRC, where you
have this type of distribution for the expected dose, by using the value
assigned by the DOE, obviously the expected dose increase for almost an
order of magnitude, which essentially has been done here in order to
refer to a starting point, the initial failure at 0 EF for both cases.
This is an indication, even at 10,000 years, that there is a
significant impact related to the number of initial failures.
The main contribution of these analysis that was presented,
the last thing that is changed that we have with DOE in performance
assessment, is that the initial failure are essentially on spheres in a
related system application. They are using many type of application,
also fuel rod from the nuclear industry, and there is a difficulty in
defining the mechanism for initial failure, because they're all lumped
A very important point is that there is no relationship
between the type of effects and what is the process that could lead to
failure starting from the initiation effect.
And the effect of experience on initial failure rate is not considered,
and I think that's a very important point, the experience in the
construction of fuel for a nuclear power plant, has demonstrated very
clear that initially the rate of failure was almost an order of
magnitude larger than the failure rate, and this reflect the experience
that was acquired in process application and improvement in the
The following figure shows the effect of the container
material on the system performance. These are calculations conducted
with the TPA code in a different version. But there are two comments I
want to say here.
One is that here we have the fraction of waste package
failure as a function of time and it's very obvious that the choice of
alloy 22 produces significant increasing failure at a given time,
especially for short times, because the failure with alloy 22 in this
case appears essentially starting from about 10,000 years.
The difference in between A-825 and 625, not significant;
therefore, significant move forward from the point of DOE who made a
choice of alloy 22. And I want to make a point here. We never believed
that alloy 825 was a good choice of a container material. Therefore,
when we study our activity, decide to look at some preliminary work with
alloy 22 that was at the time a reference material.
We have a good database, but very limited at that time on
this material, that we couldn't propose it because it was not within the
scope of our regulatory system to the NRC. It was only a reference
material for comparison.
Alloy 625 came into the picture and finally alloy 22.
Another point I want to make for here is that we able by a
simple change of input parameters in the code to deal with the choice of
three different materials, because all these materials belong to a class
of material, the base alloys.
Therefore, the corrosion failure mode were essentially
identical. We used the same approach. When we got into the laboratory,
they put parameter -- their basic mechanistic bases were fair.
We did calculations for the waste package lifetime using the
TPA code and in the baseline case that was used for the analysis of the
TSPA/VA, we felt corrosion rate that we use would essentially was
derived from wrong data and data from the literature for 825. We didn't
have enough data for C-22 at that time.
And we came out with a medium for that time of 70,000 years,
assuming the base case, which we don't have, was in the material. It's
base matter. Later on, we start to come back experiment in the lab to
measure more accurate the passive density of alloy C-22, in which this
new data, we were able to demonstrate that the life could be extended,
as I show here in the solid plot, up to 60,000 years.
And it's interesting to illustrate the point that the design
the viability assessment, waste package in which you put the C-22
outside and the carbon steel, that was the material inside, essentially
exceed the same lifetime. These two curves that are here are obtained
with the same essential passive carbon density.
But illustrate one important point, that really what
provides corrosion resistance in the waste package is C-22. And this is
why DOE made a choice of getting rid of the carbon steel, that could
lead to certain problems in that sense that it would release corrosion
product and it will affect the behavior of the near field environment
and introduce a lot of uncertainty, and use C-22 as the alloy that
What is the main consideration has been the lifetime of the
container material, to decrease release rates.
In the following one, what I wanted to stress is some of the
limitation that we see in the approach of the DOE for one side, and, on
the other side, the lack of sufficient data. This is the previous
calculation that I showed for the base case.
This is the calculation which allow to modify corrosion
data, but by the way, this is data that was obtained in the laboratory
doing electro-chemical experiment that lasts for 24 to 48 hours.
DOE has conducted tests for almost a year, in some case, two
years, to measure the corrosion rate of C-22 and they come out with the
The expected life in this case will be of the order of
50,000 years. That is here. However, if they use what they have in the
TSPA/VA as a rate of corrosion for C-22, raise it on expert elicitation,
they are -- the failure on the container, and you can see that after
100,000 years, only 20 percent of the waste packages failed.
Well, for us, this is completely unrealistic and is a
problem of the compounding of the opinion of many experts, that the
spread of range of corrosion rates over an incredible large range of
corrosion rates. There is a clearer way to distribute the dose.
How is our approach? Our approach was, first of all,
consider failure modes, corrosion, stress corrosion cracking, hydrogen
embrittlement, mechanical failures, according to the class of materials.
We have carbon steels on one side, nickel-based alloys, titanium alloys.
And on top of this, to evaluate a wide range of
environmental conditions, in terms of anion concentrations, temperature,
pH, redox potential, that are expected to contact the waste package, the
water in the waste package.
On the basis of these, to be able to develop abstracted
models for performance assessment that can be supported by mechanistic
models. And finally, to try to raise confidence on the performance
assessment by using focused experimental measurements of important
What are the important factors that affect the performance
of the so-called corrosion resistant materials? For corrosion material,
because this material, what they have is a surface fill and reaching,
most of the cases of alloys that we are considering, that provide
resistance to uniform corrosion, because corrosion rates can be
extremely slow. But over the life of these materials, this it the order
of -- therefore, this material could be extremely prone to localized
It exists through a very delicate balance, at equilibrium
with the media. The chemistry, therefore, of the water is especially
important and chloride is the main element that is able to break down
this passive thing and initiate localized corrosion.
And we come to this conclusion through experimental work in
which we demonstrate that of the various species that you have in the
ground water or in the repository, the only one that you have to
essentially concern as aggressive is chloride. Chloride, nitrate,
sulfate, and also important in thermal effectiveness of the system.
Another important factor is the redox condition in terms of
the corrosion potential, because it will be a design -- if particular,
if you reduce the wall thickness, the shielding effect that you have
already in the original VA design with the very thick carbon steel is
significantly reduced, and, therefore, you have a condition on the waste
A matter of concern I mentioned before is the microstructure
of the material in terms of the welding or eventually any heat treatment
that could be wrongly done, where you have 10,000 containers to be in
the heat treatment, and this could lead to problem. Because everything
lies on this, on the passive dissolution rate, and it's a very important
parameter to measure.
In the case of localized corrosion process, you have to
consider an active dissolution rate, this active corrosion rate has an
order of magnitude larger. And we need to include pit growth rate,
because of the geometry of the pit.
This is the methodology we applied. Some of this was
already described in the laboratory. I'm going to go very fast. It's
to measure corrosion potential, after calculation this from a basic
electro-chemical kinetic law, but doing experimental determination of
the repassivation potential as a bounding parameter for localized
corrosion in terms of temperature, pH and chloride concentration, and be
able to sustain this process of repassivation potential for localized
corrosion, to analyze the electro-chemical conditions for the stress
There is an additional requirement in terms of effects.
Later on, we can go over a little the localized corrosion rate, and we
can do some evaluation, from the experimental point of view, of the
effect of welding or thermal treatment in some of these critical
parameters, and combine all these approach in the use of parameters that
can be used in the TPA code, but size it to one point.
We are talking about, as was mentioned as a challenge here,
the long-term behavior that we need to predict. We have to have some
fundamental modeling of passivity in the localized corrosion process to
be able to, on the basis of a model and a fundamental basis, essentially
model existing -- we have to modify and adapt to get an understanding
and be able to consider factors that could affect the passivity in the
This is another way to look to the importance of repassivation potential
and the concept that is embedded as a criterion for localized corrosion
using the TPA 3.2 code.
