77. june 8 thmc model (illitization) and thm model (tphm hooke's) clay-zheng

1. Used Fuel Disposition Campaign
Integrating Coupled
Processes into PA model
Liange Zheng, Jonny Rutqvist, Jens Birkholzer
Lawrence Berkeley National Laboratory
UFD Annual Meeting, June 6-9, 2016

2. Used
Fuel
Disposition
Features and Focuses of the
coupled THMC Model
 Focuses of the model
 Illitization, i.e. the transformation of smectite to illite
 Geochemically induced swelling stress change
 Long term evolution of hydrological (e.g. permeability) and chemical (e.g sorption
capacity) changes as a results of coupled THMC evolution..
 Major features of the model
 Coupled THMC processes, especially MC coupling
 Interactions between EBS bentonite and host rock, canister and EBS bentonite
 High temperature
 Current status
 Coupled THMC model had been developed for a generic clay rock repository with
bentonite backfilled EBS.
 FY16 model incorporate canister-bentonite interaction, but chemical model needs to be
fine-tuned
 Two chemical-mechanical coupling schemes for bentonite have been tested: extended
linear swelling and dual structure BExM

3. Used
Fuel
Disposition
Snapshot of the Model
20
40
60
80
100
120
140
160
180
200
220
0.001 0.1 10 1000 100000
T(oC)
Time (year)
point A
high T
low T
-0.080
-0.070
-0.060
-0.050
-0.040
-0.030
-0.020
-0.010
0.000
0.001 0.1 10 1000 100000
Volumefractionchange
Time (year)
smectite, A
high T, Kunigel bentonite
high T, FEBEX bentonite
low T, Kunigel Bentonite
low T, FEBEX bentonite
0
0.2
0.4
0.6
0.8
1
1.2
0.001 0.01 0.1 1 10 100 1000 10000 100000
Swellingstress(MPa)
Time (year)
Point B
high T, S=f(Sl,C,Sc)
high T, S=f(Sl,C)
high T, S=f(Sl)
lowT, S=f(Sl,C,Sc)
lowT, S=f(Sl,C)
lowT, S=f(Sl)

4. Used
Fuel
Disposition
Significant R&D is needed to improve and validate the model
Difficulties for direct coupling with PA model:
PA model needs to be able to simulating THMC process
simultaneously and consider the couplings between different processes
Detailed coupling between THMC process might be too expensive to
run in PA model
Abstraction is more doable. Response surface, for example,
illitization as a function of key flow, chemical parameters and
time, can be established based on coupled THMC model and
integrated to the PA model.
Integration with PA Model

5. Used
Fuel
Disposition
Non-Darcian Flow:
Background and Proposed model
Non-Darcian flow is evidenced in laboratory experiments
and nanoscale flow simulation.
 It is critical for the advection in Excavation Damaged Zone
(EDZ) and water flow from host rock to EBS for buffer to
function (swelling) properly.
5
)]
*
,
1
(
)
1
(
[
α
α
γ
α
γ






−=
I
iI
iKq
∫
−−
=
x
ta
dtetxa
0
1
),(γ ( ) ∫
∞
−−
=
0
1
dteta ta
γ
Proposed model (Liu and Birkholzer, 2013) :

6. Used
Fuel
Disposition
Relationship between permeability and threshold hydraulic gradient
I
Permeability(m2
)
10
-1
10
0
10
1
10
2
10
3
10
4
10
5
10
-20
10
-19
10
-18
10-17
10-16
10-15
10-14
10-13
Wang et al.(2011)
Dubin and Moulin (1986)
Cui et al. (2008)
Blecker (1970)
Miller and Low (1963)
Lutz and Kemper (1959)
Zou (1996)
Curve fitting
B
AkI =
A= 4.0 E-12 and B =-0.78
6
The degree of non-Darcian flow behavior can be characterized by
permeability (k) (or pore size) and threshold gradient (I), respectively.
Liu and Birkholzer, 2013
Non-Darcian Flow:
Proposed Model

7. Used
Fuel
Disposition
Non-Darcian Flow:
Implementation and Testing
7
Non-Darcian flow model was implemented in TOUGH2 and
had been used to model the FEBEX in situ test
Observations: TH model with
Non-Darcian flow significantly
underestimates RH data. Plausible
reasons:
Non-Darcian flow VS relative
permeability
Threshold gradient is affected by
capillary pressure
Lessons learned: processes
uncertainties VS parameters
uncertainties

8. Used
Fuel
Disposition
Hydrological parameters needs to be established for Non-
Darcian flow
Coefficients for the permeability-threshold gradient needs
to be calibrated
The effect of temperature, salinity and capillary pressure
on Non-Darcian flow need to be better understood
Direct integration can be done: integration might be easy
but parameters calibration is challenging
Integration with PA Model