I put this plot as a representation of the repassivation
potential as a function of the chloride concentration, and we have only
the lines describing the behavior of 825, 625, in order to stress the
improvement that you have with alloy C-22. Only a very high chloride
concentration goes to the critical concentration corresponding to this
calculation of chloride. You have a sudden drop of the repassivation
potential to low values that are typical of other alloys.
However, in all this regime, even at very high potential,
you will have various corrosion, as indicated by the open circles. We
have very high potential there, but you don't have localized corrosion.
This potential is really associated with the solution with alloy, but
not to localized corrosion. While the point, the cross-point reflects
This means that at about 95 degrees C, the point that is
mentioned here in this ledger, there is a window of susceptibility for
C-22. C-22 is not a immune to localized corrosion. But it's
susceptible very close to the volume of water, at a very high chloride
In order to understand better this type of process that
takes place close to the boiling point, what we have done is this, to go
beyond the boiling point, and we can see very clear here that you have a
sharp decrease in the repassivation potential precisely about 95 degree,
that this is even the case for one more solution, and for .5 more
solution, even though corrosion is not at the 95 degree in .5.
But you reach one more solution in between 95 and 120, but
you can reach temperatures about boiling, you have concentration of the
chloride, you can have a sudden decrease in the repassivation potential
for alloy C-22 and it's clearly a matter of concern.
But this is for the base metal. What happened with the
welded material? The welded material showed a significant drop with
respect to the base makeup, as you can see here. And the same can be
demonstrated for a heated material for 24 hours and even for 240 hours,
the behavior of this material cannot be considered as a highly resistant
alloy. It's an alloy that is similar to the other alloy that we
considered before in the program, but we can't consider this is a
bounding value that we have here for the purpose of making sure that
there is a significant trend in terms of repassivation potential as a
sensible parameter, and we don't anticipate that this condition can be
expected in any container.
But the one here, obviously, are potentially possible and
you can see the potential values are relatively low.
The next block shows the localized corrosion propagation
rate and unfortunately, from the point of view of somebody that wants to
do experiments in the laboratory, C-22 is highly corrosion resistant.
It's very difficult to repeat it in a way that we can measure. What we
have done is an experiment, as a model alloy, with 316L, in which we
have two -- it is not plotted to demonstrate, but this reflects
ten-to-the-minus-one, the peak growth is controlled by diffusion, but
what's interesting is the fact that we were able to collect field data
obtained for alloy 625, the second alloy chosen by the DOE, that shows
the rate that we are using for localized corrosion propagation in the
TPA code is in the right range of value.
It's interesting, however, to point out that in the TSPA/VA
code, the highest value of corrosion rate for localized corrosion is one
order of magnitude lower than this. But the most important point is
that the median rate is several orders of magnitude lower.
In any circumstance, C-22 falls within this window of
stability, the life of the container cannot be assessed on the basis of
this expert elicitation value. It has to be assessed on the basis of
To close this relationship in between, laboratory data and
the parameters that are important in the TSPA code, I want to bring your
attention to this table. This table is more clear in your handout.
Since the corrosion rate of C-22 is extremely important in
determining life, you can see, and this is something like -- it's some
stretch of the imagination, but this made the point, that there is
relatively high potential in the passive region with this chloride
concentration, there is 1,000 ppa in chloride, corrosion rate can be
calculated by using an electro-chemical approach through the measurement
of anodic current densities, and the life could be of the order of one
However, if the temperature reduced, and this is shown here,
up to 95 degrees C, the current density increases; therefore, the
corrosion rate increases, and the life comes to something that is
perhaps more understandable, in the order of 60,000 years.
As you go to bigger CDPH, and this is one of the beauties of
C-22, it's an alloy that is not affected in this passive behavior by pH
effects and essentially the rate is the same.
It's in this range, the life is reduced by substantial
amount of years, but you have to take this with all the uncertainty that
is related to this type of measurement and we can't estimate
Another point I wanted to make here is if we go to the same
pH as before, but we go to a very concentrated chloride solution, again,
it's very sensitive to the chloride concentration. It's more or less in
the same range. Finally, if we go to a higher potential, that can be
reached by the presence of a species, again, we have a rate of the same
order, I would say. But these rates are very comparable with the one
that we use in the TSPA/VA code.
This rate was used without experimental data, put in the
code. Now, we are more confident to use rate in this order, mean that
the life of C-22 will be extended, if C-22 behaves as a passive matter.
But if C-22 becomes susceptible to localized corrosion, we cannot
predict that the life is this order.
And I want to point out that, as I mentioned before, this is
the order of rate that we measured in a couple of ways in the laboratory
and are very similar to the rate that were measured by the Department of
Energy in the laboratory. Immersion tests, after one year, one year and
a half period, and this indicate how probable it would be that
electro-chemical, they need to measure very low corrosion rates.
At about 400, they're difficult to obtain; however, these
materials suffer passive dissolution. Nevertheless, we didn't observe
that these condition of dissolution or localized form of corrosion, but
there is another process that leads to the solution.
In summary, the approach used by the NRC and the center is a
flexible approach, because allow us to accommodate DOE design changes in
terms of material selection. Has allowed us to use new data to update
the models, new data obtained in the laboratory, and even data that is
obtained in many different conditions, but we have a way to connect this
because we can put everything in what we call a performance map. A
performance map in terms of electro-chemical parameters, like the
corrosion potential or electro-chemical potentials in general, and
environmental factors, like the chloride concentration, the temperature,
and eventually any other species.
On a positive note, I should mention that these approaches
are being seriously considered by the DOE. They are not making material
selection here. If we decide a methodology appropriate for long-term
prediction, appropriate for long-term prediction, the sensitivity
analysis has helped us to focus this type of a study and we believe that
the assumptions in the modeling are not unduly conservative.
What are the progress that we have in resolution of all
these issues that are so important in terms of the license application?
We have several components of sub-issues that have been resolved or
became moot with the post VA design and materials changes. For
instance, I want to mention the dry air oxidation. At one point in
time, we had a conservative dry air oxidation, brought up by an NRC
staff member, and he had more reason to believe that this could be a
potential detrimental problem for the carbon steel.
What we did is we solicit the expert in the field.
Professor Bob Jott, member of the National Academy of Engineering, was
willing to do an analysis of these problem. We came to the conclusion
that this was not an important issue for the material of choice, and we
resolved this with the DOE. It was a design change and it was not
necessary to pursue it anymore.
As well as the case of aqueous corrosion, wet or dry cycles
of carbon steel, because they're changing the design made this issue
We've resolved, coming to the same conclusion regarding
thermal embrittlement of carbon steel. It's unfortunate that we have to
put this effort at that time in this issue, it became moot. But this is
life. And we deal with the issue of galvanic coupling effects and we
resolved this, and, again, ceramic coating has been a problem that has
been out of the picture due to the choice of the new design.
However, we cannot expect full resolution of all the issues
until DOE makes a final selection of the design, complete their
laboratory tests, and, above all, make a detailed plan for performance
Because here is one of the issues. We don't believe that
the DOE will be in the condition at the time of the license application
to provide sufficient data. They will provide minimal data, maybe they
provide necessary data, but we doubt that they will provide sufficient
Therefore, the issue of the performance confirmation period
is very critical for the license application process. Very critical.
We recognize that we have to complete the study of container
fabrication effects. How I am doing on time?
We recognize that we have to complete the study of the
container fabrication effects. During the last several years, we were
put in the situation that we have to defend not to do a status of welded
material and we didn't try because there was not clear decision from the
part of the DOE in which direction to move.
Now we know the material. We are following very closely
what we are learning and we can cross this issue. We have to consider
what is the most important aspect of the alternative design, the drip
shield, we have a concern with the drip shield made of titanium-16, is
apparently the alloy of choice. We have started an experimental program
in this direction. But, again, it's unfortunate.
The Department of Energy, the only alloy at the present time
is another type of titanium alloy, that is not the material of choice.
And this made very clear that they will have to use data in the
literature that is not so abundant and rely a lot in the performance
period. We need -- and this is a very critical issue -- better define
the near field environmental condition of the waste packages. This is
an integrated activity. Ron Green described in some detail what we are
doing in thermal effects on flow in relation with the near field, and I
say that this will be extremely important to define the conditions that
may lead us not to look at less corrosion of alloy C-22.
If we can't come to a clear conclusion that concentration
plus the saturation of chloride in the condition of the repository, C-22
is going to last much more longer than if susceptible to localized
And when I talk about C-22, I'm using a very broad term,
because what you're going to have in the repository is a variability.
There will be containers that perform according to what is predicted in
the laboratory on the basis of experiments, but will be in some places
of malfunction associated to wrong heat treatment that were not detected
and so on.
An important issue to continue the discussion with DOE in
terms of methodology, because this is the part that we are influence
them. Not in selecting the material, but in to try to use approach that
could be defensible. It has not to be our approach, but has to be
defensible. And the necessary data for assessing corrosion and
mechanical failure, as I mentioned, we have an Appendix 7 meeting next
week and future interaction, and this the point that I want to confess
that it is necessary to identify tools, techniques and areas of
performance confirmation testing, because this is going to become an
important issue to be seen at the time of license application.
Currently, DOE claim that in the drift heater test, they are
doing measurement, but they are doing purely and by the length and
measure emplaced on top of the drift. And what they have in term of
corrosion is a group of coupons exposed close to the waste package and
their conditions are not heat transfer conditions and without any way to
assess the behavior as a function of time. It's a matter of picking
them from time to time, if they are able to go there, and examine, they
cannot make any type of correlation that could be related very easy to
what this experimental work going on in the laboratory.
And these are a matter of concern and this is the reason why
we believe that it's extremely important to continue these discussion
with DOE in term of approach that we are going to use for assessing
corrosion and mechanical failure and safety in terms of the performance
This is -- if you look in the VA, the description of the
part, where it was supposed to be the license application plan, it's
extremely limited. Extremely limited.
I think that we are in good condition to provide support to
the license activities, the activities related to license application.
We are updating IRSR in terms of these changes in design on the basis of
data that we are obtaining in our laboratory and trying to make very
good use of what is available in the literature, because we believe that
the DOE is going to use a lot of data from the literature and we have to
evaluate these data.
And even though we all know, because we're publishing
papers, that their papers are peer-reviewed, they are peer review papers
and there are peer review papers.
MR. WYMER: Absolutely.
MR. CRAGNOLINO: We want to make sure that these fit with
this approach in terms of this performance model. It's very appropriate
to describe this domain of failure for materials in terms of
environmental and electro-chemical.
I'm open to any questions. I answer or I had the idea in
mind that I will tell you ask me question along the top, but I know I
MR. WYMER: I'll start off. I have a couple. I assume that
from the lack of mention of them, that you're not putting a whole lot of
emphasis on radiolytic effects on corrosion or biological effects.
MR. CRAGNOLINO: Let me separate the two issues. WE didn't
put -- at one time, we were giving thought to this, when they originally
decide with this thin wall container in a bore hole. By later on, when
the design evolved to a much more robust type of container, with thick
wall, we consider that the issue of radiolysis was not so important.
Now it's coming into the picture again and we are going to pay attention
Nevertheless, in the past, we have the literature and we
came to the conclusion that a dominant species produced by gamma
radiolysis was hydrogen peroxide. It's the dominant species, that they
form at a steady-state concentration, because they aren't radical, but
are not that stable. Essentially, it's defined by the concentration of
hydrogen peroxide, and we did some experiment with hydrogen peroxide to
measure the evolution potential and the corrosion potential, and we came
to a very interesting conclusion that particularly important in the case
In the literature reported that this type of alloy are not
very susceptible to the effect of metabolic product produced by
bacteria, and this is a point that is important to clarify. Many people
talk about metallurgical effects, thinking that the backfill matters.
That is not true.
What happens is that the backfill uses metabolic products
that change the local composition and in some case, when you produce a
film, the environment inside this could be significant, and this would
be extremely detrimental.
Nevertheless, look like this, high nickel, high chromium
alloys are extremely resistant to microbial influence corrosion. It is
an issue that we are in the condition to look, but it's not in the
priorities. It's in the list below the line.
If we have an indication that this becomes important, I
think that we are in the condition to address this issue, at the level
of having enough information to evaluate what DOE is doing, because
Lawrence Livermore is working very actively in this field.
MR. WYMER: Okay. Another tact now. You have one viewgraph
that caught my attention. Initial failures, TPA versus TSPA/VA, and the
second bullets on each of them were the assumptions of the failure
probabilities, and they're so different between DOE and NRC. I wondered
MR. CRAGNOLINO: It's one order of magnitude.
MR. WYMER: I'm sorry?
MR. CRAGNOLINO: One order of magnitude different.
MR. WYMER: Right.
MR. CRAGNOLINO: As I mentioned before in this presentation,
that they did the technical exchange with the DOE and I think this
counts from the fact that although implicit in the definition that we
follow for the initial failure are rather similar, what they are doing
is that they compose the probability with the assumption that it's a
double-wall container. Take data that exist for boilers and multiply
the probabilities and come out with this value. This is the rationale
In our approach, we were completely different, was to do --
at that time, we didn't know that the initial failure would play such an
important role in the dose at 10,000 years, for the reason, field
engineering evaluation that was done here in the center in '94 and it
was evaluated, but we also evaluate a steam generator tubes and
partially some boilers, and come out with this number. That was an
upper bound. We'd rather be conservative in this case and use an upper
This is something that we are going to -- it is not
satisfactory for us because it has not been a basis. DOE knows that
their approach has no technical basis either and they are trying to look
for an approach that is more justifiable.
And the approach is something like consider that we have some type of
effects and how can we -- these type of effects or initial flow has to
be evaluated also in terms of the level of the -- from non-destructive
MR. GARRICK: Given the importance of initial failures, is
there something that could be done in the operational phase -- that is,
in the pre-closure phase -- that would bring the two -- give us greater
confidence in some assumed value for such failures?
MR. CRAGNOLINO: I think that an effort has to be done
earlier than that. I think that this is something that we have in our
-- Sridhar can give more detail to you. I assume this is something that
we are planning to pay more attention and I think that the DOE has to
Unfortunately, valuable data could come from the
manufacturer of these type of component, and they don't provide this
data. When we get data, for instance, behavior of boiler material, it
will be something more close, even though the container is much more a
passive system with respect to a boiler, and it is combined because of
pure operational failure. It's not necessary what we try to consider
here in initial effects. We have to find a way to discriminate with a
more mechanistic basis.
MR. GARRICK: What about something like an on-site elaborate
MR. CRAGNOLINO: Right. There is no doubt that has to be in
place. Yes. We start from the point of view that there should be a
very elaborate place of non-destructive examination prior to put any of
these containers in the emplacement drift.
MR. GARRICK: But you don't think the inspection process can
be a basis for bringing some rationale to an assumption of a failure
MR. CRAGNOLINO: Well, I think that when the DOE are going
to do more and more --
MR. McCARTIN: This is Las Vegas. Can you guys hear us?
Could you stand in front of the mic, Gustavo?
MR. CRAGNOLINO: Yes, I'm sorry. I tend to move around.
I'm sorry. I think that will improve that process, but the -- you know,
even in cases in which a non-destructive examination technique has to
improve a lot, like in the nuclear industry for a steam generator tubes,
the fact that you have still failures, and some of them are real initial
failures, manufacturer defects and so on, many of the case that we have
to anticipate we can maybe bound better these with an improvement in
non-destructive examination technique.
But on the other side, let me tell you, this is not such an
extremely complex technology nowadays. What made it seem a little bit
complicated in this particular case is that this corrosion resistant
alloy in the design are brought to the maximum currently used and I
think Darrell can correct me if I am wrong, but I think they are very
close to what the manufacturer can produce.
MR. DUNN: No. It's 20 millimeters a day for the maximum.
The current design uses a 20 millimeter thick C-22, I think the maximum
thickness is about 55 millimeters thick.
MR. CRAGNOLINO: For the 316L, that is going to be 55
centimeters, you are right.
MR. GARRICK: For the benefit of the court reporter.
MR. DUNN: Darrell Dunn, from the Center for Nuclear Waste
MR. WYMER: I had two more questions. One is sort of
trivial and the other has maybe a little more substance to it.
ON your table, uniform corrosion rates of alloy C-22 and
values used in the TPA 3.2.
MR. CRAGNOLINO: Yes.
MR. WYMER: I believe it's page 18, I think.
MR. CRAGNOLINO: Let me project it.
MR. WYMER: I'm not sure that you need it to answer my
question. But it's page 18. It looks to me like you have chosen your
data here to make -- to enable you to make a crude pass at deriving a
rate equation and also get an activation energy from it.
Have you done that?
MR. CRAGNOLINO: Not yet, but we are in the process of doing
this and really this is going to be input in the modeling effort for a
passive film. He's trying to look at the variation of what is called a
point effect model, have that type of model for passive film, and the
data, the one that you have in this table, is going to be used for this
purpose to precisely analyze.
The most important aspect that you see here in this table is
the activation related to the variation in corrosion. We need
intermediate data. Sure. We cannot build this with --
MR. WYMER: I was just curious how far you'd gone. Finally,
the last question is on one of your final viewgraphs, issue resolution.
I guess that's page 20. The last bullet, importance of performance
confirmation testing enhanced by the anticipated lack of sufficient data
at license application.
I don't quite see how you can get much out of performance
confirmation during the pre-closure period, because the conditions would
be so different from the closure period. I don't know how you -- really
what value the observations you make pre-closure will be to you. Can
you comment on that just a little?
MR. CRAGNOLINO: Yes. This is pretty much dictated from our
own approach. It's probably an extrapolation, however. An
extrapolation of our own approach.
I think and we believe strongly that if you have a solid
model, mechanistic model, and you have data, you can input your data in
the model, you're in the condition of doing a verification. And this is
the approach you have seen in the laboratory. We have a model for
localized corrosion. We create a condition in the experiment and we are
doing test in the laboratory now for almost 40 years. We have to remove
the sample from the vessel every 30 days and we put this back and we
establish the condition.
We have done this for 40 years, a very limited period of
time with respect to performance of the repository. But this gives a
lot of confidence that our approach in terms of model and the previous
data that we obtain in short-term experiment is valid for this type of
Now, if the DOE use the same approach or at least another
approach of this type, until the period in which they start to put the
things in the repository, we have plenty of time to get good data from
other things and so and later on go to the performance under conditions
that are different, you can predict later on, if your models are
appropriate, you can predict other conditions.
You can use activation energy, as you mentioned before. You
can use parametric equation on the basis of a model that could you lead
you to understand variation on environmental effect like chloride.
MR. WYMER: So the value of your confirmatory testing is
directly related to the validity of your model. So you really need to
MR. CRAGNOLINO: Yes. And I look at it from the very narrow
point of view of container lifetime. I don't want to go beyond the
boundary of what I know as the subject.
In other areas, it would be other type of complications.
But also have an important effect if the environmental condition are
well monitored, too, because this can compliment any assessment in term
But when I --
MR. McCARTIN: I didn't mean to cut you off. I thought you
were at a stopping point. Part 63 does require the performance
confirmation period all the way through to closure and the waste
package, as specifically mentioned in the rule, has to be tested and
they have to do it under conditions that are representative of what's
going to occur at the mountain. So that there should be a -- I mean,
that's starting at construction, all the way through the waste -- to
So you're looking at at least 50 to 100 years of performance
MR. WYMER: But it does seem, if that can't be directly
related to a model, you have a lot of confidence in, it doesn't really
give you much confidence in an extrapolation of the future. So the
model is extremely important, the mechanism is extremely important.
MR. CRAGNOLINO: Yes.
MR. WYMER: That's all I have.
MR. CRAGNOLINO: We are doing -- it was mentioned by Bill in
his presentation, that these are serious technological challenges.
MR. WYMER: They are.
MR. CRAGNOLINO: And the spatial variation is not so big, so
important, I tend to believe, because in the last ten, 15, 20 years, big
project has been built that has something -- but the long extension --
you know, I think that the spatial variability can be handled. The
long-term provision, nature has an expected way to behave and who knows.
But we can do the base.
MR. HORNBERGER: Just out of curiosity, your opinion. If
the repository were designed such that the temperature never went above,
let's say, 80 degrees C, would that enhance or detract from the
MR. CRAGNOLINO: I think that for this choice of alloy, it
increases. Sometimes, when you answer these type of question, you don't
consider other ranges. For these type of value basically will be
beneficial, it will be beneficial in term of life.
MR. WYMER: Your voice is dropping a little too much.
MR. CRAGNOLINO: Will be beneficial in term of life. I take
the point. Will be beneficial.
MR. HORNBERGER: Okay.
MR. CRAGNOLINO: And what you can -- you should consider, we
didn't pay attention, because we concentrate too much in this regime of
more higher temperature close to the volume, it is a corrosion mode that
could become predominant in the other condition.
Basically, we don't expect, but if you tell me about, for instance,
other material, carbon steel or low alloy or stainless, would be much
more prone to microbial corrosion at an earlier time, and, therefore,
will not be a solution. It will be more of ability of water under such
condition and will shorten the life.
But it's something that has to be thought. I believe
personally, from what we were discussing, that go down from 85 to 60 was
a step in the right direction.
MR. GARRICK: Charles, do you have any questions?
MR. FAIRHURST: No, none. Thanks.
MR. WYMER: Thank you very much. It was extremely
interesting, to me certainly.
MR. CRAGNOLINO: Thank you.
MR. AHN: Good afternoon. My name is Tae Ahn, of the
Division of Waste Management, NMSS, of NRC headquarter.
Gustavo introduced container life and source term issues in
general. Also, presented path forward actions in general and focused on
the container life.
In my presentation, as a continuation, I would like to focus
on waste form status. And many colleagues who contributed to this talk,
Richard Codell, John Contardi, Charles Greene, V. Jain and Narasi
Sridhar. Also, I would like to acknowledge Paul Casado and Gee Cavaluno
for their later support.
As shown also in Gustavo's presentation, I would like to
highlight this flow diagram. We are in engineered system and under
engineered barrier, I will focus on the last one, EBS-4, radionuclide
release rates and the solubility limit. Gustavo addressed EBS-1 and 2
for container life.
The next slide shows the outline of the presentation. I
would like to focus on three different areas, spent nuclear fuel,
cladding performance, and high level waste glass.
For each element, I would like to address risk insights,
technical basis, and the progress and the current status, and the path
The next slide shows the spent nuclear fuel at degradation.
The risk insights may be described as follows. The dose is sensitive to
the types of spent fuel dissolution processes, and you see in these two
figures on this page three types of spent fuel dissolution models in
very high rank order when you plot peak mean dose versus for various
This vertical element of various types of contributors, such
as no retardation of plutonium in hot rock. The next three are
dissolution models of spent fuel and you see those are contributors and
contribute to dose. That applies to mean peak dose for 50,000 years as
The NRC base model for spent nuclear fuel dissolution is
more realistic compared with the DOE model. The dissolution rate from
tests in J-13 well water is much slower, as adopted by NRC. Therefore,
in NRC TPA 3.2, no credit was given to cladding and more realistic
juvenile failure of a container was allowed. This means we did not have
to introduce additional uncertainties associated with cladding credit,
as well as very low juvenile failure rate of a container by using the
realistic dissolution rate of spent nuclear fuel.
Page four describes the technical basis for spent nuclear
fuel. We considered all categories of spent nuclear fuel, including
commercial spent nuclear fuel, as well as DOE-owned spent nuclear fuel.
We identify all likely processes for spent nuclear degradation,
including matrix dissolution, colloid formation, prompt radionuclide
release, and dry oxidation. Among these processes, we have agreement
with DOE in many areas.
I would like to address later only areas where we still have
not arrived at issue resolutions.
I would like to show the typical numbers of dissolution
model for spent nuclear fuel in this table. The first one is the NRC
base case dissolution rate in the range of ten-to-the-minus-two scale
per day. On the other hand, DOE's conservative or accelerated test
result shows about two orders of magnitude higher than ours, and even
higher under drip conditions.
The basic difference is we used the data from immersion
tests in J-13 well water.
Another topic to be considered as a basis is the retardation
of plutonium. This is one potential aspect of colloidal contribution of
plutonium. The left figure is the TEDE for base case and the right
figure shows no retardation of plutonium on hot rock.
Within 10,000 years, because of the containment, we don't
see much increase of TEDE anyway. However, if you exceeded 10,000 years
and lose the integrity of a container, there is a potential for
significant contribution of plutonium released to dose. We would like
to investigate this aspect to improve the colloidal contribution model
On page six is the chemistry surrounding spent nuclear fuel.
In TSPA 3.2, we vary the chemistry of ground water on a drip to scale.
However, we have not conceded that the chemistry inside waste packages.
These three are highlights of the current understanding of
the chemistry inside a waste package. We expect very high concentration
of chloride and metal chloride complexes inside a waste package. For
instance, sustaining localized corrosion requires presence of metal
chloride complexes at the concentration of about 15 percent of
saturation. Hydrolysis of metal cations leads to extremely acidic pH
Also, we expect the dilution of this solution, which may
lead to cessation of pit growth.
Also, waste package internal environment is packed and hence
has many crevices where high concentration electrolytes will prevail.
That's one aspect of chemistry.
The other aspect is oxidizing conditions. We expect
oxidizing conditions near waste form for two reason; one, alpha
radiolysis will create highly oxidizing conditions close to the surface
of spent fuel; the second is, the packed regions near the fuel may have
other oxidized species such as Iron-3+.
However, currently, there is insufficient understanding of
the range of environments that may be present inside the waste package.
We have two comparative phenomenon. One, localized concentration, the
other one is bulk dilution.
However, we have modeling and experimental tools to
investigate the extent of variation of this internal environment, which
may affect spent fuel dissolution rate in a more conservative way.
On page seven is progress and current status, as I addressed
before. Most of the processes of degradation model was resolved between
NRC and DOE. However, we still need to resolve a few more based on TPA
sensitivity analysis which we conducted this year and last year.
We would like to evaluate further potential localized
reducing environments such as with the effect of iron and iron products.
The other one is to evaluate inter-laboratory data,
including our European data, on spent fuel dissolution rate.
We would like to sample model parameters based on these
further studies in future TPA 3.2 and we would like to assess the
chemistry inside the waste package further.
Again, the path forward, in general, was presented by
Gustavo. We will have continuous interaction with DOE and so on,
particularly to emphasize apply current models to new designs that may
result in different temperature and chemistry of ground water, and we
will elaborate with new DOE data and modify the current assessment with
the confirmatory testing program.
Page eight is on the cladding performance. Again, the risk
insight is the dose is sensitive to cladding protections. If I go back
to page three, the histogram of peak mean dose for 10,000 years and
50,000 years, you could compare the second, third and fourth spent fuel
dissolution model with a clad model, clad MI, you could see a very
sensitive protection of cladding in this plot of peak mean dose.
However, we have many uncertainties associated with the
cladding performance. Those need to be investigated in the future.
As a technical basis, we identify likely processes for
cladding degradation, such as localized corrosion, stress corrosion
cracking, hydride embrittlement, creep, mechanical failure, and initial
damage and conditions.
I would like to particularly address the initial damage and
conditions, because DOE relies on the cladding protection in their
Restrictions on temperature applies only to retrievability after
storage. What this means is during the storage, the environment could
be pretty high, increasing the temperature, causing these failure
We agreed with DOE in some of these aspects, but not all of
them. To resolve unresolved sub-issues, we conducted a TPA sensitivity
analysis and pursuing to evaluate localized corrosion.
As part of NRC rotational assignment to the center, Charles
Greene came here to initiate the localized corrosion cladding, as is
shown in figure nine, which shows the tendency of the localized
corrosion as a function of chloride concentration. At the very high
concentrated environment, there is a tendency to initiate the pitting
corrosion and cladding.
Currently, the center staff are continuing to finish this
subject. Also, Southwestern Research Institute scientist Ki Chan is
working with us to evaluate the mechanical failure of cladding.
Future path forward is, again, things presented by Gustavo
and I would like to emphasize we will apply current models to new
designs that may result in different temperature and chemistry of ground
water, and we will extend to evaluate other failure modes, like
localized corrosion or mechanical failure.
The last one is high level waste glass degradation. Here,
unlike the spent nuclear fuel and cladding performance, DOE, current DOE
models do not result in dose contribution in the TPA 3.2 exercise. We
took the DOE models and used in our TPA code and the contribution of
glass dissolution was negligible to ours.
The technical basis that we identified include various
degradation processes, such as stage three leaching processes, colloid
formation, hydride of glass surfaces prior to wetting, iron container
effect, and microbial actions.
We agreed to most of the aspects with DOE. However, we'd
still like to include three aspects. One is hydration. During the hot
humid environment, the glass can be hydrated and become soluble when
water came in. That can cause a very high dose later. We would like
DOE to consider that. And the colloidal from glass leaching, especially
plutonium release was 100 colloid, compared with plutonium release from
spent fuel. Therefore, we need to consider colloidal aspect in glass
In leaching of high level waste glass, there are three
stages. DOE considered only the first two stages. In stage three,
there is really accelerated leaching. We'd like DOE to consider that,
too. We will evaluate all these three aspects.
Also, probe some important aspect, currently the center
initiated a scoping test of high level waste glass, initiated by V. Jain
here. He would study effect of corrosion species in solution chemistry,
representative conditions near waste packages, and long-term PCT and
pressurized unsaturated flow test.
Also, we will continuously evaluate the uncertainties
associated with various added degradation modes.
The path forward is to apply current methods to new designs.
That may result in different temperature and chemistry of ground water.
In summary, using TPA 3.2, we conducted the system level
performance of waste form. In other words, we found out waste form
degradation models are very important in determining dose.
Second one is the dose is sensitive to modes of spent
nuclear fuel dissolution and cladding performance. A more realistic
model for spent nuclear fuel dissolution was used in performance
assessment of NRC. Consequently, we did not need to introduce
uncertainties associated with cladding protection or failure of
The technical bases for waste form degradation are established for
licensing review of DOE application. Uncertainties associated with
spent nuclear fuel, cladding performance, and high level waste glass are
further investigated and modeling will continue.
Tests of simulated high level waste glass and cladding have
Thank you. Any questions?
MR. WYMER: I just have a couple sort of minor ones. On the
high level waste glass degradation issue, you had three items numbered
here. One is risk insights, the second one was technical basis, and
then you had iron effects. Can you say a bit more about what iron
MR. AHN: Yes. I believe I wrote that. Yes. Page 10,
technical basis, iron effect. Iron has very low solubility. Therefore,
it tends to form colloid. It could be significant, yes.
MR. WYMER: Okay. All right.
MR. AHN: Yes. Yes. Yes. Yes.
MR. WYMER: Another-- it's not a question, it's a comment.
It seems to me you don't really need to worry too much about plutonium
in there, because there is not much plutonium left in the high level
waste glass, of course. That was what they were after in the
reprocessing, but it produced a waste that made -- there is a whole lot
less plutonium than there is in the spent fuel.
MR. AHN: Yes, absolutely. The absolute amount of release
is less compared with that from spent nuclear fuel.
MR. WYMER: By orders of magnitude.
MR. AHN: For instance, let's consider solubility limits.
Solubility, once you reach solubility limit with a limited amount of
water, it wouldn't matter how much you have inventory, unless you
deplete inventory. There is always a solubility limit. The value of
the solubility limits are the same for spent nuclear fuel and high level
Therefore, you would have a similar amount of concentration,
as long as you do not deplete the inventory of plutonium.
MR. WYMER: That's the point, of course.
MR. AHN: Yes.
MR. WYMER: You probably won't deplete the inventory in
spent fuel. You may well in glass.
MR. AHN: Yes. That takes quite a long time. I don't think
it would happen instantaneously.
MR. WYMER: Yes. Okay. I don't have any other questions
MR. GARRICK: I just wanted to comment. You indicated that
you did a more realistic analysis of the spent nuclear fuel dissolution.
MR. AHN: Yes.
MR. GARRICK: In the performance assessment. And that was
the reason you didn't take credit for cladding. And then you commented
that there is a lot of uncertainty associated with the cladding
MR. AHN: Yes.
MR. GARRICK: What if you did a realistic fuel dissolution
analysis in combination with a realistic consideration of the effect of
MR. AHN: We did that, actually, in sensitivity analysis.
You can barely see the dose in the histogram there.
MR. GARRICK: So the difference is --
MR. AHN: Right. We couldn't see the difference, yes.
MR. GARRICK: But you also make the point that they have to
reduce the uncertainties associated with the cladding analysis.
MR. AHN: Yes. Even without cladding protection, our dose
was very low. Therefore, we did not have to introduce the uncertainties
associated with the cladding performance. This is the kind of a systems
MR. GARRICK: Right.
MR. McCARTIN: Could I add one thing, Tae? When you're
saying you're adding in a realistic cladding model, I think you want to
be careful on that. I mean, that's using DOE's model and it's not to
say -- I don't think we know what a realistic cladding model is for the
behavior of cladding over thousands of years.
MR. AHN: Yes, right. As I presented it in the cladding
section, the cladding itself has about seven, eight different
degradation modes. I don't think we understand all of them at the
present time. All we know is a range of failure in a given time, even
though that is still uncertain.
We took some of those failure rates in TPA exercises and
found out it is sensitive to dose. That's what I meant.
MR. GARRICK: I'm still -- Tim maybe able to straighten me
out on this. I'm still trying to understand the difference between
understanding how to address cladding dissolution, for example, and
failure versus understanding how to address the dissolution of the fuel.
It seems to me, if you could do one, you ought to be able to
do the other.
MR. McCARTIN: I think Tae or Gustavo or some of the
materials people probably can answer that better. Some of the problem
with the cladding is you have unzipping and other failure modes beyond
dissolution and I guess what the cladding -- how it behaves over
thousands of years with potential rock fall hitting it and causing
MR. GARRICK: I guess my point, though, Tim, is you have the
same problem with the fuel. You're talking about --
MR. AHN: Yes.
MR. GARRICK: -- it and how it behaves over thousands of
MR. AHN: DOE has studied fuel at a long period of time and
accumulated a database. On the other hand, cladding credit has been
given only on a few years ago. So they have a limited database compared
with the fuel itself.
MR. GARRICK: Okay. We maybe we need to talk about that
some other time.
MR. WYMER: We have a little bit more time anyway. I'd like
to hear you say a few things more about the actual degradation
dissolution of the spent fuel material.
MR. AHN: Yes.
MR. WYMER: The oxidation and solubilization by carbonate
complexes and things like that. What more can you add in detailed
MR. AHN: There are two types of release of radionuclide.
One is the release of very highly soluble radionuclide, such as
technetium, iodine. The other one is a very low solubility radionuclide
release, such as plutonium. Neptunium is in between. In the release of
low solubility radionuclides, always releases depend on either
solubility limits or colloid concentrations, times flow rate is the
On the other hand, highly soluble radionuclides will be
released congruently with the matrix dissolution of fuels. That's why
it's so important to determine that the solution of bare spent nuclear
fuel because technetium and iodine measures contribute to dose.
MR. WYMER: I'm sure you know, because I've seen some stuff
that has been written about --
MR. AHN: Yes.
MR. WYMER: -- technetium. But you do know it, but I'll say
it for the benefit of everybody here, some people may not know that in
the spent fuel, approximately a third of the technetium is metallic
particles. I know you know that.
MR. AHN: Right, yes.
MR. WYMER: So there is -- and the rest of it is probably
dispersed maybe dioxide in the field.
MR. AHN: Yes.
MR. WYMER: And that has to be further oxidized and there is
a kinetic -- that's kinetically controlled and it's kind of refractory.
It doesn't just leap up to the --
MR. AHN: Yes.
MR. WYMER: So have you factored that into your --
MR. AHN: Our approach is conservative. In other words,
once particles are released as a matrix, we assume all particles
contribute to dose. They may not contribute to the dose.
MR. WYMER: So you're not taking any credit for about a
one-third reduction in the dose, because --
MR. AHN: No.
MR. WYMER: -- it's not going to go anywhere.
MR. AHN: Right, right, right.
MR. WYMER: Okay.
MR. GARRICK: Any other -- Charles, do you have any
MR. FAIRHURST: No.
MR. GARRICK: George?
MR. HORNBERGER: No.
MR. GARRICK: Thank you.
MR. AHN: Thank you.
MR. GARRICK: The program calls for this to be an open
discussion. I guess one way to interpret that would be to ask the
committee if -- on reflecting on the various presentations that were
made today, are there questions, comments that you either forgot earlier
and would like to pick up on now or do you have some second thoughts
about some of the things that were said? Charles, do you have any
MR. FAIRHURST: In the talk by Dr. Cragnolino, I know that
you carried out the performance calculations up to 10,000 years. Did
you carry them out any longer?
MR. CRAGNOLINO: Yes. We've done calculation for longer
time, but I don't see that we're exploring this particular issue at a
MR. FAIRHURST: As far as maximum dose, how high did it go?
Of course, it varied for a lot of other factors.
MR. CRAGNOLINO: Yes. I can't recall. I have to look in
MR. FAIRHURST: That's all right. Maybe we can just talk
MR. WYMER: While we're waiting for him to come back, I'll
make a general observation. It disturbs me some that the chemistry of
these systems are so complex and the databases, in many cases, are so
inadequate at the moment, what's in the literature is either not
directly applicable or it isn't in the literature, that you can't really
take advantage of a lot of the things that I think many of us feel are
there that would allow you to get rid of some of the conservatism that's
in these estimates.
The doses are low anyway, so maybe the conservatism isn't a
problem, but if it turns out that it is a problem, it would be nice to
be able to use some of these chemical attributes of a system that tend
to reduce the dose as a backstop, as a backup for the possible
contingency that there is something unexpected comes up.
So just a general observation, I'm disturbed that we don't
-- we aren't further along with the chemistry. We spent too much time
on the geology.
MR. GARRICK: That's true for a lot of things.
MR. AHN: Dr. Wymer, this is Tae Ahn. I presented in page
seven a current status the localized environment and the iron effects.
MR. WYMER: Yes. We need all of that we can get.
MR. AHN: As Dr. Wymer suggested, we are evaluating
potential reducing environment and iron effect for spent fuel
dissolution, as well as glass leaching.
MR. WYMER: I'm just concerned that the timeframe is such
that with the time you're having to evaluate this license application,
you still will not have been able to acquire all the information you
need or by the time the model changes again and there are other factors
that are important, there won't be time enough to patch up the holes
with data and that information.
MR. AHN: We have limited effort to conduct experiment
itself. However, we do analyze the performance assessment based on
MR. WYMER: I certainly am supporting a much increased
MR. AHN: Thank you.
MR. CAMPBELL: Tae Ahn, before you get away from the mic,
let me turn that question around a little bit. What would appear to be
seen, and certainly in the tech exchange back at the end of May, are the
microrem doses, tens of microrems typically, certainly in the 10,000
year period. Because of that, you are using and DOE is using
Would you feel that there are specific problem areas that
the NRC would have in terms of reasonable assurance that would prevent
licensing from going forward, even if you don't have all the data that
you want at licensing?
MR. AHN: I have one concern. Suppose you use any
conservative dissolution rate obtained in pure carbonate solution,
without cladding, the dose may go up to ten millirem. Then it's 25. If
you consider uncertainties and so on, then you might consider once more
the cladding protection and so on.
MR. CAMPBELL: So the issue is that you have -- you believe
that they have a very conservative dissolution model.
MR. AHN: Right.
MR. CAMPBELL: And they make up for it with a more
optimistic, if you will, approach to cladding.
MR. AHN: Right.
MR. CAMPBELL: If that --
MR. AHN: If that is the case --
MR. CAMPBELL: In support of adversarial discussion and it
becomes not accepted, then you're stuck with it.
MR. AHN: Right. That's what we are going to discuss next
MR. WYMER: Also, if you don't have enough information on
secondary phase formation that sort of stands in for the failure of
MR. AHN: Yes.
MR. WYMER: If you haven't got a convincing database and
models, then you're also in trouble.
MR. AHN: Right, yes.
MR. WYMER: So it seems to me that this is an important area
MR. AHN: Yes.
MR. WYMER: Aggressively.
MR. AHN: Yes.
MR. McCARTIN: Tim McCartin, NRC. Once again, though, DOE
is going to have to collect all that information and support their
models and we aren't trying to develop these models for the.
MR. WYMER: No. But unless you can evaluate their models
based on good information that you believe, it may not be good enough.
MR. McCARTIN: Well, their license application will have to
contain that information. There will have to be support, with a
recognition that there is a performance confirmation period that lasts a
very long time, which will supplement the information.
MR. WYMER: Okay.
MR. SRIDHAR: Can I add something to this? I hope I can add
something to it. But my -- I understood your question mainly pertaining
to spent fuel dissolution, not to the waste package.
MR. WYMER: That's right. That's what we were talking about
MR. SRIDHAR: I think there are two questions that we are
trying to answer in relation to the cladding failure. It's really how
the holes in the container are going to be situated, because that seems
to be a very critical parameter, and we have this backed up model and we
are doing sensitivity analysis on the height of the hole and so on, but
that is something we need to know and that would actually have more of
an effect in eventual release.
We don't want to be prisoners of the chemistry issue in some
way. I think in the container corrosion issue, we have addressed that
by saying that based on our experience, on experimental data and so on,
we limit the number of species that we need to understand. Otherwise,
we would be chasing all kinds of cats and dogs in the mountain, and we
are saying in the container corrosion, we are mainly interested in the
chloride and the corrosion potential.
I believe that in doing spent fuel dissolution type of
modeling, we have been using mainly DOE data based on the immersion
corrosion tests, but we haven't really taken a mechanistic approach to
predicting spent fuel dissolution rate.
One of the things the Canadians have done a lot better job
than we have is really approach it from a mechanistic point of view on
why spent fuel corrodes the way it does. They're using an
electro-chemical approach, because spent fuel does dissolve and is
electro-chemically active. I think that would provide us a mechanism,
that's my hope, to eliminate a lot of the unnecessary information needs
and help us focus on the few chemistry details that we need to know.
For example, redox potential, we need to know that and
perhaps carbonate and sulfate and so on. But we may not need to know
everything about the environment.
MR. WYMER: That gets back to the earlier point, I think, of
the importance of understanding mechanisms as opposed to just
Your confidence in extrapolations into the future and your
confidence in the information you gather in the pre-closure period is
pretty much based on your understanding of the mechanisms that really
are controlling things rather than just taking data.
I know you're going after mechanisms and I think that's the
right thing to do.
MR. SRIDHAR: Right, and we need to do more of that for
MR. WYMER: Yes. And it needs to be more broadly applied,
that concept. That's right. Thank you.
MR. GARRICK: George, anything?
MR. HORNBERGER: No.
MR. GARRICK: We have a sidebar going on here.
MR. CRAGNOLINO: We're getting a copy of that. This
clarifies only one question that I had before from Dr. Fairhurst, and
this referred to the multiple realization to calculate radionuclide
dose, and obviously it was not presented in detail here, but for the
technical exchange with the DOE on May 27, we showed the calculation up
to 100,000 years for the base case and the peak expected dose was four
millirem per year. Answering your question.
Let me make a point. I have a little disagreement with what
Sridhar said before regarding the spent fuel. I think that it is very
important in this approach, the electro-chemical approach to the
dissolution rate of the spent fuel. But one of the conclusions that
they have is that even though this approach gives you the maximum rate
of dissolution, and allows you to map out the behavior in terms of the
redox condition, there was the issue of alteration products on the spent
fuel surface that could come.
The potential effect of the slowing down the process, but also later on,
by exfoliation, produce release, and I think that this is an issue that
has to be addressed mostly by the DOE, obviously, but we have to have a
more solid type of approach.
I don't think they have the standard that we have here about
this problem and I think that this is something that has to be -- and it
has to be done also in relation to the models that we have for spent
fuel. We need to keep looking at these and trying to improve these and
improve the technical basis for these.
MR. FAIRHURST: Could you help clarify for me something?
Because in the VA analysis, these doses were up to 25 millirem or more.
MR. HORNBERGER: That's DOE's analysis.
MR. FAIRHURST: I know. But I'm trying to say what is --
could you identify roughly what was different in those realizations than
these. I see you've got some specific statements about travel time in
the unsaturated zone and so on.
MR. WYMER: Time scale.
MR. CRAGNOLINO: I think that I -- this was my intention to
answer your question, because he is more in the condition of this, but
that Gordon might clarify this particular issue.
MR. FAIRHURST: Okay. Good.
MR. WITTMEYER: Gordon Wittmeyer, from the center.
Sitiganta has graciously allowed me to answer instead. Really, it's
very hard to compare the VA and TSPA/VA to the TPA results here because
there are so many different approaches, that they mask what the
essential differences are.
One thing I could point to, and, Tim, please chime in when
you feel it's appropriate. One thing is we have a much different
approach to computing the concentration in the dose at the receptor
location. That alone could account for a difference that if we're
following the numbers here, the difference between four millirem and, I
believe, 25 millirem. A factor of roughly six.
But there are so may things. It's very hard to tell you
what the differences are unless we go through and kind of do one-up
analyses on each one of the models very carefully.
So that's a very good non-answer, I think.
MR. FAIRHURST: Let me come back to the second question. If
your answer is not quite an order of magnitude different, four to 25 is
getting close, first, that yours -- I'm saying if this is not the real
license application, but yours is lower, but what -- you're the one that
has to decide whether it's acceptable or not, right? So how are you
going to get at these hidden assumptions that are made in the
MR. WITTMEYER: We get in there and we actually probe very
carefully each of the model assumptions used in the TSPA/VA. We have
done that. In some cases, we'll take our best understanding of their
models, try and mimic them with our models by changing parameters.
We always get -- we can't always do that in a
straightforward way, but we think --
MR. FAIRHURST: I understand.
MR. WITTMEYER: But we think try to see what those
differences are. It's just hard, dirty work going through there one at
MR. FAIRHURST: Thank you.
MR. GARRICK: Now, before we sign off, I want to ask
Washington and Las Vegas if they have any comments that they would care
to make. Let's start with Washington, since it's later there.
MR. McCARTIN: I think we're all pretty happy here, I guess.
Dick, did you have something that you wanted to add?
MR. CODELL: This is Dick Codell from NRC. I would be
interested in getting the committee's view on the importance of
colloids, if they have anything to add to our performance assessments.
I personally feel that it's blown out of proportion, but the little
studies that we've done at NRC are on very over-simplified modeling.
I don't think that they're realistic at all. They're put
out there really to generate some controversy so we can get some
discussion going. When you see concerns about colloids, like the recent
discovery of plutonium colloids at the Nevada test site from the Benham
nuclear test, that raises public's concern about it. That's why we need
to put it to rest, if we can.
So does the committee have any thoughts on this issue?
MR. GARRICK: I don't know if we've advanced our thinking on
it to a point of -- but, Ray, I know you have some thoughts about it.
MR. WYMER: These are just off the cuff, but that's all
you're going to expect to get.
I think that primary colloids, like americium and plutonium,
are not too likely to make any real contribution. The importance of
secondary colloids, like the solutions material or clays or the iron
hydroxide type colloids, I think you've got to take a pretty good look
at those because they do absorb things pretty strongly and you'll have
conditions where the colloids will form. But I don't think you can
write it off without a little bit more attention.
MR. GARRICK: Do you have any comments about colloids?
MR. HORNBERGER: Well, there is no doubt that certainly in
the saturated zone, as you indicated, Dick, in the Benham test, there is
a very clear indication that there has been movement over a kilometer or
so, several hundred meters anyway.
The real question really is whether or not any of the
colloids will be mobile to any extent in the unsaturated zone and there
are -- I think there are serious doubts that the colloids will be an
issue because of that.
Again, how do you lay it to rest? I think that there is
some work going on, there is some work going on at Lawrence Berkeley and
elsewhere, I guess at Sandia. DOE is certainly doing, I think, doing
enough experimental work, that data should be available.
MR. GARRICK: Any other comments?
MR. McCARTIN: We have one more comment from Washington and
it's just somewhat of a supplement to what Gordon said about comparing
the DOE and the NRC results.
He's absolutely correct, there are many, many different
aspects to the two analyses that are different and it's hard to get a
one-to-one correspondence. However, in doing the -- when we do a
license application, we are going to have to understand where the
differences are and what it means, whether we ask DOE to do some
additional analyses with slight changes in parameters, or we do them.
Obviously, we'll have to wait and see what the analyses look
like, but we will -- we certainly have a good idea of where our
differences are and the fact that we have four versus 25, we're
certainly in the same ballpark.
But let me say that one thing that really makes a comparison
difficult is the fact that there is a strict time limit there. And you
can change that dose by moving a little bit further or a shorter time
period or a longer time period and there are things that delay the dose.
If we're going all to a peak no matter when it occurred, it
might be that the comparison would be easier, but I think some of that
could be just due to where in time we decide to cut off the dose.
MR. GARRICK: That's a good comment. Any other comments
with Washington? How about Las Vegas? Any comments from Las Vegas?
Did we drop them or lose them?
MR. LEE: Not to my knowledge.
MR. GARRICK: Just no comment. At least there ought to be a
comment that there is no comment. I think they've had enough. All
right. Any other comments from the room or from the staff?
I know that Andy has an announcement he wants to make, do
you not? Or Lynn.
MS. DEERING: I just wanted to announce that there is a
one-on-one informal discussion planned with George Hornberger tomorrow.
It's going to take place at 7:30 a.m. here and the subjects would
include thermal effects, saturated/unsaturated zone or igneous activity,
those are possible topics. We don't have anything particular planned.
But I think Dr. Fairhurst will also join us.
The plan was to get a room, and I don't have a room yet
here, but we were going to contact Robert Johnson in the morning and let
them know what the bridge number is, if there is anybody in Washington
who would like to join that discussion.
MR. CAMPBELL: And as far as the other individual meetings,
those will be at 1:00 tomorrow.
MR. GARRICK: Central time. All right. Any parting
comments from anybody? I thought we had a very good day, covered a lot
of material, saw some very interesting experiments, heard some
especially insightful remarks, and with that, I think we will adjourn.
[Whereupon, at 5:00 p.m., the meeting was recessed, to
reconvene at 8:30 a.m., Wednesday, June 30, 1999.]
Page Last Reviewed/Updated Friday, September 29, 2017